Abstract
Order-driven market simulation mimics the trader behaviors to generate order streams to support interactive studies of financial strategies. In market simulator, the multi-agent approach is commonly adopted due to its explainability. Existing multi-agent systems employ heuristic search to generate order streams, which is inefficient for large-scale simulation. Furthermore, the search-based behavior calibration often leads to inconsistent trader actions under the same general market condition, making the simulation results unstable and difficult to interpret. We propose CaliSim, the first search-free calibration approach multi-agent market simulator which achieves large-scale efficiency and behavior consistency. CaliSim uses meta-learning and devises a surrogate trading system with a consistency loss function for the reproducibility of order stream and trader behaviors. Extensive experiments in the market replay and case studies show that CaliSim achieves state-of-the-art in terms of order stream reproduction with consistent trader behavior and can capture patterns of real markets.
An Explainable Geometric-Weighted Graph Attention Network for Identifying Functional Networks Associated with Gait Impairment
Authors: Favour Nerrise (1), Qingyu Zhao (2), Kathleen L. Poston (3), Kilian M. Pohl (2), Ehsan Adeli (2) ((1) Department of Electrical Engineering, Stanford University, Stanford, CA, USA, (2) Dept. of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA, USA, (3) Dept. of Neurology and Neurological Sciences, Stanford University, Stanford, CA, USA)
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Image and Video Processing (eess.IV); Neurons and Cognition (q-bio.NC)
Abstract
One of the hallmark symptoms of Parkinson's Disease (PD) is the progressive loss of postural reflexes, which eventually leads to gait difficulties and balance problems. Identifying disruptions in brain function associated with gait impairment could be crucial in better understanding PD motor progression, thus advancing the development of more effective and personalized therapeutics. In this work, we present an explainable, geometric, weighted-graph attention neural network (xGW-GAT) to identify functional networks predictive of the progression of gait difficulties in individuals with PD. xGW-GAT predicts the multi-class gait impairment on the MDS Unified PD Rating Scale (MDS-UPDRS). Our computational- and data-efficient model represents functional connectomes as symmetric positive definite (SPD) matrices on a Riemannian manifold to explicitly encode pairwise interactions of entire connectomes, based on which we learn an attention mask yielding individual- and group-level explainability. Applied to our resting-state functional MRI (rs-fMRI) dataset of individuals with PD, xGW-GAT identifies functional connectivity patterns associated with gait impairment in PD and offers interpretable explanations of functional subnetworks associated with motor impairment. Our model successfully outperforms several existing methods while simultaneously revealing clinically-relevant connectivity patterns. The source code is available at https://github.com/favour-nerrise/xGW-GAT .
Multi-UAV Speed Control with Collision Avoidance and Handover-aware Cell Association: DRL with Action Branching
Authors: Zijiang Yan, Wael Jaafar, Bassant Selim, Hina Tabassum
Subjects: Machine Learning (cs.LG); Robotics (cs.RO); Systems and Control (eess.SY)
Abstract
This paper presents a deep reinforcement learning solution for optimizing multi-UAV cell-association decisions and their moving velocity on a 3D aerial highway. The objective is to enhance transportation and communication performance, including collision avoidance, connectivity, and handovers. The problem is formulated as a Markov decision process (MDP) with UAVs' states defined by velocities and communication data rates. We propose a neural architecture with a shared decision module and multiple network branches, each dedicated to a specific action dimension in a 2D transportation-communication space. This design efficiently handles the multi-dimensional action space, allowing independence for individual action dimensions. We introduce two models, Branching Dueling Q-Network (BDQ) and Branching Dueling Double Deep Q-Network (Dueling DDQN), to demonstrate the approach. Simulation results show a significant improvement of 18.32% compared to existing benchmarks.
Template-Based Static Posterior Inference for Bayesian Probabilistic Programming
Authors: Peixin Wang, Hongfei Fu, Tengshun Yang, Guanyan Li, Luke Ong
Abstract
In Bayesian probabilistic programming, a central problem is to estimate the normalised posterior distribution (NPD) of a probabilistic program with conditioning. Prominent approximate approaches to address this problem include Markov chain Monte Carlo and variational inference, but neither can generate guaranteed outcomes within limited time. Moreover, most existing formal approaches that perform exact inference for NPD are restricted to programs with closed-form solutions or bounded loops/recursion. A recent work (Beutner et al., PLDI 2022) derived guaranteed bounds for NPD over programs with unbounded recursion. However, as this approach requires recursion unrolling, it suffers from the path explosion problem. Furthermore, previous approaches do not consider score-recursive probabilistic programs that allow score statements inside loops, which is non-trivial and requires careful treatment to ensure the integrability of the normalising constant in NPD. In this work, we propose a novel automated approach to derive bounds for NPD via polynomial templates. Our approach can handle probabilistic programs with unbounded while loops and continuous distributions with infinite supports. The novelties in our approach are three-fold: First, we use polynomial templates to circumvent the path explosion problem from recursion unrolling; Second, we derive a novel multiplicative variant of Optional Stopping Theorem that addresses the integrability issue in score-recursive programs; Third, to increase the accuracy of the derived bounds via polynomial templates, we propose a novel technique of truncation that truncates a program into a bounded range of program values. Experiments over a wide range of benchmarks demonstrate that our approach is time-efficient and can derive bounds for NPD that are comparable with (or tighter than) the recursion-unrolling approach (Beutner et al., PLDI 2022).
Navigating the Web of Misinformation: A Framework for Misinformation Domain Detection Using Browser Traffic
Authors: Mayana Pereira, Kevin Greene, Nilima Pisharody, Rahul Dodhia, Jacob N. Shapiro, Juan Lavista
Abstract
The proliferation of misinformation and propaganda is a global challenge, with profound effects during major crises such as the COVID-19 pandemic and the Russian invasion of Ukraine. Understanding the spread of misinformation and its social impacts requires identifying the news sources spreading false information. While machine learning (ML) techniques have been proposed to address this issue, ML models have failed to provide an efficient implementation scenario that yields useful results. In prior research, the precision of deployment in real traffic deteriorates significantly, experiencing a decrement up to ten times compared to the results derived from benchmark data sets. Our research addresses this gap by proposing a graph-based approach to capture navigational patterns and generate traffic-based features which are used to train a classification model. These navigational and traffic-based features result in classifiers that present outstanding performance when evaluated against real traffic. Moreover, we also propose graph-based filtering techniques to filter out models to be classified by our framework. These filtering techniques increase the signal-to-noise ratio of the models to be classified, greatly reducing false positives and the computational cost of deploying the model. Our proposed framework for the detection of misinformation domains achieves a precision of 0.78 when evaluated in real traffic. This outcome represents an improvement factor of over ten times over those achieved in previous studies.
Abstract
Emotion regulation is the process of consciously altering one's affective state, that is the underlying emotional state such as happiness, confidence, guilt, anger etc. The ability to effectively regulate emotions is necessary for functioning efficiently in everyday life. Today, the pervasiveness of digital technology is being purposefully employed to modify our affective states, a process known as digital emotion regulation. Understanding digital emotion regulation can help support the rise of ethical technology design, development, and deployment. This article presents an overview of digital emotion regulation in social media applications, as well as a synthesis of recent research on emotion regulation interventions for social media. We share our findings from analysing state-of-the-art literature on how different social media applications are utilised at different stages in the process of emotion regulation.
Multilevel Large Language Models for Everyone
Authors: Yuanhao Gong
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Distributed, Parallel, and Cluster Computing (cs.DC); General Economics (econ.GN)
Abstract
Large language models have made significant progress in the past few years. However, they are either generic {\it or} field specific, splitting the community into different groups. In this paper, we unify these large language models into a larger map, where the generic {\it and} specific models are linked together and can improve each other, based on the user personal input and information from the internet. The idea of linking several large language models together is inspired by the functionality of human brain. The specific regions on the brain cortex are specific for certain low level functionality. And these regions can jointly work together to achieve more complex high level functionality. Such behavior on human brain cortex sheds the light to design the multilevel large language models that contain global level, field level and user level models. The user level models run on local machines to achieve efficient response and protect the user's privacy. Such multilevel models reduce some redundancy and perform better than the single level models. The proposed multilevel idea can be applied in various applications, such as natural language processing, computer vision tasks, professional assistant, business and healthcare.
Rank Optimization for MIMO systems with RIS: Simulation and Measurement
Authors: Shengguo Meng, Wankai Tang, Weicong Chen, Jifeng Lan, Qun Yan Zhou, Yu Han, Xiao Li, Shi Jin
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
Reconfigurable intelligent surface (RIS) is a promising technology that can reshape the electromagnetic environment in wireless networks, offering various possibilities for enhancing wireless channels. Motivated by this, we investigate the channel optimization for multiple-input multiple-output (MIMO) systems assisted by RIS. In this paper, an efficient RIS optimization method is proposed to enhance the effective rank of the MIMO channel for achievable rate improvement. Numerical results are presented to verify the effectiveness of RIS in improving MIMO channels. Additionally, we construct a 2$\times$2 RIS-assisted MIMO prototype to perform experimental measurements and validate the performance of our proposed algorithm. The results reveal a significant increase in effective rank and achievable rate for the RIS-assisted MIMO channel compared to the MIMO channel without RIS.
Social Optimum Equilibrium Selection for Distributed Multi-Agent Optimization
Abstract
We study the open question of how players learn to play a social optimum pure-strategy Nash equilibrium (PSNE) through repeated interactions in general-sum coordination games. A social optimum of a game is the stable Pareto-optimal state that provides a maximum return in the sum of all players' payoffs (social welfare) and always exists. We consider finite repeated games where each player only has access to its own utility (or payoff) function but is able to exchange information with other players. We develop a novel regret matching (RM) based algorithm for computing an efficient PSNE solution that could approach a desired Pareto-optimal outcome yielding the highest social welfare among all the attainable equilibria in the long run. Our proposed learning procedure follows the regret minimization framework but extends it in three major ways: (1) agents use global, instead of local, utility for calculating regrets, (2) each agent maintains a small and diminishing exploration probability in order to explore various PSNEs, and (3) agents stay with the actions that achieve the best global utility thus far, regardless of regrets. We prove that these three extensions enable the algorithm to select the stable social optimum equilibrium instead of converging to an arbitrary or cyclic equilibrium as in the conventional RM approach. We demonstrate the effectiveness of our approach through a set of applications in multi-agent distributed control, including a large-scale resource allocation game and a hard combinatorial task assignment problem for which no efficient (polynomial) solution exists.
A Model Predictive Capture Point Control Framework for Robust Humanoid Balancing via Ankle, Hip, and Stepping Strategies
Authors: Myeong-Ju Kim, Daegyu Lim, Gyeongjae Park, Jaeheung Park
Abstract
The robust balancing capability of humanoid robots against disturbances has been considered as one of the crucial requirements for their practical mobility in real-world environments. In particular, many studies have been devoted to the efficient implementation of the three balance strategies, inspired by human balance strategies involving ankle, hip, and stepping strategies, to endow humanoid robots with human-level balancing capability. In this paper, a robust balance control framework for humanoid robots is proposed. Firstly, a novel Model Predictive Control (MPC) framework is proposed for Capture Point (CP) tracking control, enabling the integration of ankle, hip, and stepping strategies within a single framework. Additionally, a variable weighting method is introduced that adjusts the weighting parameters of the Centroidal Angular Momentum (CAM) damping control over the time horizon of MPC to improve the balancing performance. Secondly, a hierarchical structure of the MPC and a stepping controller was proposed, allowing for the step time optimization. The robust balancing performance of the proposed method is validated through extensive simulations and real robot experiments. Furthermore, a superior balancing performance is demonstrated, particularly in the presence of disturbances, compared to a state-of-the-art Quadratic Programming (QP)-based CP controller that employs the ankle, hip, and stepping strategies. The supplementary video is available at https://youtu.be/CrD75UbYzdc
Federated Split Learning with Only Positive Labels for resource-constrained IoT environment
Abstract
Distributed collaborative machine learning (DCML) is a promising method in the Internet of Things (IoT) domain for training deep learning models, as data is distributed across multiple devices. A key advantage of this approach is that it improves data privacy by removing the necessity for the centralized aggregation of raw data but also empowers IoT devices with low computational power. Among various techniques in a DCML framework, federated split learning, known as splitfed learning (SFL), is the most suitable for efficient training and testing when devices have limited computational capabilities. Nevertheless, when resource-constrained IoT devices have only positive labeled data, multiclass classification deep learning models in SFL fail to converge or provide suboptimal results. To overcome these challenges, we propose splitfed learning with positive labels (SFPL). SFPL applies a random shuffling function to the smashed data received from clients before supplying it to the server for model training. Additionally, SFPL incorporates the local batch normalization for the client-side model portion during the inference phase. Our results demonstrate that SFPL outperforms SFL: (i) by factors of 51.54 and 32.57 for ResNet-56 and ResNet-32, respectively, with the CIFAR-100 dataset, and (ii) by factors of 9.23 and 8.52 for ResNet-32 and ResNet-8, respectively, with CIFAR-10 dataset. Overall, this investigation underscores the efficacy of the proposed SFPL framework in DCML.
Abstract
A biologically plausible method for training an Artificial Neural Network (ANN) involves treating each unit as a stochastic Reinforcement Learning (RL) agent, thereby considering the network as a team of agents. Consequently, all units can learn via REINFORCE, a local learning rule modulated by a global reward signal, which aligns more closely with biologically observed forms of synaptic plasticity. Nevertheless, this learning method is often slow and scales poorly with network size due to inefficient structural credit assignment, since a single reward signal is broadcast to all units without considering individual contributions. Weight Maximization, a proposed solution, replaces a unit's reward signal with the norm of its outgoing weight, thereby allowing each hidden unit to maximize the norm of the outgoing weight instead of the global reward signal. In this research report, we analyze the theoretical properties of Weight Maximization and propose a variant, Unbiased Weight Maximization. This new approach provides an unbiased learning rule that increases learning speed and improves asymptotic performance. Notably, to our knowledge, this is the first learning rule for a network of Bernoulli-logistic units that is unbiased and scales well with the number of network's units in terms of learning speed.
QuIP: 2-Bit Quantization of Large Language Models With Guarantees
Authors: Jerry Chee, Yaohui Cai, Volodymyr Kuleshov, Christopher De Sa
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
This work studies post-training parameter quantization in large language models (LLMs). We introduce quantization with incoherence processing (QuIP), a new method based on the insight that quantization benefits from incoherent weight and Hessian matrices, i.e., from the weights and the directions in which it is important to round them accurately being unaligned with the coordinate axes. QuIP consists of two steps: (1) an adaptive rounding procedure minimizing a quadratic proxy objective; (2) efficient pre- and post-processing that ensures weight and Hessian incoherence via multiplication by random orthogonal matrices. We complement QuIP with the first theoretical analysis for an LLM-scale quantization algorithm, and show that our theory also applies to an existing method, OPTQ. Empirically, we find that our incoherence preprocessing improves several existing quantization algorithms and yields the first LLM quantization methods that produce viable results using only two bits per weight. Our code can be found at https://github.com/jerry-chee/QuIP .
Federated Heavy Hitter Recovery under Linear Sketching
Authors: Adria Gascon, Peter Kairouz, Ziteng Sun, Ananda Theertha Suresh
Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR); Information Theory (cs.IT)
Abstract
Motivated by real-life deployments of multi-round federated analytics with secure aggregation, we investigate the fundamental communication-accuracy tradeoffs of the heavy hitter discovery and approximate (open-domain) histogram problems under a linear sketching constraint. We propose efficient algorithms based on local subsampling and invertible bloom look-up tables (IBLTs). We also show that our algorithms are information-theoretically optimal for a broad class of interactive schemes. The results show that the linear sketching constraint does increase the communication cost for both tasks by introducing an extra linear dependence on the number of users in a round. Moreover, our results also establish a separation between the communication cost for heavy hitter discovery and approximate histogram in the multi-round setting. The dependence on the number of rounds $R$ is at most logarithmic for heavy hitter discovery whereas that of approximate histogram is $\Theta(\sqrt{R})$. We also empirically demonstrate our findings.
Learning Regions of Interest for Bayesian Optimization with Adaptive Level-Set Estimation
Authors: Fengxue Zhang, Jialin Song, James Bowden, Alexander Ladd, Yisong Yue, Thomas A. Desautels, Yuxin Chen
Abstract
We study Bayesian optimization (BO) in high-dimensional and non-stationary scenarios. Existing algorithms for such scenarios typically require extensive hyperparameter tuning, which limits their practical effectiveness. We propose a framework, called BALLET, which adaptively filters for a high-confidence region of interest (ROI) as a superlevel-set of a nonparametric probabilistic model such as a Gaussian process (GP). Our approach is easy to tune, and is able to focus on local region of the optimization space that can be tackled by existing BO methods. The key idea is to use two probabilistic models: a coarse GP to identify the ROI, and a localized GP for optimization within the ROI. We show theoretically that BALLET can efficiently shrink the search space, and can exhibit a tighter regret bound than standard BO without ROI filtering. We demonstrate empirically the effectiveness of BALLET on both synthetic and real-world optimization tasks.
Scaff-PD: Communication Efficient Fair and Robust Federated Learning
Authors: Yaodong Yu, Sai Praneeth Karimireddy, Yi Ma, Michael I. Jordan
Subjects: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC); Optimization and Control (math.OC); Machine Learning (stat.ML)
Abstract
We present Scaff-PD, a fast and communication-efficient algorithm for distributionally robust federated learning. Our approach improves fairness by optimizing a family of distributionally robust objectives tailored to heterogeneous clients. We leverage the special structure of these objectives, and design an accelerated primal dual (APD) algorithm which uses bias corrected local steps (as in Scaffold) to achieve significant gains in communication efficiency and convergence speed. We evaluate Scaff-PD on several benchmark datasets and demonstrate its effectiveness in improving fairness and robustness while maintaining competitive accuracy. Our results suggest that Scaff-PD is a promising approach for federated learning in resource-constrained and heterogeneous settings.
Counterfactual Explanation via Search in Gaussian Mixture Distributed Latent Space
Authors: Xuan Zhao, Klaus Broelemann, Gjergji Kasneci
Abstract
Counterfactual Explanations (CEs) are an important tool in Algorithmic Recourse for addressing two questions: 1. What are the crucial factors that led to an automated prediction/decision? 2. How can these factors be changed to achieve a more favorable outcome from a user's perspective? Thus, guiding the user's interaction with AI systems by proposing easy-to-understand explanations and easy-to-attain feasible changes is essential for the trustworthy adoption and long-term acceptance of AI systems. In the literature, various methods have been proposed to generate CEs, and different quality measures have been suggested to evaluate these methods. However, the generation of CEs is usually computationally expensive, and the resulting suggestions are unrealistic and thus non-actionable. In this paper, we introduce a new method to generate CEs for a pre-trained binary classifier by first shaping the latent space of an autoencoder to be a mixture of Gaussian distributions. CEs are then generated in latent space by linear interpolation between the query sample and the centroid of the target class. We show that our method maintains the characteristics of the input sample during the counterfactual search. In various experiments, we show that the proposed method is competitive based on different quality measures on image and tabular datasets -- efficiently returns results that are closer to the original data manifold compared to three state-of-the-art methods, which are essential for realistic high-dimensional machine learning applications.
Solving Odd-Fair Parity Games
Authors: Irmak Sağlam, Anne-Kathrin Schmuck
Subjects: Computer Science and Game Theory (cs.GT); Systems and Control (eess.SY)
Abstract
This paper discusses the problem of efficiently solving parity games where player Odd has to obey an additional 'strong transition fairness constraint' on its vertices -- given that a player Odd vertex $v$ is visited infinitely often, a particular subset of the outgoing edges (called live edges) of $v$ has to be taken infinitely often. Such games, which we call 'Odd-fair parity games', naturally arise from abstractions of cyber-physical systems for planning and control. In this paper, we present a new Zielonka-type algorithm for solving Odd-fair parity games. This algorithm not only shares 'the same worst-case time complexity' as Zielonka's algorithm for (normal) parity games but also preserves the algorithmic advantage Zielonka's algorithm possesses over other parity solvers with exponential time complexity. We additionally introduce a formalization of Odd player winning strategies in such games, which were unexplored previous to this work. This formalization serves dual purposes: firstly, it enables us to prove our Zielonka-type algorithm; secondly, it stands as a noteworthy contribution in its own right, augmenting our understanding of additional fairness assumptions in two-player games.
Scoring Cycling Environments Perceived Safety using Pairwise Image Comparisons
Authors: Miguel Costa, Manuel Marques, Felix Wilhelm Siebert, Carlos Lima Azevedo, Filipe Moura
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Today, many cities seek to transition to more sustainable transportation systems. Cycling is critical in this transition for shorter trips, including first-and-last-mile links to transit. Yet, if individuals perceive cycling as unsafe, they will not cycle and choose other transportation modes. This study presents a novel approach to identifying how the perception of cycling safety can be analyzed and understood and the impact of the built environment and cycling contexts on such perceptions. We base our work on other perception studies and pairwise comparisons, using real-world images to survey respondents. We repeatedly show respondents two road environments and ask them to select the one they perceive as safer for cycling. We compare several methods capable of rating cycling environments from pairwise comparisons and classify cycling environments perceived as safe or unsafe. Urban planning can use this score to improve interventions' effectiveness and improve cycling promotion campaigns. Furthermore, this approach facilitates the continuous assessment of changing cycling environments, allows for a short-term evaluation of measures, and is efficiently deployed in different locations or contexts.
Mitigating Memory Wall Effects in CNN Engines with On-the-Fly Weights Generation
Authors: Stylianos I. Venieris, Javier Fernandez-Marques, Nicholas D. Lane
Abstract
The unprecedented accuracy of convolutional neural networks (CNNs) across a broad range of AI tasks has led to their widespread deployment in mobile and embedded settings. In a pursuit for high-performance and energy-efficient inference, significant research effort has been invested in the design of FPGA-based CNN accelerators. In this context, single computation engines constitute a popular approach to support diverse CNN modes without the overhead of fabric reconfiguration. Nevertheless, this flexibility often comes with significantly degraded performance on memory-bound layers and resource underutilisation due to the suboptimal mapping of certain layers on the engine's fixed configuration. In this work, we investigate the implications in terms of CNN engine design for a class of models that introduce a pre-convolution stage to decompress the weights at run time. We refer to these approaches as on-the-fly. This paper presents unzipFPGA, a novel CNN inference system that counteracts the limitations of existing CNN engines. The proposed framework comprises a novel CNN hardware architecture that introduces a weights generator module that enables the on-chip on-the-fly generation of weights, alleviating the negative impact of limited bandwidth on memory-bound layers. We further enhance unzipFPGA with an automated hardware-aware methodology that tailors the weights generation mechanism to the target CNN-device pair, leading to an improved accuracy-performance balance. Finally, we introduce an input selective processing element (PE) design that balances the load between PEs in suboptimally mapped layers. The proposed framework yields hardware designs that achieve an average of 2.57x performance efficiency gain over highly optimised GPU designs for the same power constraints and up to 3.94x higher performance density over a diverse range of state-of-the-art FPGA-based CNN accelerators.
Communication-Efficient Orchestrations for URLLC Service via Hierarchical Reinforcement Learning
Abstract
Ultra-reliable low latency communications (URLLC) service is envisioned to enable use cases with strict reliability and latency requirements in 5G. One approach for enabling URLLC services is to leverage Reinforcement Learning (RL) to efficiently allocate wireless resources. However, with conventional RL methods, the decision variables (though being deployed at various network layers) are typically optimized in the same control loop, leading to significant practical limitations on the control loop's delay as well as excessive signaling and energy consumption. In this paper, we propose a multi-agent Hierarchical RL (HRL) framework that enables the implementation of multi-level policies with different control loop timescales. Agents with faster control loops are deployed closer to the base station, while the ones with slower control loops are at the edge or closer to the core network providing high-level guidelines for low-level actions. On a use case from the prior art, with our HRL framework, we optimized the maximum number of retransmissions and transmission power of industrial devices. Our extensive simulation results on the factory automation scenario show that the HRL framework achieves better performance as the baseline single-agent RL method, with significantly less overhead of signal transmissions and delay compared to the one-agent RL methods.
A signal processing interpretation of noise-reduction convolutional neural networks
Authors: Luis A. Zavala-Mondragón, Peter H.N. de With, Fons van der Sommen
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Image and Video Processing (eess.IV); Signal Processing (eess.SP)
Abstract
Encoding-decoding CNNs play a central role in data-driven noise reduction and can be found within numerous deep-learning algorithms. However, the development of these CNN architectures is often done in ad-hoc fashion and theoretical underpinnings for important design choices is generally lacking. Up to this moment there are different existing relevant works that strive to explain the internal operation of these CNNs. Still, these ideas are either scattered and/or may require significant expertise to be accessible for a bigger audience. In order to open up this exciting field, this article builds intuition on the theory of deep convolutional framelets and explains diverse ED CNN architectures in a unified theoretical framework. By connecting basic principles from signal processing to the field of deep learning, this self-contained material offers significant guidance for designing robust and efficient novel CNN architectures.
Achieving Linear Speedup in Decentralized Stochastic Compositional Minimax Optimization
Authors: Hongchang Gao
Subjects: Machine Learning (cs.LG); Optimization and Control (math.OC)
Abstract
The stochastic compositional minimax problem has attracted a surge of attention in recent years since it covers many emerging machine learning models. Meanwhile, due to the emergence of distributed data, optimizing this kind of problem under the decentralized setting becomes badly needed. However, the compositional structure in the loss function brings unique challenges to designing efficient decentralized optimization algorithms. In particular, our study shows that the standard gossip communication strategy cannot achieve linear speedup for decentralized compositional minimax problems due to the large consensus error about the inner-level function. To address this issue, we developed a novel decentralized stochastic compositional gradient descent ascent with momentum algorithm to reduce the consensus error in the inner-level function. As such, our theoretical results demonstrate that it is able to achieve linear speedup with respect to the number of workers. We believe this novel algorithmic design could benefit the development of decentralized compositional optimization. Finally, we applied our methods to the imbalanced classification problem. The extensive experimental results provide evidence for the effectiveness of our algorithm.
Combinatorial Auctions and Graph Neural Networks for Local Energy Flexibility Markets
Authors: Awadelrahman M. A. Ahmed, Frank Eliassen, Yan Zhang
Subjects: Machine Learning (cs.LG); Computer Science and Game Theory (cs.GT)
Abstract
This paper proposes a new combinatorial auction framework for local energy flexibility markets, which addresses the issue of prosumers' inability to bundle multiple flexibility time intervals. To solve the underlying NP-complete winner determination problems, we present a simple yet powerful heterogeneous tri-partite graph representation and design graph neural network-based models. Our models achieve an average optimal value deviation of less than 5\% from an off-the-shelf optimization tool and show linear inference time complexity compared to the exponential complexity of the commercial solver. Contributions and results demonstrate the potential of using machine learning to efficiently allocate energy flexibility resources in local markets and solving optimization problems in general.
Exploring MLOps Dynamics: An Experimental Analysis in a Real-World Machine Learning Project
Abstract
This article presents an experiment focused on optimizing the MLOps (Machine Learning Operations) process, a crucial aspect of efficiently implementing machine learning projects. The objective is to identify patterns and insights to enhance the MLOps workflow, considering its iterative and interdependent nature in real-world model development scenarios. The experiment involves a comprehensive MLOps workflow, covering essential phases like problem definition, data acquisition, data preparation, model development, model deployment, monitoring, management, scalability, and governance and compliance. Practical tips and recommendations are derived from the results, emphasizing proactive planning and continuous improvement for the MLOps workflow. The experimental investigation was strategically integrated within a real-world ML project which followed essential phases of the MLOps process in a production environment, handling large-scale structured data. A systematic tracking approach was employed to document revisits to specific phases from a main phase under focus, capturing the reasons for such revisits. By constructing a matrix to quantify the degree of overlap between phases, the study unveils the dynamic and iterative nature of the MLOps workflow. The resulting data provides visual representations of the MLOps process's interdependencies and iterative characteristics within the experimental framework, offering valuable insights for optimizing the workflow and making informed decisions in real-world scenarios. This analysis contributes to enhancing the efficiency and effectiveness of machine learning projects through an improved MLOps process. Keywords: MLOps, Machine Learning Operations, Optimization, Experimental Analysis, Iterative Process, Pattern Identification.
Finding Money Launderers Using Heterogeneous Graph Neural Networks
Abstract
Current anti-money laundering (AML) systems, predominantly rule-based, exhibit notable shortcomings in efficiently and precisely detecting instances of money laundering. As a result, there has been a recent surge toward exploring alternative approaches, particularly those utilizing machine learning. Since criminals often collaborate in their money laundering endeavors, accounting for diverse types of customer relations and links becomes crucial. In line with this, the present paper introduces a graph neural network (GNN) approach to identify money laundering activities within a large heterogeneous network constructed from real-world bank transactions and business role data belonging to DNB, Norway's largest bank. Specifically, we extend the homogeneous GNN method known as the Message Passing Neural Network (MPNN) to operate effectively on a heterogeneous graph. As part of this procedure, we propose a novel method for aggregating messages across different edges of the graph. Our findings highlight the importance of using an appropriate GNN architecture when combining information in heterogeneous graphs. The performance results of our model demonstrate great potential in enhancing the quality of electronic surveillance systems employed by banks to detect instances of money laundering. To the best of our knowledge, this is the first published work applying GNN on a large real-world heterogeneous network for anti-money laundering purposes.
Preliminary Design of the Dragonfly Navigation Filter
Authors: Ben Schilling, Timothy G. McGee, Ryan Mitch, Ryan Watson
Abstract
Dragonfly is scheduled to begin exploring Titan by 2034 using a series of multi-kilometer surface flights. This paper outlines the preliminary design of the navigation filter for the Dragonfly Mobility subsystem. The software architecture and filter formulation for lidar, visual odometry, pressure sensors, and redundant IMUs are described in detail. Special discussion is given to developments to achieve multi-kilometer surface flights, including optimizing sequential image baselines, modeling correlating image processing errors, and an efficient approximation to the Simultaneous Localization and Mapping (SLAM) problem.
INFINITY: Neural Field Modeling for Reynolds-Averaged Navier-Stokes Equations
Authors: Louis Serrano, Leon Migus, Yuan Yin, Jocelyn Ahmed Mazari, Patrick Gallinari
Abstract
For numerical design, the development of efficient and accurate surrogate models is paramount. They allow us to approximate complex physical phenomena, thereby reducing the computational burden of direct numerical simulations. We propose INFINITY, a deep learning model that utilizes implicit neural representations (INRs) to address this challenge. Our framework encodes geometric information and physical fields into compact representations and learns a mapping between them to infer the physical fields. We use an airfoil design optimization problem as an example task and we evaluate our approach on the challenging AirfRANS dataset, which closely resembles real-world industrial use-cases. The experimental results demonstrate that our framework achieves state-of-the-art performance by accurately inferring physical fields throughout the volume and surface. Additionally we demonstrate its applicability in contexts such as design exploration and shape optimization: our model can correctly predict drag and lift coefficients while adhering to the equations.
Group Activity Recognition in Computer Vision: A Comprehensive Review, Challenges, and Future Perspectives
Authors: Chuanchuan Wang, Ahmad Sufril Azlan Mohamed
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Group activity recognition is a hot topic in computer vision. Recognizing activities through group relationships plays a vital role in group activity recognition. It holds practical implications in various scenarios, such as video analysis, surveillance, automatic driving, and understanding social activities. The model's key capabilities encompass efficiently modeling hierarchical relationships within a scene and accurately extracting distinctive spatiotemporal features from groups. Given this technology's extensive applicability, identifying group activities has garnered significant research attention. This work examines the current progress in technology for recognizing group activities, with a specific focus on global interactivity and activities. Firstly, we comprehensively review the pertinent literature and various group activity recognition approaches, from traditional methodologies to the latest methods based on spatial structure, descriptors, non-deep learning, hierarchical recurrent neural networks (HRNN), relationship models, and attention mechanisms. Subsequently, we present the relational network and relational architectures for each module. Thirdly, we investigate methods for recognizing group activity and compare their performance with state-of-the-art technologies. We summarize the existing challenges and provide comprehensive guidance for newcomers to understand group activity recognition. Furthermore, we review emerging perspectives in group activity recognition to explore new directions and possibilities.
On Solving the Rubik's Cube with Domain-Independent Planners Using Standard Representations
Abstract
Rubik's Cube (RC) is a well-known and computationally challenging puzzle that has motivated AI researchers to explore efficient alternative representations and problem-solving methods. The ideal situation for planning here is that a problem be solved optimally and efficiently represented in a standard notation using a general-purpose solver and heuristics. The fastest solver today for RC is DeepCubeA with a custom representation, and another approach is with Scorpion planner with State-Action-Space+ (SAS+) representation. In this paper, we present the first RC representation in the popular PDDL language so that the domain becomes more accessible to PDDL planners, competitions, and knowledge engineering tools, and is more human-readable. We then bridge across existing approaches and compare performance. We find that in one comparable experiment, DeepCubeA solves all problems with varying complexities, albeit only 18\% are optimal plans. For the same problem set, Scorpion with SAS+ representation and pattern database heuristics solves 61.50\% problems, while FastDownward with PDDL representation and FF heuristic solves 56.50\% problems, out of which all the plans generated were optimal. Our study provides valuable insights into the trade-offs between representational choice and plan optimality that can help researchers design future strategies for challenging domains combining general-purpose solving methods (planning, reinforcement learning), heuristics, and representations (standard or custom).
A threshold dislocation dynamics method
Authors: Xiaoxue Qin, Alfonso H.W. Ngan, Yang Xiang
Abstract
The Merriman-Bence-Osher threshold dynamics method is an efficient algorithm to simulate the motion by mean curvature, in which two steps of convolution with diffusion kernel and thresholding alternate. It has the advantages of being easy to implement and with high efficiency. In this paper, we propose an efficient threshold dynamics method for dislocation dynamics in a slip plane. We show that this proposed threshold dislocation dynamics method is able to give correct two leading orders in dislocation velocity, including both the $O(\log \varepsilon)$ local curvature force and the $O(1)$ nonlocal force due to the long-range stress field generated by the dislocations, where $\varepsilon$ is the dislocation core size. This is different from the available threshold dynamics methods in the literature which only give the leading order local velocities associated with mean curvature or its anisotropic generalizations of the moving fronts. We also propose a numerical method based on spatial variable stretching to overcome the numerical limitations brought by physical settings in this threshold dislocation dynamics method. Specifically, this variable stretching method is able to correct the mobility and to rescale the velocity, which can be applied generally to any threshold dynamics method. We validate the proposed threshold dislocation dynamics method by numerical simulations of various motions and interaction of dislocations.
Smartpick: Workload Prediction for Serverless-enabled Scalable Data Analytics Systems
Authors: Anshuman Das Mohapatra, Kwangsung Oh
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
Many data analytic systems have adopted a newly emerging compute resource, serverless (SL), to handle data analytics queries in a timely and cost-efficient manner, i.e., serverless data analytics. While these systems can start processing queries quickly thanks to the agility and scalability of SL, they may encounter performance- and cost-bottlenecks based on workloads due to SL's worse performance and more expensive cost than traditional compute resources, e.g., virtual machine (VM). In this project, we introduce Smartpick, a SL-enabled scalable data analytics system that exploits SL and VM together to realize composite benefits, i.e., agility from SL and better performance with reduced cost from VM. Smartpick uses a machine learning prediction scheme, decision-tree based Random Forest with Bayesian Optimizer, to determine SL and VM configurations, i.e., how many SL and VM instances for queries, that meet cost-performance goals. Smartpick offers a knob for applications to allow them to explore a richer cost-performance tradeoff space opened by exploiting SL and VM together. To maximize the benefits of SL, Smartpick supports a simple but strong mechanism, called relay-instances. Smartpick also supports event-driven prediction model retraining to deal with workload dynamics. A Smartpick prototype was implemented on Spark and deployed on live test-beds, Amazon AWS and Google Cloud Platform. Evaluation results indicate 97.05% and 83.49% prediction accuracies respectively with up to 50% cost reduction as opposed to the baselines. The results also confirm that Smartpick allows data analytics applications to navigate the richer cost-performance tradeoff space efficiently and to handle workload dynamics effectively and automatically.
RED CoMETS: An ensemble classifier for symbolically represented multivariate time series
Abstract
Multivariate time series classification is a rapidly growing research field with practical applications in finance, healthcare, engineering, and more. The complexity of classifying multivariate time series data arises from its high dimensionality, temporal dependencies, and varying lengths. This paper introduces a novel ensemble classifier called RED CoMETS (Random Enhanced Co-eye for Multivariate Time Series), which addresses these challenges. RED CoMETS builds upon the success of Co-eye, an ensemble classifier specifically designed for symbolically represented univariate time series, and extends its capabilities to handle multivariate data. The performance of RED CoMETS is evaluated on benchmark datasets from the UCR archive, where it demonstrates competitive accuracy when compared to state-of-the-art techniques in multivariate settings. Notably, it achieves the highest reported accuracy in the literature for the 'HandMovementDirection' dataset. Moreover, the proposed method significantly reduces computation time compared to Co-eye, making it an efficient and effective choice for multivariate time series classification.
A Compact DAG for Storing and Searching Maximal Common Subsequences
Authors: Alessio Conte, Roberto Grossi, Giulia Punzi, Takeaki Uno
Abstract
Maximal Common Subsequences (MCSs) between two strings X and Y are subsequences of both X and Y that are maximal under inclusion. MCSs relax and generalize the well known and widely used concept of Longest Common Subsequences (LCSs), which can be seen as MCSs of maximum length. While the number both LCSs and MCSs can be exponential in the length of the strings, LCSs have been long exploited for string and text analysis, as simple compact representations of all LCSs between two strings, built via dynamic programming or automata, have been known since the '70s. MCSs appear to have a more challenging structure: even listing them efficiently was an open problem open until recently, thus narrowing the complexity difference between the two problems, but the gap remained significant. In this paper we close the complexity gap: we show how to build DAG of polynomial size-in polynomial time-which allows for efficient operations on the set of all MCSs such as enumeration in Constant Amortized Time per solution (CAT), counting, and random access to the i-th element (i.e., rank and select operations). Other than improving known algorithmic results, this work paves the way for new sequence analysis methods based on MCSs.
Keyword: faster
Communication-Efficient Orchestrations for URLLC Service via Hierarchical Reinforcement Learning
Abstract
Ultra-reliable low latency communications (URLLC) service is envisioned to enable use cases with strict reliability and latency requirements in 5G. One approach for enabling URLLC services is to leverage Reinforcement Learning (RL) to efficiently allocate wireless resources. However, with conventional RL methods, the decision variables (though being deployed at various network layers) are typically optimized in the same control loop, leading to significant practical limitations on the control loop's delay as well as excessive signaling and energy consumption. In this paper, we propose a multi-agent Hierarchical RL (HRL) framework that enables the implementation of multi-level policies with different control loop timescales. Agents with faster control loops are deployed closer to the base station, while the ones with slower control loops are at the edge or closer to the core network providing high-level guidelines for low-level actions. On a use case from the prior art, with our HRL framework, we optimized the maximum number of retransmissions and transmission power of industrial devices. Our extensive simulation results on the factory automation scenario show that the HRL framework achieves better performance as the baseline single-agent RL method, with significantly less overhead of signal transmissions and delay compared to the one-agent RL methods.
A behavioural transformer for effective collaboration between a robot and a non-stationary human
Abstract
A key challenge in human-robot collaboration is the non-stationarity created by humans due to changes in their behaviour. This alters environmental transitions and hinders human-robot collaboration. We propose a principled meta-learning framework to explore how robots could better predict human behaviour, and thereby deal with issues of non-stationarity. On the basis of this framework, we developed Behaviour-Transform (BeTrans). BeTrans is a conditional transformer that enables a robot agent to adapt quickly to new human agents with non-stationary behaviours, due to its notable performance with sequential data. We trained BeTrans on simulated human agents with different systematic biases in collaborative settings. We used an original customisable environment to show that BeTrans effectively collaborates with simulated human agents and adapts faster to non-stationary simulated human agents than SOTA techniques.
Monte-Carlo Tree Search for Multi-Agent Pathfinding: Preliminary Results
Authors: Yelisey Pitanov, Alexey Skrynnik, Anton Andreychuk, Konstantin Yakovlev, Aleksandr Panov
Abstract
In this work we study a well-known and challenging problem of Multi-agent Pathfinding, when a set of agents is confined to a graph, each agent is assigned a unique start and goal vertices and the task is to find a set of collision-free paths (one for each agent) such that each agent reaches its respective goal. We investigate how to utilize Monte-Carlo Tree Search (MCTS) to solve the problem. Although MCTS was shown to demonstrate superior performance in a wide range of problems like playing antagonistic games (e.g. Go, Chess etc.), discovering faster matrix multiplication algorithms etc., its application to the problem at hand was not well studied before. To this end we introduce an original variant of MCTS, tailored to multi-agent pathfinding. The crux of our approach is how the reward, that guides MCTS, is computed. Specifically, we use individual paths to assist the agents with the the goal-reaching behavior, while leaving them freedom to get off the track if it is needed to avoid collisions. We also use a dedicated decomposition technique to reduce the branching factor of the tree search procedure. Empirically we show that the suggested method outperforms the baseline planning algorithm that invokes heuristic search, e.g. A*, at each re-planning step.
Spectrum-guided Multi-granularity Referring Video Object Segmentation
Authors: Bo Miao, Mohammed Bennamoun, Yongsheng Gao, Ajmal Mian
Abstract
Current referring video object segmentation (R-VOS) techniques extract conditional kernels from encoded (low-resolution) vision-language features to segment the decoded high-resolution features. We discovered that this causes significant feature drift, which the segmentation kernels struggle to perceive during the forward computation. This negatively affects the ability of segmentation kernels. To address the drift problem, we propose a Spectrum-guided Multi-granularity (SgMg) approach, which performs direct segmentation on the encoded features and employs visual details to further optimize the masks. In addition, we propose Spectrum-guided Cross-modal Fusion (SCF) to perform intra-frame global interactions in the spectral domain for effective multimodal representation. Finally, we extend SgMg to perform multi-object R-VOS, a new paradigm that enables simultaneous segmentation of multiple referred objects in a video. This not only makes R-VOS faster, but also more practical. Extensive experiments show that SgMg achieves state-of-the-art performance on four video benchmark datasets, outperforming the nearest competitor by 2.8% points on Ref-YouTube-VOS. Our extended SgMg enables multi-object R-VOS, runs about 3 times faster while maintaining satisfactory performance. Code is available at https://github.com/bo-miao/SgMg.
Keyword: mobile
Our Nudges, Our Selves: Tailoring Mobile User Engagement Using Personality
Abstract
To increase mobile user engagement, current apps employ a variety of behavioral nudges, but these engagement techniques are applied in a one-size-fits-all approach. Yet the very same techniques may be perceived differently by different individuals. To test this, we developed HarrySpotter, a location-based AR app that embedded six engagement techniques. We deployed it in a 2-week study involving 29 users who also took the Big-Five personality test. Preferences for specific engagement techniques are not only descriptive but also predictive of personality traits. The Adj. $R^2$ ranges from 0.16 for conscientious users (encouraged by competition) to 0.32 for neurotic users (self-centered and focused on their own achievements), and even up to 0.61 for extroverts (motivated by both exploration of objects and places). These findings suggest that these techniques need to be personalized in the future.
Abstract
We investigate a combinatorial optimization problem that involves patrolling the edges of an acute triangle using a unit-speed agent. The goal is to minimize the maximum (1-gap) idle time of any edge, which is defined as the time gap between consecutive visits to that edge. This problem has roots in a centuries-old optimization problem posed by Fagnano in 1775, who sought to determine the inscribed triangle of an acute triangle with the minimum perimeter. It is well-known that the orthic triangle, giving rise to a periodic and cyclic trajectory obeying the laws of geometric optics, is the optimal solution to Fagnano's problem. Such trajectories are known as Fagnano orbits, or more generally as billiard trajectories. We demonstrate that the orthic triangle is also an optimal solution to the patrolling problem. Our main contributions pertain to new connections between billiard trajectories and optimal patrolling schedules in combinatorial optimization. In particular, as an artifact of our arguments, we introduce a novel 2-gap patrolling problem that seeks to minimize the visitation time of objects every three visits. We prove that there exist infinitely many well-structured billiard-type optimal trajectories for this problem, including the orthic trajectory, which has the special property of minimizing the visitation time gap between any two consecutively visited edges. Complementary to that, we also examine the cost of dynamic, sub-optimal trajectories to the 1-gap patrolling optimization problem. These trajectories result from a greedy algorithm and can be implemented by a computationally primitive mobile agent.
Abstract
Autonomous robots have real-world applications in diverse fields, such as mobile manipulation and environmental exploration, and many such tasks benefit from a hands-off approach in terms of human user involvement over a long task horizon. However, the level of autonomy achievable by a deployment is limited in part by the problem definition or task specification required by the system. Task specifications often require technical, low-level information that is unintuitive to describe and may result in generic solutions, burdening the user technically both before and after task completion. In this thesis, we aim to advance task specification abstraction toward the goal of increasing robot autonomy in real-world scenarios. We do so by tackling problems that address several different angles of this goal. First, we develop a way for the automatic discovery of optimal transition points between subtasks in the context of constrained mobile manipulation, removing the need for the human to hand-specify these in the task specification. We further propose a way to automatically describe constraints on robot motion by using demonstrated data as opposed to manually-defined constraints. Then, within the context of environmental exploration, we propose a flexible task specification framework, requiring just a set of quantiles of interest from the user that allows the robot to directly suggest locations in the environment for the user to study. We next systematically study the effect of including a robot team in the task specification and show that multirobot teams have the ability to improve performance under certain specification conditions, including enabling inter-robot communication. Finally, we propose methods for a communication protocol that autonomously selects useful but limited information to share with the other robots.
Mitigating Memory Wall Effects in CNN Engines with On-the-Fly Weights Generation
Authors: Stylianos I. Venieris, Javier Fernandez-Marques, Nicholas D. Lane
Abstract
The unprecedented accuracy of convolutional neural networks (CNNs) across a broad range of AI tasks has led to their widespread deployment in mobile and embedded settings. In a pursuit for high-performance and energy-efficient inference, significant research effort has been invested in the design of FPGA-based CNN accelerators. In this context, single computation engines constitute a popular approach to support diverse CNN modes without the overhead of fabric reconfiguration. Nevertheless, this flexibility often comes with significantly degraded performance on memory-bound layers and resource underutilisation due to the suboptimal mapping of certain layers on the engine's fixed configuration. In this work, we investigate the implications in terms of CNN engine design for a class of models that introduce a pre-convolution stage to decompress the weights at run time. We refer to these approaches as on-the-fly. This paper presents unzipFPGA, a novel CNN inference system that counteracts the limitations of existing CNN engines. The proposed framework comprises a novel CNN hardware architecture that introduces a weights generator module that enables the on-chip on-the-fly generation of weights, alleviating the negative impact of limited bandwidth on memory-bound layers. We further enhance unzipFPGA with an automated hardware-aware methodology that tailors the weights generation mechanism to the target CNN-device pair, leading to an improved accuracy-performance balance. Finally, we introduce an input selective processing element (PE) design that balances the load between PEs in suboptimally mapped layers. The proposed framework yields hardware designs that achieve an average of 2.57x performance efficiency gain over highly optimised GPU designs for the same power constraints and up to 3.94x higher performance density over a diverse range of state-of-the-art FPGA-based CNN accelerators.
An Explainable Model-Agnostic Algorithm for CNN-based Biometrics Verification
Authors: Fernando Alonso-Fernandez, Kevin Hernandez-Diaz, Jose M. Buades, Prayag Tiwari, Josef Bigun
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
This paper describes an adaptation of the Local Interpretable Model-Agnostic Explanations (LIME) AI method to operate under a biometric verification setting. LIME was initially proposed for networks with the same output classes used for training, and it employs the softmax probability to determine which regions of the image contribute the most to classification. However, in a verification setting, the classes to be recognized have not been seen during training. In addition, instead of using the softmax output, face descriptors are usually obtained from a layer before the classification layer. The model is adapted to achieve explainability via cosine similarity between feature vectors of perturbated versions of the input image. The method is showcased for face biometrics with two CNN models based on MobileNetv2 and ResNet50.
Abstract
The utilization of Large Language Models (LLMs) for the construction of AI systems has garnered significant attention across diverse fields. The extension of LLMs to the domain of fashion holds substantial commercial potential but also inherent challenges due to the intricate semantic interactions in fashion-related generation. To address this issue, we developed a hierarchical AI system called Fashion Matrix dedicated to editing photos by just talking. This system facilitates diverse prompt-driven tasks, encompassing garment or accessory replacement, recoloring, addition, and removal. Specifically, Fashion Matrix employs LLM as its foundational support and engages in iterative interactions with users. It employs a range of Semantic Segmentation Models (e.g., Grounded-SAM, MattingAnything, etc.) to delineate the specific editing masks based on user instructions. Subsequently, Visual Foundation Models (e.g., Stable Diffusion, ControlNet, etc.) are leveraged to generate edited images from text prompts and masks, thereby facilitating the automation of fashion editing processes. Experiments demonstrate the outstanding ability of Fashion Matrix to explores the collaborative potential of functionally diverse pre-trained models in the domain of fashion editing.
Not with my name! Inferring artists' names of input strings employed by Diffusion Models
Authors: Roberto Leotta, Oliver Giudice, Luca Guarnera, Sebastiano Battiato
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Diffusion Models (DM) are highly effective at generating realistic, high-quality images. However, these models lack creativity and merely compose outputs based on their training data, guided by a textual input provided at creation time. Is it acceptable to generate images reminiscent of an artist, employing his name as input? This imply that if the DM is able to replicate an artist's work then it was trained on some or all of his artworks thus violating copyright. In this paper, a preliminary study to infer the probability of use of an artist's name in the input string of a generated image is presented. To this aim we focused only on images generated by the famous DALL-E 2 and collected images (both original and generated) of five renowned artists. Finally, a dedicated Siamese Neural Network was employed to have a first kind of probability. Experimental results demonstrate that our approach is an optimal starting point and can be employed as a prior for predicting a complete input string of an investigated image. Dataset and code are available at: https://github.com/ictlab-unict/not-with-my-name .
XDLM: Cross-lingual Diffusion Language Model for Machine Translation
Authors: Linyao Chen, Aosong Feng, Boming Yang, Zihui Li
Abstract
Recently, diffusion models have excelled in image generation tasks and have also been applied to neural language processing (NLP) for controllable text generation. However, the application of diffusion models in a cross-lingual setting is less unexplored. Additionally, while pretraining with diffusion models has been studied within a single language, the potential of cross-lingual pretraining remains understudied. To address these gaps, we propose XDLM, a novel Cross-lingual diffusion model for machine translation, consisting of pretraining and fine-tuning stages. In the pretraining stage, we propose TLDM, a new training objective for mastering the mapping between different languages; in the fine-tuning stage, we build up the translation system based on the pretrained model. We evaluate the result on several machine translation benchmarks and outperformed both diffusion and Transformer baselines.
Fake It Without Making It: Conditioned Face Generation for Accurate 3D Face Shape Estimation
Authors: Will Rowan, Patrik Huber, Nick Pears, Andrew Keeling
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Accurate 3D face shape estimation is an enabling technology with applications in healthcare, security, and creative industries, yet current state-of-the-art methods either rely on self-supervised training with 2D image data or supervised training with very limited 3D data. To bridge this gap, we present a novel approach which uses a conditioned stable diffusion model for face image generation, leveraging the abundance of 2D facial information to inform 3D space. By conditioning stable diffusion on depth maps sampled from a 3D Morphable Model (3DMM) of the human face, we generate diverse and shape-consistent images, forming the basis of SynthFace. We introduce this large-scale synthesised dataset of 250K photorealistic images and corresponding 3DMM parameters. We further propose ControlFace, a deep neural network, trained on SynthFace, which achieves competitive performance on the NoW benchmark, without requiring 3D supervision or manual 3D asset creation.
A threshold dislocation dynamics method
Authors: Xiaoxue Qin, Alfonso H.W. Ngan, Yang Xiang
Abstract
The Merriman-Bence-Osher threshold dynamics method is an efficient algorithm to simulate the motion by mean curvature, in which two steps of convolution with diffusion kernel and thresholding alternate. It has the advantages of being easy to implement and with high efficiency. In this paper, we propose an efficient threshold dynamics method for dislocation dynamics in a slip plane. We show that this proposed threshold dislocation dynamics method is able to give correct two leading orders in dislocation velocity, including both the $O(\log \varepsilon)$ local curvature force and the $O(1)$ nonlocal force due to the long-range stress field generated by the dislocations, where $\varepsilon$ is the dislocation core size. This is different from the available threshold dynamics methods in the literature which only give the leading order local velocities associated with mean curvature or its anisotropic generalizations of the moving fronts. We also propose a numerical method based on spatial variable stretching to overcome the numerical limitations brought by physical settings in this threshold dislocation dynamics method. Specifically, this variable stretching method is able to correct the mobility and to rescale the velocity, which can be applied generally to any threshold dynamics method. We validate the proposed threshold dislocation dynamics method by numerical simulations of various motions and interaction of dislocations.
Keyword: adaptive
Adaptive Certified Training: Towards Better Accuracy-Robustness Tradeoffs
Authors: Zhakshylyk Nurlanov, Frank R. Schmidt, Florian Bernard
Abstract
As deep learning models continue to advance and are increasingly utilized in real-world systems, the issue of robustness remains a major challenge. Existing certified training methods produce models that achieve high provable robustness guarantees at certain perturbation levels. However, the main problem of such models is a dramatically low standard accuracy, i.e. accuracy on clean unperturbed data, that makes them impractical. In this work, we consider a more realistic perspective of maximizing the robustness of a model at certain levels of (high) standard accuracy. To this end, we propose a novel certified training method based on a key insight that training with adaptive certified radii helps to improve both the accuracy and robustness of the model, advancing state-of-the-art accuracy-robustness tradeoffs. We demonstrate the effectiveness of the proposed method on MNIST, CIFAR-10, and TinyImageNet datasets. Particularly, on CIFAR-10 and TinyImageNet, our method yields models with up to two times higher robustness, measured as an average certified radius of a test set, at the same levels of standard accuracy compared to baseline approaches.
Text-oriented Modality Reinforcement Network for Multimodal Sentiment Analysis from Unaligned Multimodal Sequences
Abstract
Multimodal Sentiment Analysis (MSA) aims to mine sentiment information from text, visual, and acoustic modalities. Previous works have focused on representation learning and feature fusion strategies. However, most of these efforts ignored the disparity in the semantic richness of different modalities and treated each modality in the same manner. That may lead to strong modalities being neglected and weak modalities being overvalued. Motivated by these observations, we propose a Text-oriented Modality Reinforcement Network (TMRN), which focuses on the dominance of the text modality in MSA. More specifically, we design a Text-Centered Cross-modal Attention (TCCA) module to make full interaction for text/acoustic and text/visual pairs, and a Text-Gated Self-Attention (TGSA) module to guide the self-reinforcement of the other two modalities. Furthermore, we present an adaptive fusion mechanism to decide the proportion of different modalities involved in the fusion process. Finally, we combine the feature matrices into vectors to get the final representation for the downstream tasks. Experimental results show that our TMRN outperforms the state-of-the-art methods on two MSA benchmarks.
Multi-Granularity Prediction with Learnable Fusion for Scene Text Recognition
Authors: Cheng Da, Peng Wang, Cong Yao
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Due to the enormous technical challenges and wide range of applications, scene text recognition (STR) has been an active research topic in computer vision for years. To tackle this tough problem, numerous innovative methods have been successively proposed, and incorporating linguistic knowledge into STR models has recently become a prominent trend. In this work, we first draw inspiration from the recent progress in Vision Transformer (ViT) to construct a conceptually simple yet functionally powerful vision STR model, which is built upon ViT and a tailored Adaptive Addressing and Aggregation (A$^3$) module. It already outperforms most previous state-of-the-art models for scene text recognition, including both pure vision models and language-augmented methods. To integrate linguistic knowledge, we further propose a Multi-Granularity Prediction strategy to inject information from the language modality into the model in an implicit way, \ie, subword representations (BPE and WordPiece) widely used in NLP are introduced into the output space, in addition to the conventional character level representation, while no independent language model (LM) is adopted. To produce the final recognition results, two strategies for effectively fusing the multi-granularity predictions are devised. The resultant algorithm (termed MGP-STR) is able to push the performance envelope of STR to an even higher level. Specifically, MGP-STR achieves an average recognition accuracy of $94\%$ on standard benchmarks for scene text recognition. Moreover, it also achieves state-of-the-art results on widely-used handwritten benchmarks as well as more challenging scene text datasets, demonstrating the generality of the proposed MGP-STR algorithm. The source code and models will be available at: \url{https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/OCR/MGP-STR}.
GaitFormer: Revisiting Intrinsic Periodicity for Gait Recognition
Abstract
Gait recognition aims to distinguish different walking patterns by analyzing video-level human silhouettes, rather than relying on appearance information. Previous research on gait recognition has primarily focused on extracting local or global spatial-temporal representations, while overlooking the intrinsic periodic features of gait sequences, which, when fully utilized, can significantly enhance performance. In this work, we propose a plug-and-play strategy, called Temporal Periodic Alignment (TPA), which leverages the periodic nature and fine-grained temporal dependencies of gait patterns. The TPA strategy comprises two key components. The first component is Adaptive Fourier-transform Position Encoding (AFPE), which adaptively converts features and discrete-time signals into embeddings that are sensitive to periodic walking patterns. The second component is the Temporal Aggregation Module (TAM), which separates embeddings into trend and seasonal components, and extracts meaningful temporal correlations to identify primary components, while filtering out random noise. We present a simple and effective baseline method for gait recognition, based on the TPA strategy. Extensive experiments conducted on three popular public datasets (CASIA-B, OU-MVLP, and GREW) demonstrate that our proposed method achieves state-of-the-art performance on multiple benchmark tests.
Online Maximum Independent Set of Hyperrectangles
Authors: Rishi Advani, Abolfazl Asudeh
Subjects: Data Structures and Algorithms (cs.DS); Computational Geometry (cs.CG)
Abstract
The maximum independent set problem is a classical NP-hard problem in theoretical computer science. In this work, we study a special case where the family of graphs considered is restricted to intersection graphs of sets of axis-aligned hyperrectangles and the input is provided in an online fashion. We prove bounds on the competitive ratio of an optimal online algorithm under the adaptive offline, adaptive online, and oblivious adversary models, for several classes of hyperrectangles and restrictions on the order of the input. We are the first to present results on this problem under the oblivious adversary model. We prove bounds on the competitive ratio for unit hypercubes, $\sigma$-bounded hypercubes, unit-volume hypercubes, arbitrary hypercubes, and arbitrary hyperrectangles, in both arbitrary and non-dominated order. We are also the first to present results under the adaptive offline and adaptive online adversary models with input in non-dominated order, proving bounds on the competitive ratio for the same classes of hyperrectangles; for input in arbitrary order, we present the first results on $\sigma$-bounded hypercubes, unit-volume hyperrectangles, arbitrary hypercubes, and arbitrary hyperrectangles. For input in dominating order, we show that the performance of the naive greedy algorithm matches the performance of an optimal offline algorithm in all cases. We also give lower bounds on the competitive ratio of a probabilistic greedy algorithm under the oblivious adversary model. We conclude by discussing several promising directions for future work.
QuIP: 2-Bit Quantization of Large Language Models With Guarantees
Authors: Jerry Chee, Yaohui Cai, Volodymyr Kuleshov, Christopher De Sa
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
This work studies post-training parameter quantization in large language models (LLMs). We introduce quantization with incoherence processing (QuIP), a new method based on the insight that quantization benefits from incoherent weight and Hessian matrices, i.e., from the weights and the directions in which it is important to round them accurately being unaligned with the coordinate axes. QuIP consists of two steps: (1) an adaptive rounding procedure minimizing a quadratic proxy objective; (2) efficient pre- and post-processing that ensures weight and Hessian incoherence via multiplication by random orthogonal matrices. We complement QuIP with the first theoretical analysis for an LLM-scale quantization algorithm, and show that our theory also applies to an existing method, OPTQ. Empirically, we find that our incoherence preprocessing improves several existing quantization algorithms and yields the first LLM quantization methods that produce viable results using only two bits per weight. Our code can be found at https://github.com/jerry-chee/QuIP .
CT-Net: Arbitrary-Shaped Text Detection via Contour Transformer
Authors: Zhiwen Shao, Yuchen Su, Yong Zhou, Fanrong Meng, Hancheng Zhu, Bing Liu, Rui Yao
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Contour based scene text detection methods have rapidly developed recently, but still suffer from inaccurate frontend contour initialization, multi-stage error accumulation, or deficient local information aggregation. To tackle these limitations, we propose a novel arbitrary-shaped scene text detection framework named CT-Net by progressive contour regression with contour transformers. Specifically, we first employ a contour initialization module that generates coarse text contours without any post-processing. Then, we adopt contour refinement modules to adaptively refine text contours in an iterative manner, which are beneficial for context information capturing and progressive global contour deformation. Besides, we propose an adaptive training strategy to enable the contour transformers to learn more potential deformation paths, and introduce a re-score mechanism that can effectively suppress false positives. Extensive experiments are conducted on four challenging datasets, which demonstrate the accuracy and efficiency of our CT-Net over state-of-the-art methods. Particularly, CT-Net achieves F-measure of 86.1 at 11.2 frames per second (FPS) and F-measure of 87.8 at 10.1 FPS for CTW1500 and Total-Text datasets, respectively.
Learning Autonomous Ultrasound via Latent Task Representation and Robotic Skills Adaptation
Authors: Xutian Deng, Junnan Jiang, Wen Cheng, Miao Li
Abstract
As medical ultrasound is becoming a prevailing examination approach nowadays, robotic ultrasound systems can facilitate the scanning process and prevent professional sonographers from repetitive and tedious work. Despite the recent progress, it is still a challenge to enable robots to autonomously accomplish the ultrasound examination, which is largely due to the lack of a proper task representation method, and also an adaptation approach to generalize learned skills across different patients. To solve these problems, we propose the latent task representation and the robotic skills adaptation for autonomous ultrasound in this paper. During the offline stage, the multimodal ultrasound skills are merged and encapsulated into a low-dimensional probability model through a fully self-supervised framework, which takes clinically demonstrated ultrasound images, probe orientations, and contact forces into account. During the online stage, the probability model will select and evaluate the optimal prediction. For unstable singularities, the adaptive optimizer fine-tunes them to near and stable predictions in high-confidence regions. Experimental results show that the proposed approach can generate complex ultrasound strategies for diverse populations and achieve significantly better quantitative results than our previous method.
New Adaptive Low-Dissipation Central-Upwind Schemes
Abstract
We introduce new second-order adaptive low-dissipation central-upwind (LDCU) schemes for the one- and two-dimensional hyperbolic systems of conservation laws. The new adaptive LDCU schemes employ the LDCU numerical fluxes (recently proposed in [{\sc A. Kurganov and R. Xin}, J. Sci. Comput., 96 (2023), Paper No. 56]) computed using the point values reconstructed with the help of adaptively selected nonlinear limiters. To this end, we use a smoothness indicator to detect rough'' parts of the computed solution, where the piecewise linear reconstruction is performed using an overcompressive limiter, which leads to extremely sharp resolution of shock and contact waves. In thesmooth'' areas, we use a more dissipative limiter to prevent the appearance of artificial kinks and staircase-like structures there. In order to avoid oscillations, we perform the reconstruction in the local characteristic variables obtained using the local characteristic decomposition. We test two different smoothness indicators and apply the developed schemes to the one- and two-dimensional Euler equations of gas dynamics. The obtained numerical results clearly demonstrate that the new adaptive LDCU schemes outperform the original ones.
Kefa: A Knowledge Enhanced and Fine-grained Aligned Speaker for Navigation Instruction Generation
Authors: Haitian Zeng, Xiaohan Wang, Wenguan Wang, Yi Yang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
We introduce a novel speaker model \textsc{Kefa} for navigation instruction generation. The existing speaker models in Vision-and-Language Navigation suffer from the large domain gap of vision features between different environments and insufficient temporal grounding capability. To address the challenges, we propose a Knowledge Refinement Module to enhance the feature representation with external knowledge facts, and an Adaptive Temporal Alignment method to enforce fine-grained alignment between the generated instructions and the observation sequences. Moreover, we propose a new metric SPICE-D for navigation instruction evaluation, which is aware of the correctness of direction phrases. The experimental results on R2R and UrbanWalk datasets show that the proposed KEFA speaker achieves state-of-the-art instruction generation performance for both indoor and outdoor scenes.
Learning Regions of Interest for Bayesian Optimization with Adaptive Level-Set Estimation
Authors: Fengxue Zhang, Jialin Song, James Bowden, Alexander Ladd, Yisong Yue, Thomas A. Desautels, Yuxin Chen
Abstract
We study Bayesian optimization (BO) in high-dimensional and non-stationary scenarios. Existing algorithms for such scenarios typically require extensive hyperparameter tuning, which limits their practical effectiveness. We propose a framework, called BALLET, which adaptively filters for a high-confidence region of interest (ROI) as a superlevel-set of a nonparametric probabilistic model such as a Gaussian process (GP). Our approach is easy to tune, and is able to focus on local region of the optimization space that can be tackled by existing BO methods. The key idea is to use two probabilistic models: a coarse GP to identify the ROI, and a localized GP for optimization within the ROI. We show theoretically that BALLET can efficiently shrink the search space, and can exhibit a tighter regret bound than standard BO without ROI filtering. We demonstrate empirically the effectiveness of BALLET on both synthetic and real-world optimization tasks.
Rational kernel-based interpolation for complex-valued frequency response functions
Authors: Julien Bect, Niklas Georg, Ulrich Römer, Sebastian Schöps
Abstract
This work is concerned with the kernel-based approximation of a complex-valued function from data, where the frequency response function of a partial differential equation in the frequency domain is of particular interest. In this setting, kernel methods are employed more and more frequently, however, standard kernels do not perform well. Moreover, the role and mathematical implications of the underlying pair of kernels, which arises naturally in the complex-valued case, remain to be addressed. We introduce new reproducing kernel Hilbert spaces of complex-valued functions, and formulate the problem of complex-valued interpolation with a kernel pair as minimum norm interpolation in these spaces. Moreover, we combine the interpolant with a low-order rational function, where the order is adaptively selected based on a new model selection criterion. Numerical results on examples from different fields, including electromagnetics and acoustic examples, illustrate the performance of the method, also in comparison to available rational approximation methods.
Model Calibration in Dense Classification with Adaptive Label Perturbation
Authors: Jiawei Liu, Changkun Ye, Shan Wang, Ruikai Cui, Jing Zhang, Kaihao Zhang, Nick Barnes
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
For safety-related applications, it is crucial to produce trustworthy deep neural networks whose prediction is associated with confidence that can represent the likelihood of correctness for subsequent decision-making. Existing dense binary classification models are prone to being over-confident. To improve model calibration, we propose Adaptive Stochastic Label Perturbation (ASLP) which learns a unique label perturbation level for each training image. ASLP employs our proposed Self-Calibrating Binary Cross Entropy (SC-BCE) loss, which unifies label perturbation processes including stochastic approaches (like DisturbLabel), and label smoothing, to correct calibration while maintaining classification rates. ASLP follows Maximum Entropy Inference of classic statistical mechanics to maximise prediction entropy with respect to missing information. It performs this while: (1) preserving classification accuracy on known data as a conservative solution, or (2) specifically improves model calibration degree by minimising the gap between the prediction accuracy and expected confidence of the target training label. Extensive results demonstrate that ASLP can significantly improve calibration degrees of dense binary classification models on both in-distribution and out-of-distribution data. The code is available on https://github.com/Carlisle-Liu/ASLP.
Abstract
End-to-end region-based object detectors like Sparse R-CNN usually have multiple cascade bounding box decoding stages, which refine the current predictions according to their previous results. Model parameters within each stage are independent, evolving a huge cost. In this paper, we find the general setting of decoding stages is actually redundant. By simply sharing parameters and making a recursive decoder, the detector already obtains a significant improvement. The recursive decoder can be further enhanced by positional encoding (PE) of the proposal box, which makes it aware of the exact locations and sizes of input bounding boxes, thus becoming adaptive to proposals from different stages during the recursion. Moreover, we also design centerness-based PE to distinguish the RoI feature element and dynamic convolution kernels at different positions within the bounding box. To validate the effectiveness of the proposed method, we conduct intensive ablations and build the full model on three recent mainstream region-based detectors. The RecusiveDet is able to achieve obvious performance boosts with even fewer model parameters and slightly increased computation cost. Codes are available at https://github.com/bravezzzzzz/RecursiveDet.
High Probability Analysis for Non-Convex Stochastic Optimization with Clipping
Abstract
Gradient clipping is a commonly used technique to stabilize the training process of neural networks. A growing body of studies has shown that gradient clipping is a promising technique for dealing with the heavy-tailed behavior that emerged in stochastic optimization as well. While gradient clipping is significant, its theoretical guarantees are scarce. Most theoretical guarantees only provide an in-expectation analysis and only focus on optimization performance. In this paper, we provide high probability analysis in the non-convex setting and derive the optimization bound and the generalization bound simultaneously for popular stochastic optimization algorithms with gradient clipping, including stochastic gradient descent and its variants of momentum and adaptive stepsizes. With the gradient clipping, we study a heavy-tailed assumption that the gradients only have bounded $\alpha$-th moments for some $\alpha \in (1, 2]$, which is much weaker than the standard bounded second-moment assumption. Overall, our study provides a relatively complete picture for the theoretical guarantee of stochastic optimization algorithms with clipping.
Keyword: quantization
Fourier-Domain Inversion for the Modulo Radon Transform
Authors: Matthias Beckmann, Ayush Bhandari, Meira Iske
Subjects: Numerical Analysis (math.NA); Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
Inspired by the multiple-exposure fusion approach in computational photography, recently, several practitioners have explored the idea of high dynamic range (HDR) X-ray imaging and tomography. While establishing promising results, these approaches inherit the limitations of multiple-exposure fusion strategy. To overcome these disadvantages, the modulo Radon transform (MRT) has been proposed. The MRT is based on a co-design of hardware and algorithms. In the hardware step, Radon transform projections are folded using modulo non-linearities. Thereon, recovery is performed by algorithmically inverting the folding, thus enabling a single-shot, HDR approach to tomography. The first steps in this topic established rigorous mathematical treatment to the problem of reconstruction from folded projections. This paper takes a step forward by proposing a new, Fourier domain recovery algorithm that is backed by mathematical guarantees. The advantages include recovery at lower sampling rates while being agnostic to modulo threshold, lower computational complexity and empirical robustness to system noise. Beyond numerical simulations, we use prototype modulo ADC based hardware experiments to validate our claims. In particular, we report image recovery based on hardware measurements up to 10 times larger than the sensor's dynamic range while benefiting with lower quantization noise ($\sim$12 dB).
High-Resolution Volumetric Reconstruction for Clothed Humans
Authors: Sicong Tang, Guangyuan Wang, Qing Ran, Lingzhi Li, Li Shen, Ping Tan
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
We present a novel method for reconstructing clothed humans from a sparse set of, e.g., 1 to 6 RGB images. Despite impressive results from recent works employing deep implicit representation, we revisit the volumetric approach and demonstrate that better performance can be achieved with proper system design. The volumetric representation offers significant advantages in leveraging 3D spatial context through 3D convolutions, and the notorious quantization error is largely negligible with a reasonably large yet affordable volume resolution, e.g., 512. To handle memory and computation costs, we propose a sophisticated coarse-to-fine strategy with voxel culling and subspace sparse convolution. Our method starts with a discretized visual hull to compute a coarse shape and then focuses on a narrow band nearby the coarse shape for refinement. Once the shape is reconstructed, we adopt an image-based rendering approach, which computes the colors of surface points by blending input images with learned weights. Extensive experimental results show that our method significantly reduces the mean point-to-surface (P2S) precision of state-of-the-art methods by more than 50% to achieve approximately 2mm accuracy with a 512 volume resolution. Additionally, images rendered from our textured model achieve a higher peak signal-to-noise ratio (PSNR) compared to state-of-the-art methods.
CQNV: A combination of coarsely quantized bitstream and neural vocoder for low rate speech coding
Abstract
Recently, speech codecs based on neural networks have proven to perform better than traditional methods. However, redundancy in traditional parameter quantization is visible within the codec architecture of combining the traditional codec with the neural vocoder. In this paper, we propose a novel framework named CQNV, which combines the coarsely quantized parameters of a traditional parametric codec to reduce the bitrate with a neural vocoder to improve the quality of the decoded speech. Furthermore, we introduce a parameters processing module into the neural vocoder to enhance the application of the bitstream of traditional speech coding parameters to the neural vocoder, further improving the reconstructed speech's quality. In the experiments, both subjective and objective evaluations demonstrate the effectiveness of the proposed CQNV framework. Specifically, our proposed method can achieve higher quality reconstructed speech at 1.1 kbps than Lyra and Encodec at 3 kbps.
QuIP: 2-Bit Quantization of Large Language Models With Guarantees
Authors: Jerry Chee, Yaohui Cai, Volodymyr Kuleshov, Christopher De Sa
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
This work studies post-training parameter quantization in large language models (LLMs). We introduce quantization with incoherence processing (QuIP), a new method based on the insight that quantization benefits from incoherent weight and Hessian matrices, i.e., from the weights and the directions in which it is important to round them accurately being unaligned with the coordinate axes. QuIP consists of two steps: (1) an adaptive rounding procedure minimizing a quadratic proxy objective; (2) efficient pre- and post-processing that ensures weight and Hessian incoherence via multiplication by random orthogonal matrices. We complement QuIP with the first theoretical analysis for an LLM-scale quantization algorithm, and show that our theory also applies to an existing method, OPTQ. Empirically, we find that our incoherence preprocessing improves several existing quantization algorithms and yields the first LLM quantization methods that produce viable results using only two bits per weight. Our code can be found at https://github.com/jerry-chee/QuIP .
Overcoming Distribution Mismatch in Quantizing Image Super-Resolution Networks
Authors: Cheeun Hong, Kyoung Mu Lee
Subjects: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
Abstract
Quantization is a promising approach to reduce the high computational complexity of image super-resolution (SR) networks. However, compared to high-level tasks like image classification, low-bit quantization leads to severe accuracy loss in SR networks. This is because feature distributions of SR networks are significantly divergent for each channel or input image, and is thus difficult to determine a quantization range. Existing SR quantization works approach this distribution mismatch problem by dynamically adapting quantization ranges to the variant distributions during test time. However, such dynamic adaptation incurs additional computational costs that limit the benefits of quantization. Instead, we propose a new quantization-aware training framework that effectively Overcomes the Distribution Mismatch problem in SR networks without the need for dynamic adaptation. Intuitively, the mismatch can be reduced by directly regularizing the variance in features during training. However, we observe that variance regularization can collide with the reconstruction loss during training and adversely impact SR accuracy. Thus, we avoid the conflict between two losses by regularizing the variance only when the gradients of variance regularization are cooperative with that of reconstruction. Additionally, to further reduce the distribution mismatch, we introduce distribution offsets to layers with a significant mismatch, which either scales or shifts channel-wise features. Our proposed algorithm, called ODM, effectively reduces the mismatch in distributions with minimal computational overhead. Experimental results show that ODM effectively outperforms existing SR quantization approaches with similar or fewer computations, demonstrating the importance of reducing the distribution mismatch problem. Our code is available at https://github.com/Cheeun/ODM.
Keyword: efficient
Efficient Behavior-consistent Calibration for Multi-agent Market Simulation
An Explainable Geometric-Weighted Graph Attention Network for Identifying Functional Networks Associated with Gait Impairment
Multi-UAV Speed Control with Collision Avoidance and Handover-aware Cell Association: DRL with Action Branching
Template-Based Static Posterior Inference for Bayesian Probabilistic Programming
Navigating the Web of Misinformation: A Framework for Misinformation Domain Detection Using Browser Traffic
Digital Emotion Regulation on Social Media
Multilevel Large Language Models for Everyone
Rank Optimization for MIMO systems with RIS: Simulation and Measurement
Social Optimum Equilibrium Selection for Distributed Multi-Agent Optimization
A Model Predictive Capture Point Control Framework for Robust Humanoid Balancing via Ankle, Hip, and Stepping Strategies
Federated Split Learning with Only Positive Labels for resource-constrained IoT environment
Unbiased Weight Maximization
QuIP: 2-Bit Quantization of Large Language Models With Guarantees
Federated Heavy Hitter Recovery under Linear Sketching
Learning Regions of Interest for Bayesian Optimization with Adaptive Level-Set Estimation
Scaff-PD: Communication Efficient Fair and Robust Federated Learning
Counterfactual Explanation via Search in Gaussian Mixture Distributed Latent Space
Solving Odd-Fair Parity Games
Scoring Cycling Environments Perceived Safety using Pairwise Image Comparisons
Mitigating Memory Wall Effects in CNN Engines with On-the-Fly Weights Generation
Communication-Efficient Orchestrations for URLLC Service via Hierarchical Reinforcement Learning
A signal processing interpretation of noise-reduction convolutional neural networks
Achieving Linear Speedup in Decentralized Stochastic Compositional Minimax Optimization
Combinatorial Auctions and Graph Neural Networks for Local Energy Flexibility Markets
Exploring MLOps Dynamics: An Experimental Analysis in a Real-World Machine Learning Project
Finding Money Launderers Using Heterogeneous Graph Neural Networks
Preliminary Design of the Dragonfly Navigation Filter
INFINITY: Neural Field Modeling for Reynolds-Averaged Navier-Stokes Equations
Group Activity Recognition in Computer Vision: A Comprehensive Review, Challenges, and Future Perspectives
On Solving the Rubik's Cube with Domain-Independent Planners Using Standard Representations
A threshold dislocation dynamics method
Smartpick: Workload Prediction for Serverless-enabled Scalable Data Analytics Systems
RED CoMETS: An ensemble classifier for symbolically represented multivariate time series
A Compact DAG for Storing and Searching Maximal Common Subsequences
Keyword: faster
Communication-Efficient Orchestrations for URLLC Service via Hierarchical Reinforcement Learning
A behavioural transformer for effective collaboration between a robot and a non-stationary human
Monte-Carlo Tree Search for Multi-Agent Pathfinding: Preliminary Results
Spectrum-guided Multi-granularity Referring Video Object Segmentation
Keyword: mobile
Our Nudges, Our Selves: Tailoring Mobile User Engagement Using Personality
The Fagnano Triangle Patrolling Problem
Advancing Robot Autonomy for Long-Horizon Tasks
Mitigating Memory Wall Effects in CNN Engines with On-the-Fly Weights Generation
An Explainable Model-Agnostic Algorithm for CNN-based Biometrics Verification
Keyword: pruning
There is no result
Keyword: diffusion
Fashion Matrix: Editing Photos by Just Talking
Not with my name! Inferring artists' names of input strings employed by Diffusion Models
XDLM: Cross-lingual Diffusion Language Model for Machine Translation
Fake It Without Making It: Conditioned Face Generation for Accurate 3D Face Shape Estimation
A threshold dislocation dynamics method
Keyword: adaptive
Adaptive Certified Training: Towards Better Accuracy-Robustness Tradeoffs
Text-oriented Modality Reinforcement Network for Multimodal Sentiment Analysis from Unaligned Multimodal Sequences
Multi-Granularity Prediction with Learnable Fusion for Scene Text Recognition
GaitFormer: Revisiting Intrinsic Periodicity for Gait Recognition
Online Maximum Independent Set of Hyperrectangles
QuIP: 2-Bit Quantization of Large Language Models With Guarantees
CT-Net: Arbitrary-Shaped Text Detection via Contour Transformer
Learning Autonomous Ultrasound via Latent Task Representation and Robotic Skills Adaptation
New Adaptive Low-Dissipation Central-Upwind Schemes
rough'' parts of the computed solution, where the piecewise linear reconstruction is performed using an overcompressive limiter, which leads to extremely sharp resolution of shock and contact waves. In the
smooth'' areas, we use a more dissipative limiter to prevent the appearance of artificial kinks and staircase-like structures there. In order to avoid oscillations, we perform the reconstruction in the local characteristic variables obtained using the local characteristic decomposition. We test two different smoothness indicators and apply the developed schemes to the one- and two-dimensional Euler equations of gas dynamics. The obtained numerical results clearly demonstrate that the new adaptive LDCU schemes outperform the original ones.Kefa: A Knowledge Enhanced and Fine-grained Aligned Speaker for Navigation Instruction Generation
Learning Regions of Interest for Bayesian Optimization with Adaptive Level-Set Estimation
Rational kernel-based interpolation for complex-valued frequency response functions
Model Calibration in Dense Classification with Adaptive Label Perturbation
RecursiveDet: End-to-End Region-based Recursive Object Detection
High Probability Analysis for Non-Convex Stochastic Optimization with Clipping
Keyword: quantization
Fourier-Domain Inversion for the Modulo Radon Transform
High-Resolution Volumetric Reconstruction for Clothed Humans
CQNV: A combination of coarsely quantized bitstream and neural vocoder for low rate speech coding
QuIP: 2-Bit Quantization of Large Language Models With Guarantees
Overcoming Distribution Mismatch in Quantizing Image Super-Resolution Networks