Abstract
High penetration of renewable resources results in a power system with lower inertia and higher frequency sensitivity to power imbalances. Such systems are becoming increasingly susceptible to frequency collapse during extreme disturbances. Under-Frequency Load Shedding (UFLS) is a last-resort protection scheme and acts as an emergency brake by shedding load to arrest frequency decline. Current and emerging efforts to optimize UFLS settings and frequency thresholds are mostly network agnostic, ignoring network spatial information. With the prevalence of Distributed Energy Resources (DERs) in the high-renewable paradigm, the power grid is becoming more bidirectional, making some locations in the network less effective for UFLS action than others. This work proposes a Mixed Integer Linear Program that optimizes the UFLS setpoints (prioritizing one location over another) to minimize frequency deviation and load-shed for a given disturbance. The formulation considers system information and DER generation mix at different network locations, increasing model fidelity. The formulation also captures the discrete nature and practical time delays and deadbands associated with UFLS using a minimal set of binary variables, reducing problem complexity. We empirically validate the optimization approach on the dynamic IEEE 39-bus system for performance metrics, including frequency nadir, steady-state frequency and total load shed.
Abstract
Standard approaches for global optimization of non-convex functions, such as branch-and-bound, maintain partition trees to systematically prune the domain. The tree size grows exponentially in the number of dimensions. We propose new sampling-based methods for non-convex optimization that adapts Monte Carlo Tree Search (MCTS) to improve efficiency. Instead of the standard use of visitation count in Upper Confidence Bounds, we utilize numerical overapproximations of the objective as an uncertainty metric, and also take into account of sampled estimates of first-order and second-order information. The Monte Carlo tree in our approach avoids the usual fixed combinatorial patterns in growing the tree, and aggressively zooms into the promising regions, while still balancing exploration and exploitation. We evaluate the proposed algorithms on high-dimensional non-convex optimization benchmarks against competitive baselines and analyze the effects of the hyper parameters.
Graph Learning-based Fleet Scheduling for Urban Air Mobility under Operational Constraints, Varying Demand & Uncertainties
Authors: Authors: Steve Paul, Jhoel Witter, Souma Chowdhury
Subjects: Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Abstract
This paper develops a graph reinforcement learning approach to online planning of the schedule and destinations of electric aircraft that comprise an urban air mobility (UAM) fleet operating across multiple vertiports. This fleet scheduling problem is formulated to consider time-varying demand, constraints related to vertiport capacity, aircraft capacity and airspace safety guidelines, uncertainties related to take-off delay, weather-induced route closures, and unanticipated aircraft downtime. Collectively, such a formulation presents greater complexity, and potentially increased realism, than in existing UAM fleet planning implementations. To address these complexities, a new policy architecture is constructed, primary components of which include: graph capsule conv-nets for encoding vertiport and aircraft-fleet states both abstracted as graphs; transformer layers encoding time series information on demand and passenger fare; and a Multi-head Attention-based decoder that uses the encoded information to compute the probability of selecting each available destination for an aircraft. Trained with Proximal Policy Optimization, this policy architecture shows significantly better performance in terms of daily averaged profits on unseen test scenarios involving 8 vertiports and 40 aircraft, when compared to a random baseline and genetic algorithm-derived optimal solutions, while being nearly 1000 times faster in execution than the latter.
A Universal Cooperative Decision-Making Framework for Connected Autonomous Vehicles with Generic Road Topologies
Authors: Authors: Zhenmin Huang, Shaojie Shen, Jun Ma
Subjects: Robotics (cs.RO); Systems and Control (eess.SY)
Abstract
Cooperative decision-making of Connected Autonomous Vehicles (CAVs) presents a longstanding challenge due to its inherent nonlinearity, non-convexity, and discrete characteristics, compounded by the diverse road topologies encountered in real-world traffic scenarios. The majority of current methodologies are only applicable to a single and specific scenario, predicated on scenario-specific assumptions. Consequently, their application in real-world environments is restricted by the innumerable nature of traffic scenarios. In this study, we propose a unified optimization approach that exhibits the potential to address cooperative decision-making problems related to traffic scenarios with generic road topologies. This development is grounded in the premise that the topologies of various traffic scenarios can be universally represented as Directed Acyclic Graphs (DAGs). Particularly, the reference paths and time profiles for all involved CAVs are determined in a fully cooperative manner, taking into account factors such as velocities, accelerations, conflict resolutions, and overall traffic efficiency. The cooperative decision-making of CAVs is approximated as a mixed-integer linear programming (MILP) problem building on the DAGs of road topologies. This favorably facilitates the use of standard numerical solvers and the global optimality can be attained through the optimization. Case studies corresponding to different multi-lane traffic scenarios featuring diverse topologies are scheduled as the test itineraries, and the efficacy of our proposed methodology is corroborated.
Some Critical Thinking on EV Battery Reliability: from Enhancement to Optimization -- comprehensive perspectives, lifecycle innovation, system cognation, and strategic insights
Abstract
In the era of sustainable transportation, the significance of electric vehicles (EVs) and their battery technology is becoming increasingly paramount. This study addresses the critical aspect of EV battery reliability, an essential factor in the vehicles' sustainability, performance, and longevity. Current efforts to enhance EV battery reliability tend to focus on isolated areas, often missing the broader, interconnected challenges within the system. This research investigates these challenges across micro, meso, and macro levels, presenting a novel lifecycle framework that includes "Zero"-Life reliability and phases such as use, reuse, repurpose, and recycling. By adopting a holistic approach and delving into system cognition, the study aims to bridge the gap between isolated improvements and comprehensive system optimization, aligning with global sustainability goals and contributing to the advancement of sustainable transportation and EV technology.
An Information Theoretic Approach to Interaction-Grounded Learning
Abstract
Reinforcement learning (RL) problems where the learner attempts to infer an unobserved reward from some feedback variables have been studied in several recent papers. The setting of Interaction-Grounded Learning (IGL) is an example of such feedback-based reinforcement learning tasks where the learner optimizes the return by inferring latent binary rewards from the interaction with the environment. In the IGL setting, a relevant assumption used in the RL literature is that the feedback variable $Y$ is conditionally independent of the context-action $(X,A)$ given the latent reward $R$. In this work, we propose Variational Information-based IGL (VI-IGL) as an information-theoretic method to enforce the conditional independence assumption in the IGL-based RL problem. The VI-IGL framework learns a reward decoder using an information-based objective based on the conditional mutual information (MI) between the context-action $(X,A)$ and the feedback variable $Y$ observed from the environment. To estimate and optimize the information-based terms for the continuous random variables in the RL problem, VI-IGL leverages the variational representation of mutual information and results in a min-max optimization problem. Furthermore, we extend the VI-IGL framework to general $f$-Information measures in the information theory literature, leading to the generalized $f$-VI-IGL framework to address the RL problem under the IGL condition. Finally, we provide the empirical results of applying the VI-IGL method to several reinforcement learning settings, which indicate an improved performance in comparison to the previous IGL-based RL algorithm.
Abstract
Existing Global Path Planning (GPP) algorithms predominantly presume planning in a static environment. This assumption immensely limits their applications to Unmanned Surface Vehicles (USVs) that typically navigate in dynamic environments. To address this limitation, we present OkayPlan, a GPP algorithm capable of generating safe and short paths in dynamic scenarios at a real-time executing speed (125 Hz on a desktop-class computer). Specifically, we approach the challenge of dynamic obstacle avoidance by formulating the path planning problem as an obstacle kinematics augmented optimization problem, which can be efficiently resolved through a PSO-based optimizer at a real-time speed. Meanwhile, a Dynamic Prioritized Initialization (DPI) mechanism that adaptively initializes potential solutions for the optimization problem is established to further ameliorate the solution quality. Additionally, a relaxation strategy that facilitates the autonomous tuning of OkayPlan's hyperparameters in dynamic environments is devised. Comparative experiments involving canonical and contemporary GPP algorithms, along with ablation studies, have been conducted to substantiate the efficacy of our approach. Results indicate that OkayPlan outstrips existing methods in terms of path safety, length optimality, and computational efficiency, establishing it as a potent GPP technique for dynamic environments. The video and code associated with this paper are accessible at https://github.com/XinJingHao/OkayPlan.
OTAS: An Elastic Transformer Serving System via Token Adaptation
Authors: Authors: Jinyu Chen, Wenchao Xu, Zicong Hong, Song Guo, Haozhao Wang, Jie Zhang, Deze Zeng
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
Transformer model empowered architectures have become a pillar of cloud services that keeps reshaping our society. However, the dynamic query loads and heterogeneous user requirements severely challenge current transformer serving systems, which rely on pre-training multiple variants of a foundation model, i.e., with different sizes, to accommodate varying service demands. Unfortunately, such a mechanism is unsuitable for large transformer models due to the additional training costs and excessive I/O delay. In this paper, we introduce OTAS, the first elastic serving system specially tailored for transformer models by exploring lightweight token management. We develop a novel idea called token adaptation that adds prompting tokens to improve accuracy and removes redundant tokens to accelerate inference. To cope with fluctuating query loads and diverse user requests, we enhance OTAS with application-aware selective batching and online token adaptation. OTAS first batches incoming queries with similar service-level objectives to improve the ingress throughput. Then, to strike a tradeoff between the overhead of token increment and the potentials for accuracy improvement, OTAS adaptively adjusts the token execution strategy by solving an optimization problem. We implement and evaluate a prototype of OTAS with multiple datasets, which show that OTAS improves the system utility by at least 18.2%.
Accelerating Maximal Biclique Enumeration on GPUs
Authors: Authors: Chou-Ying Hsieh, Chia-Ming Chang, Po-Hsiu Cheng, Sy-Yen Kuo
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
Maximal Biclique Enumeration (MBE) holds critical importance in graph theory with applications extending across fields such as bioinformatics, social networks, and recommendation systems. However, its computational complexity presents barriers for efficiently scaling to large graphs. To address these challenges, we introduce cuMBE, a GPU-optimized parallel algorithm for MBE. Utilizing a unique data structure, called compact array, cuMBE eradicates the need for recursion, thereby significantly minimizing dynamic memory requirements and computational overhead. The algorithm utilizes a hybrid parallelism approach, in which GPU thread blocks handle coarse-grained tasks associated with part of the search process. Besides, we implement three fine-grained optimizations within each thread block to enhance performance. Further, we integrate a work-stealing mechanism to mitigate workload imbalances among thread blocks. Our experiments reveal that cuMBE achieves an geometric mean speedup of 4.02x and 4.13x compared to the state-of-the-art serial algorithm and parallel CPU-based algorithm on both common and real-world datasets, respectively.
An Optimizing Framework on MLIR for Efficient FPGA-based Accelerator Generation
Abstract
With the increasing demand for computing capability given limited resource and power budgets, it is crucial to deploy applications to customized accelerators like FPGAs. However, FPGA programming is non-trivial. Although existing high-level synthesis (HLS) tools improve productivity to a certain extent, they are limited in scope and capability to support sufficient FPGA-oriented optimizations. This paper focuses on FPGA-based accelerators and proposes POM, an optimizing framework built on multi-level intermediate representation (MLIR). POM has several features which demonstrate its scope and capability of performance optimization. First, most HLS tools depend exclusively on a single-level IR to perform all the optimizations, introducing excessive information into the IR and making debugging an arduous task. In contrast, POM introduces three layers of IR to perform operations at suitable abstraction levels, streamlining the implementation and debugging process and exhibiting better flexibility, extensibility, and systematicness. Second, POM integrates the polyhedral model into MLIR, enabling advanced dependence analysis and various FPGA-oriented loop transformations. By representing nested loops with integer sets and maps, loop transformations can be conducted conveniently through manipulations on polyhedral semantics. Finally, to further relieve design effort, POM has a user-friendly programming interface (DSL) that allows a concise description of computation and includes a rich collection of scheduling primitives. An automatic design space exploration (DSE) engine is provided to search for high-performance optimization schemes efficiently and generate optimized accelerators automatically. Experimental results show that POM achieves a $6.46\times$ average speedup on typical benchmark suites and a $6.06\times$ average speedup on real-world applications compared to the state-of-the-art.
Integrated Sensing and Communication with Reconfigurable Distributed Antenna and Reflecting Surface: Joint Beamforming and Mode Selection
Authors: Authors: Pingping Zhang, Jintao Wang, Yulin Shao, Shaodan Ma
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
This paper presents a new integrated sensing and communication (ISAC) framework, leveraging the recent advancements of reconfigurable distributed antenna and reflecting surface (RDARS). RDARS is a programmable surface structure comprising numerous elements, each of which can be flexibly configured to operate either in a reflection mode, resembling a passive reconfigurable intelligent surface (RIS), or in a connected mode, functioning as a remote transmit or receive antenna. Our RDARS-aided ISAC framework effectively mitigates the adverse impact of multiplicative fading when compared to the passive RIS-aided ISAC, and reduces cost and energy consumption when compared to the active RIS-aided ISAC. Within our RDARS-aided ISAC framework, we consider a radar output signal-to-noise ratio (SNR) maximization problem under communication constraints to jointly optimize the active transmit beamforming matrix of the base station (BS), the reflection and mode selection matrices of RDARS, and the receive filter. To tackle the inherent non-convexity and the binary integer optimization introduced by the mode selection in this optimization challenge, we propose an efficient iterative algorithm with proved convergence based on majorization minimization (MM) and penalty-based methods.Numerical and simulation results demonstrate the superior performance of our new framework, and clearly verify substantial distribution, reflection as well as selection gains obtained by properly configuring the RDARS.
Closed-loop Identification of a MSW Grate Incinerator using Bayesian Optimization for Selecting Model Inputs and Structure
Authors: Authors: Johannes Lips, Stefan DeYoung, Max Schönsteiner, Hendrik Lens
Abstract
The creation of low-order dynamic models for complex industrial systems is complicated by disturbances and limited sensor accuracy. This work presents a system identification procedure that uses machine learning methods and process knowledge to robustly identify a low-order closed-loop model of a municipal solid waste (MSW) grate incineration plant. These types of plants are known for their strong disturbances coming from fuel composition fluctuations. Using Bayesian optimization, the algorithm ranks and selects inputs from the available sensor data and chooses the model structure. This results in accurate models with low complexity while avoiding overfitting. The method is applied and validated using data of an industrial MSW incineration plant. The obtained models give excellent predictions and confidence intervals for the steam capacity and intermediate quantities such as supply air flow and flue gas temperature. The identified continuous-time models are fully given, and their step-response dynamics are discussed. The models can be used to develop model-based unit control schemes for grate incineration plants. The presented method shows great potential for the identification of over-actuated systems or disturbed systems with many sensors.
Do Vision and Language Encoders Represent the World Similarly?
Abstract
Aligned text-image encoders such as CLIP have become the de facto model for vision-language tasks. Furthermore, modality-specific encoders achieve impressive performances in their respective domains. This raises a central question: does an alignment exist between uni-modal vision and language encoders since they fundamentally represent the same physical world? Analyzing the latent spaces structure of vision and language models on image-caption benchmarks using the Centered Kernel Alignment (CKA), we find that the representation spaces of unaligned and aligned encoders are semantically similar. In the absence of statistical similarity in aligned encoders like CLIP, we show that a possible matching of unaligned encoders exists without any training. We frame this as a seeded graph-matching problem exploiting the semantic similarity between graphs and propose two methods - a Fast Quadratic Assignment Problem optimization, and a novel localized CKA metric-based matching/retrieval. We demonstrate the effectiveness of this on several downstream tasks including cross-lingual, cross-domain caption matching and image classification.
A Survey on Optimization Studies of Group Centrality Metrics
Authors: Authors: Mustafa Can Camur, Chrysafis Vogiatzis
Subjects: Social and Information Networks (cs.SI); Optimization and Control (math.OC)
Abstract
Centrality metrics have become a popular concept in network science and optimization. Over the years, centrality has been used to assign importance and identify influential elements in various settings, including transportation, infrastructure, biological, and social networks, among others. That said, most of the literature has focused on nodal versions of centrality. Recently, group counterparts of centrality have started attracting scientific and practitioner interest. The identification of sets of nodes that are influential within a network is becoming increasingly more important. This is even more pronounced when these sets of nodes are required to induce a certain motif or structure. In this study, we review group centrality metrics from an operations research and optimization perspective for the first time. This is particularly interesting due to the rapid evolution and development of this area in the operations research community over the last decade. We first present a historical overview of how we have reached this point in the study of group centrality. We then discuss the different structures and motifs that appear prominently in the literature, alongside the techniques and methodologies that are popular. We finally present possible avenues and directions for future work, mainly in three areas: (i) probabilistic metrics to account for randomness along with stochastic optimization techniques; (ii) structures and relaxations that have not been yet studied; and (iii) new emerging applications that can take advantage of group centrality. Our survey offers a concise review of group centrality and its intersection with network analysis and optimization.
Score Distillation Sampling with Learned Manifold Corrective
Abstract
Score Distillation Sampling (SDS) is a recent but already widely popular method that relies on an image diffusion model to control optimization problems using text prompts. In this paper, we conduct an in-depth analysis of the SDS loss function, identify an inherent problem with its formulation, and propose a surprisingly easy but effective fix. Specifically, we decompose the loss into different factors and isolate the component responsible for noisy gradients. In the original formulation, high text guidance is used to account for the noise, leading to unwanted side effects. Instead, we train a shallow network mimicking the timestep-dependent denoising deficiency of the image diffusion model in order to effectively factor it out. We demonstrate the versatility and the effectiveness of our novel loss formulation through several qualitative and quantitative experiments, including optimization-based image synthesis and editing, zero-shot image translation network training, and text-to-3D synthesis.
Keyword: adam
Hyperbolic Machine Learning Moment Closures for the BGK Equations
Authors: Authors: Andrew J. Christlieb, Mingchang Ding, Juntao Huang, Nicholas A. Krupansky
Abstract
We introduce a hyperbolic closure for the Grad moment expansion of the Bhatnagar-Gross-Krook's (BGK) kinetic model using a neural network (NN) trained on BGK's moment data. This closure is motivated by the exact closure for the free streaming limit that we derived in our paper on closures in transport \cite{Huang2022-RTE1}. The exact closure relates the gradient of the highest moment to the gradient of four lower moments. As with our past work, the model presented here learns the gradient of the highest moment in terms of the coefficients of gradients for all lower ones. By necessity, this means that the resulting hyperbolic system is not conservative in the highest moment. For stability, the output layers of the NN are designed to enforce hyperbolicity and Galilean invariance. This ensures the model can be run outside of the training window of the NN. Unlike our previous work on radiation transport that dealt with linear models, the BGK model's nonlinearity demanded advanced training tools. These comprised an optimal learning rate discovery, one cycle training, batch normalization in each neural layer, and the use of the \texttt{AdamW} optimizer. To address the non-conservative structure of the hyperbolic model, we adopt the FORCE numerical method to achieve robust solutions. This results in a comprehensive computing model combining learned closures with methods for solving hyperbolic models. The proposed model can capture accurate moment solutions across a broad spectrum of Knudsen numbers. Our paper details the multi-scale model construction and is run on a range of test problems.
Keyword: gradient
Music Genre Classification: A Comparative Analysis of CNN and XGBoost Approaches with Mel-frequency cepstral coefficients and Mel Spectrograms
Abstract
In recent years, various well-designed algorithms have empowered music platforms to provide content based on one's preferences. Music genres are defined through various aspects, including acoustic features and cultural considerations. Music genre classification works well with content-based filtering, which recommends content based on music similarity to users. Given a considerable dataset, one premise is automatic annotation using machine learning or deep learning methods that can effectively classify audio files. The effectiveness of systems largely depends on feature and model selection, as different architectures and features can facilitate each other and yield different results. In this study, we conduct a comparative study investigating the performances of three models: a proposed convolutional neural network (CNN), the VGG16 with fully connected layers (FC), and an eXtreme Gradient Boosting (XGBoost) approach on different features: 30-second Mel spectrogram and 3-second Mel-frequency cepstral coefficients (MFCCs). The results show that the MFCC XGBoost model outperformed the others. Furthermore, applying data segmentation in the data preprocessing phase can significantly enhance the performance of the CNNs.
Hyperbolic Machine Learning Moment Closures for the BGK Equations
Authors: Authors: Andrew J. Christlieb, Mingchang Ding, Juntao Huang, Nicholas A. Krupansky
Abstract
We introduce a hyperbolic closure for the Grad moment expansion of the Bhatnagar-Gross-Krook's (BGK) kinetic model using a neural network (NN) trained on BGK's moment data. This closure is motivated by the exact closure for the free streaming limit that we derived in our paper on closures in transport \cite{Huang2022-RTE1}. The exact closure relates the gradient of the highest moment to the gradient of four lower moments. As with our past work, the model presented here learns the gradient of the highest moment in terms of the coefficients of gradients for all lower ones. By necessity, this means that the resulting hyperbolic system is not conservative in the highest moment. For stability, the output layers of the NN are designed to enforce hyperbolicity and Galilean invariance. This ensures the model can be run outside of the training window of the NN. Unlike our previous work on radiation transport that dealt with linear models, the BGK model's nonlinearity demanded advanced training tools. These comprised an optimal learning rate discovery, one cycle training, batch normalization in each neural layer, and the use of the \texttt{AdamW} optimizer. To address the non-conservative structure of the hyperbolic model, we adopt the FORCE numerical method to achieve robust solutions. This results in a comprehensive computing model combining learned closures with methods for solving hyperbolic models. The proposed model can capture accurate moment solutions across a broad spectrum of Knudsen numbers. Our paper details the multi-scale model construction and is run on a range of test problems.
Relaxed Contrastive Learning for Federated Learning
Authors: Authors: Seonguk Seo, Jinkyu Kim, Geeho Kim, Bohyung Han
Abstract
We propose a novel contrastive learning framework to effectively address the challenges of data heterogeneity in federated learning. We first analyze the inconsistency of gradient updates across clients during local training and establish its dependence on the distribution of feature representations, leading to the derivation of the supervised contrastive learning (SCL) objective to mitigate local deviations. In addition, we show that a na\"ive adoption of SCL in federated learning leads to representation collapse, resulting in slow convergence and limited performance gains. To address this issue, we introduce a relaxed contrastive learning loss that imposes a divergence penalty on excessively similar sample pairs within each class. This strategy prevents collapsed representations and enhances feature transferability, facilitating collaborative training and leading to significant performance improvements. Our framework outperforms all existing federated learning approaches by huge margins on the standard benchmarks through extensive experimental results.
AdaFed: Fair Federated Learning via Adaptive Common Descent Direction
Authors: Authors: Shayan Mohajer Hamidi, En-Hui Yang
Abstract
Federated learning (FL) is a promising technology via which some edge devices/clients collaboratively train a machine learning model orchestrated by a server. Learning an unfair model is known as a critical problem in federated learning, where the trained model may unfairly advantage or disadvantage some of the devices. To tackle this problem, in this work, we propose AdaFed. The goal of AdaFed is to find an updating direction for the server along which (i) all the clients' loss functions are decreasing; and (ii) more importantly, the loss functions for the clients with larger values decrease with a higher rate. AdaFed adaptively tunes this common direction based on the values of local gradients and loss functions. We validate the effectiveness of AdaFed on a suite of federated datasets, and demonstrate that AdaFed outperforms state-of-the-art fair FL methods.
Distributed Experimental Design Networks
Authors: Authors: Yuanyuan Li, Lili Su, Carlee Joe-Wong, Edmund Yeh, Stratis Ioannidis
Subjects: Networking and Internet Architecture (cs.NI)
Abstract
As edge computing capabilities increase, model learning deployments in diverse edge environments have emerged. In experimental design networks, introduced recently, network routing and rate allocation are designed to aid the transfer of data from sensors to heterogeneous learners. We design efficient experimental design network algorithms that are (a) distributed and (b) use multicast transmissions. This setting poses significant challenges as classic decentralization approaches often operate on (strictly) concave objectives under differentiable constraints. In contrast, the problem we study here has a non-convex, continuous DR-submodular objective, while multicast transmissions naturally result in non-differentiable constraints. From a technical standpoint, we propose a distributed Frank-Wolfe and a distributed projected gradient ascent algorithm that, coupled with a relaxation of non-differentiable constraints, yield allocations within a $1-1/e$ factor from the optimal. Numerical evaluations show that our proposed algorithms outperform competitors with respect to model learning quality.
Sampled-Data Primal-Dual Gradient Dynamics in Model Predictive Control
Abstract
Model Predictive Control (MPC) is a versatile approach capable of accommodating diverse control requirements, holding significant promise for a broad spectrum of industrial applications. Noteworthy challenges associated with MPC include the substantial computational burden and the inherent difficulty in ensuring system stability. Recently, a rapid computation technique has been introduced as a potential solution. This method guides the input toward convergence with the optimal control problem solution by employing the primal-dual gradient (PDG) dynamics as a controller. Furthermore, stability assurances grounded in dissipativity theory have been established. However, these assurances are applicable solely to continuous-time feedback systems. As a consequence, when the controller undergoes discretization and is implemented as a sampled-data system, stability cannot be guaranteed. In this paper, we propose a discrete-time dynamical controller, incorporating specific modifications to the PDG approach, and present stability conditions relevant to the resulting sampled-data system. Additionally, we introduce an extension designed to enhance control performance. Numerical examples substantiate that our proposed method not only enhances control effectiveness but also effectively discerns stability degradation resulting from discretization, a nuance often overlooked by conventional methods.
Exploring Vulnerabilities of No-Reference Image Quality Assessment Models: A Query-Based Black-Box Method
Abstract
No-Reference Image Quality Assessment (NR-IQA) aims to predict image quality scores consistent with human perception without relying on pristine reference images, serving as a crucial component in various visual tasks. Ensuring the robustness of NR-IQA methods is vital for reliable comparisons of different image processing techniques and consistent user experiences in recommendations. The attack methods for NR-IQA provide a powerful instrument to test the robustness of NR-IQA. However, current attack methods of NR-IQA heavily rely on the gradient of the NR-IQA model, leading to limitations when the gradient information is unavailable. In this paper, we present a pioneering query-based black box attack against NR-IQA methods. We propose the concept of \emph{score boundary} and leverage an adaptive iterative approach with multiple score boundaries. Meanwhile, the initial attack directions are also designed to leverage the characteristics of the Human Visual System (HVS). Experiments show our attack method outperforms all compared state-of-the-art methods and is far ahead of previous black-box methods. The effective DBCNN model suffers a Spearman rank-order correlation coefficient (SROCC) decline of $0.6972$ attacked by our method, revealing the vulnerability of NR-IQA to black-box attacks. The proposed attack method also provides a potent tool for further exploration into NR-IQA robustness.
Score Distillation Sampling with Learned Manifold Corrective
Abstract
Score Distillation Sampling (SDS) is a recent but already widely popular method that relies on an image diffusion model to control optimization problems using text prompts. In this paper, we conduct an in-depth analysis of the SDS loss function, identify an inherent problem with its formulation, and propose a surprisingly easy but effective fix. Specifically, we decompose the loss into different factors and isolate the component responsible for noisy gradients. In the original formulation, high text guidance is used to account for the noise, leading to unwanted side effects. Instead, we train a shallow network mimicking the timestep-dependent denoising deficiency of the image diffusion model in order to effectively factor it out. We demonstrate the versatility and the effectiveness of our novel loss formulation through several qualitative and quantitative experiments, including optimization-based image synthesis and editing, zero-shot image translation network training, and text-to-3D synthesis.
Keyword: sgd
There is no result
Keyword: optimization
Optimization-based Framework for Selecting Under-frequency Load Shedding Parameters
Sample-and-Bound for Non-Convex Optimization
Graph Learning-based Fleet Scheduling for Urban Air Mobility under Operational Constraints, Varying Demand & Uncertainties
A Universal Cooperative Decision-Making Framework for Connected Autonomous Vehicles with Generic Road Topologies
Some Critical Thinking on EV Battery Reliability: from Enhancement to Optimization -- comprehensive perspectives, lifecycle innovation, system cognation, and strategic insights
An Information Theoretic Approach to Interaction-Grounded Learning
OkayPlan: Obstacle Kinematics Augmented Dynamic Real-time Path Planning via Particle Swarm Optimization
OTAS: An Elastic Transformer Serving System via Token Adaptation
Accelerating Maximal Biclique Enumeration on GPUs
An Optimizing Framework on MLIR for Efficient FPGA-based Accelerator Generation
Integrated Sensing and Communication with Reconfigurable Distributed Antenna and Reflecting Surface: Joint Beamforming and Mode Selection
Closed-loop Identification of a MSW Grate Incinerator using Bayesian Optimization for Selecting Model Inputs and Structure
Do Vision and Language Encoders Represent the World Similarly?
A Survey on Optimization Studies of Group Centrality Metrics
Score Distillation Sampling with Learned Manifold Corrective
Keyword: adam
Hyperbolic Machine Learning Moment Closures for the BGK Equations
Keyword: gradient
Music Genre Classification: A Comparative Analysis of CNN and XGBoost Approaches with Mel-frequency cepstral coefficients and Mel Spectrograms
Hyperbolic Machine Learning Moment Closures for the BGK Equations
Relaxed Contrastive Learning for Federated Learning
AdaFed: Fair Federated Learning via Adaptive Common Descent Direction
Distributed Experimental Design Networks
Sampled-Data Primal-Dual Gradient Dynamics in Model Predictive Control
Exploring Vulnerabilities of No-Reference Image Quality Assessment Models: A Query-Based Black-Box Method
Score Distillation Sampling with Learned Manifold Corrective
Keyword: super-resolution
There is no result