Abstract
We propose a novel approach for semi-supervised learning (SSL) designed to overcome distribution shifts between training and real-world data arising in the keyword spotting (KWS) task. Shifts from training data distribution are a key challenge for real-world KWS tasks: when a new model is deployed on device, the gating of the accepted data undergoes a shift in distribution, making the problem of timely updates via subsequent deployments hard. Despite the shift, we assume that the marginal distributions on labels do not change. We utilize a modified teacher/student training framework, where labeled training data is augmented with unlabeled data. Note that the teacher does not have access to the new distribution as well. To train effectively with a mix of human and teacher labeled data, we develop a teacher labeling strategy based on confidence heuristics to reduce entropy on the label distribution from the teacher model; the data is then sampled to match the marginal distribution on the labels. Large scale experimental results show that a convolutional neural network (CNN) trained on far-field audio, and evaluated on far-field audio drawn from a different distribution, obtains a 14.3% relative improvement in false discovery rate (FDR) at equal false reject rate (FRR), while yielding a 5% improvement in FDR under no distribution shift. Under a more severe distribution shift from far-field to near-field audio with a smaller fully connected network (FCN) our approach achieves a 52% relative improvement in FDR at equal FRR, while yielding a 20% relative improvement in FDR on the original distribution.
Graph Neural Network Bandits
Authors: Parnian Kassraie, Andreas Krause, Ilija Bogunovic
Abstract
We consider the bandit optimization problem with the reward function defined over graph-structured data. This problem has important applications in molecule design and drug discovery, where the reward is naturally invariant to graph permutations. The key challenges in this setting are scaling to large domains, and to graphs with many nodes. We resolve these challenges by embedding the permutation invariance into our model. In particular, we show that graph neural networks (GNNs) can be used to estimate the reward function, assuming it resides in the Reproducing Kernel Hilbert Space of a permutation-invariant additive kernel. By establishing a novel connection between such kernels and the graph neural tangent kernel (GNTK), we introduce the first GNN confidence bound and use it to design a phased-elimination algorithm with sublinear regret. Our regret bound depends on the GNTK's maximum information gain, which we also provide a bound for. While the reward function depends on all $N$ node features, our guarantees are independent of the number of graph nodes $N$. Empirically, our approach exhibits competitive performance and scales well on graph-structured domains.
T-RECX: Tiny-Resource Efficient Convolutional Neural Networks with Early-Exit
Authors: Nikhil P Ghanathe, Steve Wilton
Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
Abstract
Deploying Machine learning (ML) on the milliwatt-scale edge devices (tinyML) is gaining popularity due to recent breakthroughs in ML and IoT. However, the capabilities of tinyML are restricted by strict power and compute constraints. The majority of the contemporary research in tinyML focuses on model compression techniques such as model pruning and quantization to fit ML models on low-end devices. Nevertheless, the improvements in energy consumption and inference time obtained by existing techniques are limited because aggressive compression quickly shrinks model capacity and accuracy. Another approach to improve inference time and/or reduce power while preserving its model capacity is through early-exit networks. These networks place intermediate classifiers along a baseline neural network that facilitate early exit from neural network computation if an intermediate classifier exhibits sufficient confidence in its prediction. Previous work on early-exit networks have focused on large networks, beyond what would typically be used for tinyML applications. In this paper, we discuss the challenges of adding early-exits to state-of-the-art tiny-CNNs and devise an early-exit architecture, T-RECX, that addresses these challenges. In addition, we develop a method to alleviate the effect of network overthinking at the final exit by leveraging the high-level representations learned by the early-exit. We evaluate T-RECX on three CNNs from the MLPerf tiny benchmark suite for image classification, keyword spotting and visual wake word detection tasks. Our results demonstrate that T-RECX improves the accuracy of baseline network and significantly reduces the average inference time of tiny-CNNs. T-RECX achieves 32.58% average reduction in FLOPS in exchange for 1% accuracy across all evaluated models. Also, our techniques increase the accuracy of baseline network in two out of three models we evaluate
Detecting Volunteer Cotton Plants in a Corn Field with Deep Learning on UAV Remote-Sensing Imagery
Authors: Pappu Kumar Yadav, J. Alex Thomasson, Robert Hardin, Stephen W. Searcy, Ulisses Braga-Neto, Sorin C. Popescu, Daniel E. Martin, Roberto Rodriguez, Karem Meza, Juan Enciso, Jorge Solorzano Diaz, Tianyi Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
The cotton boll weevil, Anthonomus grandis Boheman is a serious pest to the U.S. cotton industry that has cost more than 16 billion USD in damages since it entered the United States from Mexico in the late 1800s. This pest has been nearly eradicated; however, southern part of Texas still faces this issue and is always prone to the pest reinfestation each year due to its sub-tropical climate where cotton plants can grow year-round. Volunteer cotton (VC) plants growing in the fields of inter-seasonal crops, like corn, can serve as hosts to these pests once they reach pin-head square stage (5-6 leaf stage) and therefore need to be detected, located, and destroyed or sprayed . In this paper, we present a study to detect VC plants in a corn field using YOLOv3 on three band aerial images collected by unmanned aircraft system (UAS). The two-fold objectives of this paper were : (i) to determine whether YOLOv3 can be used for VC detection in a corn field using RGB (red, green, and blue) aerial images collected by UAS and (ii) to investigate the behavior of YOLOv3 on images at three different scales (320 x 320, S1; 416 x 416, S2; and 512 x 512, S3 pixels) based on average precision (AP), mean average precision (mAP) and F1-score at 95% confidence level. No significant differences existed for mAP among the three scales, while a significant difference was found for AP between S1 and S3 (p = 0.04) and S2 and S3 (p = 0.02). A significant difference was also found for F1-score between S2 and S3 (p = 0.02). The lack of significant differences of mAP at all the three scales indicated that the trained YOLOv3 model can be used on a computer vision-based remotely piloted aerial application system (RPAAS) for VC detection and spray application in near real-time.
Robot Swarms as Hybrid Systems: Modelling and Verification
Authors: Stefan Schupp (TU Wien), Francesco Leofante (Imperial College London), Leander Behr (RWTH Aachen University), Erika Ábrahám (RWTH Aachen University), Armando Taccella (University of Genoa)
Subjects: Systems and Control (eess.SY); Multiagent Systems (cs.MA)
Abstract
A swarm robotic system consists of a team of robots performing cooperative tasks without any centralized coordination. In principle, swarms enable flexible and scalable solutions; however, designing individual control algorithms that can guarantee a required global behavior is difficult. Formal methods have been suggested by several researchers as a mean to increase confidence in the behavior of the swarm. In this work, we propose to model swarms as hybrid systems and use reachability analysis to verify their properties. We discuss challenges and report on the experience gained from applying hybrid formalisms to the verification of a swarm robotic system.
Confident Adaptive Language Modeling
Authors: Tal Schuster, Adam Fisch, Jai Gupta, Mostafa Dehghani, Dara Bahri, Vinh Q. Tran, Yi Tay, Donald Metzler
Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Recent advances in Transformer-based large language models (LLMs) have led to significant performance improvements across many tasks. These gains come with a drastic increase in the models' size, potentially leading to slow and costly use at inference time. In practice, however, the series of generations made by LLMs is composed of varying levels of difficulty. While certain predictions truly benefit from the models' full capacity, other continuations are more trivial and can be solved with reduced compute. In this work, we introduce Confident Adaptive Language Modeling (CALM), a framework for dynamically allocating different amounts of compute per input and generation timestep. Early exit decoding involves several challenges that we address here, such as: (1) what confidence measure to use; (2) connecting sequence-level constraints to local per-token exit decisions; and (3) attending back to missing hidden representations due to early exits in previous tokens. Through theoretical analysis and empirical experiments on three diverse text generation tasks, we demonstrate the efficacy of our framework in reducing compute -- potential speedup of up to $\times 3$ -- while provably maintaining high performance.
On the Strong Correlation Between Model Invariance and Generalization
Abstract
Generalization and invariance are two essential properties of any machine learning model. Generalization captures a model's ability to classify unseen data while invariance measures consistency of model predictions on transformations of the data. Existing research suggests a positive relationship: a model generalizing well should be invariant to certain visual factors. Building on this qualitative implication we make two contributions. First, we introduce effective invariance (EI), a simple and reasonable measure of model invariance which does not rely on image labels. Given predictions on a test image and its transformed version, EI measures how well the predictions agree and with what level of confidence. Second, using invariance scores computed by EI, we perform large-scale quantitative correlation studies between generalization and invariance, focusing on rotation and grayscale transformations. From a model-centric view, we observe generalization and invariance of different models exhibit a strong linear relationship, on both in-distribution and out-of-distribution datasets. From a dataset-centric view, we find a certain model's accuracy and invariance linearly correlated on different test sets. Apart from these major findings, other minor but interesting insights are also discussed.
Keyword: scaling
Graph Neural Network Bandits
Authors: Parnian Kassraie, Andreas Krause, Ilija Bogunovic
Abstract
We consider the bandit optimization problem with the reward function defined over graph-structured data. This problem has important applications in molecule design and drug discovery, where the reward is naturally invariant to graph permutations. The key challenges in this setting are scaling to large domains, and to graphs with many nodes. We resolve these challenges by embedding the permutation invariance into our model. In particular, we show that graph neural networks (GNNs) can be used to estimate the reward function, assuming it resides in the Reproducing Kernel Hilbert Space of a permutation-invariant additive kernel. By establishing a novel connection between such kernels and the graph neural tangent kernel (GNTK), we introduce the first GNN confidence bound and use it to design a phased-elimination algorithm with sublinear regret. Our regret bound depends on the GNTK's maximum information gain, which we also provide a bound for. While the reward function depends on all $N$ node features, our guarantees are independent of the number of graph nodes $N$. Empirically, our approach exhibits competitive performance and scales well on graph-structured domains.
Rethinking Super-Resolution as Text-Guided Details Generation
Abstract
Deep neural networks have greatly promoted the performance of single image super-resolution (SISR). Conventional methods still resort to restoring the single high-resolution (HR) solution only based on the input of image modality. However, the image-level information is insufficient to predict adequate details and photo-realistic visual quality facing large upscaling factors (x8, x16). In this paper, we propose a new perspective that regards the SISR as a semantic image detail enhancement problem to generate semantically reasonable HR image that are faithful to the ground truth. To enhance the semantic accuracy and the visual quality of the reconstructed image, we explore the multi-modal fusion learning in SISR by proposing a Text-Guided Super-Resolution (TGSR) framework, which can effectively utilize the information from the text and image modalities. Different from existing methods, the proposed TGSR could generate HR image details that match the text descriptions through a coarse-to-fine process. Extensive experiments and ablation studies demonstrate the effect of the TGSR, which exploits the text reference to recover realistic images.
Precision Attitude Stabilization with Intermittent External Torque
Abstract
The attitude stabilization of a micro-satellite employing a variable-amplitude cold gas thruster which reflects as a time varying gain on the control input is considered. Existing literature uses a persistence filter based approach that typically leads to large control gains and torque inputs during specific time intervals corresponding to the 'on' phase of the external actuation. This work aims at reducing the transient spikes placed upon the torque commands by the judicious introduction of an additional time varying scaling signal as part of the control law. The time update mechanism for the new scaling factor and overall closed-loop stability are established through a Lyapunov-like analysis. Numerical simulations highlight the various features of this new control algorithm for spacecraft attitude stabilization subject to torque intermittence.
Keyword: out of distribution detection
There is no result
Keyword: out-of-distribution detection
There is no result
Keyword: expected calibration error
There is no result
Keyword: overconfident
There is no result
Keyword: overconfidence
There is no result
Keyword: confidence
Wakeword Detection under Distribution Shifts
Graph Neural Network Bandits
T-RECX: Tiny-Resource Efficient Convolutional Neural Networks with Early-Exit
Detecting Volunteer Cotton Plants in a Corn Field with Deep Learning on UAV Remote-Sensing Imagery
Robot Swarms as Hybrid Systems: Modelling and Verification
Confident Adaptive Language Modeling
On the Strong Correlation Between Model Invariance and Generalization
Keyword: scaling
Graph Neural Network Bandits
Rethinking Super-Resolution as Text-Guided Details Generation
Precision Attitude Stabilization with Intermittent External Torque
Keyword: calibration
There is no result