Abstract
Representation learning has become a practical family of methods for building rich parametric codifications of massive high-dimensional data while succeeding in the reconstruction side. When considering unsupervised tasks with test-train distribution shifts, the probabilistic viewpoint helps for addressing overconfidence and poor calibration of predictions. However, the direct introduction of Bayesian inference on top of neural networks weights is still an ardous problem for multiple reasons, i.e. the curse of dimensionality or intractability issues. The Laplace approximation (LA) offers a solution here, as one may build Gaussian approximations of the posterior density of weights via second-order Taylor expansions in certain locations of the parameter space. In this work, we present a Bayesian autoencoder for unsupervised representation learning inspired in LA. Our method implements iterative Laplace updates to obtain a novel variational lower-bound of the autoencoder evidence. The vast computational burden of the second-order partial derivatives is skipped via approximations of the Hessian matrix. Empirically, we demonstrate the scalability and performance of the Laplacian autoencoder by providing well-calibrated uncertainties for out-of-distribution detection, geodesics for differential geometry and missing data imputations.
Keyword: expected calibration error
There is no result
Keyword: overconfident
There is no result
Keyword: overconfidence
Laplacian Autoencoders for Learning Stochastic Representations
Authors: Marco Miani, Frederik Warburg, Pablo Moreno-Muñoz, Nicke Skafte Detlefsen, Søren Hauberg
Abstract
Representation learning has become a practical family of methods for building rich parametric codifications of massive high-dimensional data while succeeding in the reconstruction side. When considering unsupervised tasks with test-train distribution shifts, the probabilistic viewpoint helps for addressing overconfidence and poor calibration of predictions. However, the direct introduction of Bayesian inference on top of neural networks weights is still an ardous problem for multiple reasons, i.e. the curse of dimensionality or intractability issues. The Laplace approximation (LA) offers a solution here, as one may build Gaussian approximations of the posterior density of weights via second-order Taylor expansions in certain locations of the parameter space. In this work, we present a Bayesian autoencoder for unsupervised representation learning inspired in LA. Our method implements iterative Laplace updates to obtain a novel variational lower-bound of the autoencoder evidence. The vast computational burden of the second-order partial derivatives is skipped via approximations of the Hessian matrix. Empirically, we demonstrate the scalability and performance of the Laplacian autoencoder by providing well-calibrated uncertainties for out-of-distribution detection, geodesics for differential geometry and missing data imputations.
Keyword: confidence
Benchmarking the Robustness of Deep Neural Networks to Common Corruptions in Digital Pathology
Authors: Yunlong Zhang, Yuxuan Sun, Honglin Li, Sunyi Zheng, Chenglu Zhu, Lin Yang
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
When designing a diagnostic model for a clinical application, it is crucial to guarantee the robustness of the model with respect to a wide range of image corruptions. Herein, an easy-to-use benchmark is established to evaluate how deep neural networks perform on corrupted pathology images. Specifically, corrupted images are generated by injecting nine types of common corruptions into validation images. Besides, two classification and one ranking metrics are designed to evaluate the prediction and confidence performance under corruption. Evaluated on two resulting benchmark datasets, we find that (1) a variety of deep neural network models suffer from a significant accuracy decrease (double the error on clean images) and the unreliable confidence estimation on corrupted images; (2) A low correlation between the validation and test errors while replacing the validation set with our benchmark can increase the correlation. Our codes are available on https://github.com/superjamessyx/robustness_benchmark.
Laplacian Autoencoders for Learning Stochastic Representations
Authors: Marco Miani, Frederik Warburg, Pablo Moreno-Muñoz, Nicke Skafte Detlefsen, Søren Hauberg
Abstract
Representation learning has become a practical family of methods for building rich parametric codifications of massive high-dimensional data while succeeding in the reconstruction side. When considering unsupervised tasks with test-train distribution shifts, the probabilistic viewpoint helps for addressing overconfidence and poor calibration of predictions. However, the direct introduction of Bayesian inference on top of neural networks weights is still an ardous problem for multiple reasons, i.e. the curse of dimensionality or intractability issues. The Laplace approximation (LA) offers a solution here, as one may build Gaussian approximations of the posterior density of weights via second-order Taylor expansions in certain locations of the parameter space. In this work, we present a Bayesian autoencoder for unsupervised representation learning inspired in LA. Our method implements iterative Laplace updates to obtain a novel variational lower-bound of the autoencoder evidence. The vast computational burden of the second-order partial derivatives is skipped via approximations of the Hessian matrix. Empirically, we demonstrate the scalability and performance of the Laplacian autoencoder by providing well-calibrated uncertainties for out-of-distribution detection, geodesics for differential geometry and missing data imputations.
Submission to Generic Event Boundary Detection Challenge@CVPR 2022: Local Context Modeling and Global Boundary Decoding Approach
Authors: Jiaqi Tang, Zhaoyang Liu, Jing Tan, Chen Qian, Wayne Wu, Limin Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Generic event boundary detection (GEBD) is an important yet challenging task in video understanding, which aims at detecting the moments where humans naturally perceive event boundaries. In this paper, we present a local context modeling and global boundary decoding approach for GEBD task. Local context modeling sub-network is proposed to perceive diverse patterns of generic event boundaries, and it generates powerful video representations and reliable boundary confidence. Based on them, global boundary decoding sub-network is exploited to decode event boundaries from a global view. Our proposed method achieves 85.13% F1-score on Kinetics-GEBD testing set, which achieves a more than 22% F1-score boost compared to the baseline method. The code is available at https://github.com/JackyTown/GEBD_Challenge_CVPR2022.
Keyword: scaling
Learnable Model-Driven Performance Prediction and Optimization for Imperfect MIMO System: Framework and Application
Authors: Fan Meng, Shengheng Liu, Yongming Huang, Zhaohua Lu
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
State-of-the-art schemes for performance analysis and optimization of multiple-input multiple-output systems generally experience degradation or even become invalid in dynamic complex scenarios with unknown interference and channel state information (CSI) uncertainty. To adapt to the challenging settings and better accomplish these network auto-tuning tasks, we propose a generic learnable model-driven framework in this paper. To explain how the proposed framework works, we consider regularized zero-forcing precoding as a usage instance and design a light-weight neural network for refined prediction of sum rate and detection error based on coarse model-driven approximations. Then, we estimate the CSI uncertainty on the learned predictor in an iterative manner and, on this basis, optimize the transmit regularization term and subsequent receive power scaling factors. A deep unfolded projected gradient descent based algorithm is proposed for power scaling, which achieves favorable trade-off between convergence rate and robustness.
A Time Series Forecasting Approach to Minimize Cold Start Time in Cloud-Serverless Platform
Authors: Akash Puliyadi Jegannathan, Rounak Saha, Sourav Kanti Addya
Subjects: Networking and Internet Architecture (cs.NI)
Abstract
Serverless computing is a buzzword that is being used commonly in the world of technology and among developers and businesses. Using the Function-as-a-Service (FaaS) model of serverless, one can easily deploy their applications to the cloud and go live in a matter of days, it facilitates the developers to focus on their core business logic and the backend process such as managing the infrastructure, scaling of the application, updation of software and other dependencies is handled by the Cloud Service Provider. One of the features of serverless computing is ability to scale the containers to zero, which results in a problem called cold start. The challenging part is to reduce the cold start latency without the consumption of extra resources. In this paper, we use SARIMA (Seasonal Auto Regressive Integrated Moving Average), one of the classical time series forecasting models to predict the time at which the incoming request comes, and accordingly increase or decrease the amount of required containers to minimize the resource wastage, thus reducing the function launching time. Finally, we implement PBA (Prediction Based Autoscaler) and compare it with the default HPA (Horizontal Pod Autoscaler), which comes inbuilt with kubernetes. The results showed that PBA performs fairly better than the default HPA, while reducing the wastage of resources.
The Body Scaling Effect and Its Impact on Physics Plausibility
Authors: Matti Pouke, Evan G. Center, Alexis P. Chambers, Sakaria Pouke, Timo Ojala, Steven M. LaValle
Abstract
In this study we investigated the effect of body ownership illusion-based body scaling on physics plausibility in Virtual Reality (VR). Our interest was in examining whether body ownership illusion-based body scaling could affect the plausibility of rigid body dynamics similarly to altering VR users' scale by manipulating their virtual interpupillary distance and viewpoint height. The procedure involved the conceptual replication of two previous studies. We investigated physics plausibility with 40 participants under two conditions. In our synchronous condition, we used visuo-tactile stimuli to elicit a body ownership illusion of inhabiting an invisible doll-sized body on participants reclining on an exam table. Our asynchronous condition was otherwise similar, but the visuo-tactile stimuli were provided asynchronously to prevent the onset of the body ownership illusion. We were interested in whether the correct approximation of physics (true physics) or physics that are incorrect and appearing as if the environment is five times larger instead (movie physics) appear more realistic to participants as a function of body scale. We found that movie physics did appear more realistic to participants under the body ownership illusion condition. However, our hypothesis that true physics would appear more realistic in the asynchronous condition was unsupported. Our exploratory analyses revealed that movie physics were perceived as plausible under both conditions. Moreover, we were not able to replicate previous findings from literature concerning object size estimations while inhabiting a small invisible body. However, we found a significant opposite effect regarding size estimations; the object sizes were on average underestimated during the synchronous visuo-tactile condition when compared to the asynchronous condition.
On-Device Training Under 256KB Memory
Authors: Ji Lin, Ligeng Zhu, Wei-Ming Chen, Wei-Chen Wang, Chuang Gan, Song Han
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
On-device training enables the model to adapt to new data collected from the sensors by fine-tuning a pre-trained model. However, the training memory consumption is prohibitive for IoT devices that have tiny memory resources. We propose an algorithm-system co-design framework to make on-device training possible with only 256KB of memory. On-device training faces two unique challenges: (1) the quantized graphs of neural networks are hard to optimize due to mixed bit-precision and the lack of normalization; (2) the limited hardware resource (memory and computation) does not allow full backward computation. To cope with the optimization difficulty, we propose Quantization-Aware Scaling to calibrate the gradient scales and stabilize quantized training. To reduce the memory footprint, we propose Sparse Update to skip the gradient computation of less important layers and sub-tensors. The algorithm innovation is implemented by a lightweight training system, Tiny Training Engine, which prunes the backward computation graph to support sparse updates and offloads the runtime auto-differentiation to compile time. Our framework is the first practical solution for on-device transfer learning of visual recognition on tiny IoT devices (e.g., a microcontroller with only 256KB SRAM), using less than 1/100 of the memory of existing frameworks while matching the accuracy of cloud training+edge deployment for the tinyML application VWW. Our study enables IoT devices to not only perform inference but also continuously adapt to new data for on-device lifelong learning.
Keyword: calibration
Laplacian Autoencoders for Learning Stochastic Representations
Authors: Marco Miani, Frederik Warburg, Pablo Moreno-Muñoz, Nicke Skafte Detlefsen, Søren Hauberg
Abstract
Representation learning has become a practical family of methods for building rich parametric codifications of massive high-dimensional data while succeeding in the reconstruction side. When considering unsupervised tasks with test-train distribution shifts, the probabilistic viewpoint helps for addressing overconfidence and poor calibration of predictions. However, the direct introduction of Bayesian inference on top of neural networks weights is still an ardous problem for multiple reasons, i.e. the curse of dimensionality or intractability issues. The Laplace approximation (LA) offers a solution here, as one may build Gaussian approximations of the posterior density of weights via second-order Taylor expansions in certain locations of the parameter space. In this work, we present a Bayesian autoencoder for unsupervised representation learning inspired in LA. Our method implements iterative Laplace updates to obtain a novel variational lower-bound of the autoencoder evidence. The vast computational burden of the second-order partial derivatives is skipped via approximations of the Hessian matrix. Empirically, we demonstrate the scalability and performance of the Laplacian autoencoder by providing well-calibrated uncertainties for out-of-distribution detection, geodesics for differential geometry and missing data imputations.
Hierarchical Mask Calibration for Unified Domain Adaptive Panoptic Segmentation
Authors: Jingyi Zhang, Jiaxing Huang, Shijian Lu
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Domain adaptive panoptic segmentation aims to mitigate data annotation challenge by leveraging off-the-shelf annotated data in one or multiple related source domains. However, existing studies employ two networks for instance segmentation and semantic segmentation separately which lead to a large amount of network parameters with complicated and computationally intensive training and inference processes. We design UniDAPS, a Unified Domain Adaptive Panoptic Segmentation network that is simple but capable of achieving domain adaptive instance segmentation and semantic segmentation simultaneously within a single network. UniDAPS introduces Hierarchical Mask Calibration (HMC) that rectifies the predicted pseudo masks, pseudo superpixels and pseudo pixels and performs network re-training via an online self-training process on the fly. It has three unique features: 1) it enables unified domain adaptive panoptic adaptation; 2) it mitigates false predictions and improves domain adaptive panoptic segmentation effectively; 3) it is end-to-end trainable with much less parameters and simpler training and inference pipeline. Extensive experiments over multiple public benchmarks show that UniDAPS achieves superior domain adaptive panoptic segmentation as compared with the state-of-the-art.
Forecasting Future World Events with Neural Networks
Authors: Andy Zou, Tristan Xiao, Ryan Jia, Joe Kwon, Mantas Mazeika, Richard Li, Dawn Song, Jacob Steinhardt, Owain Evans, Dan Hendrycks
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
Forecasting future world events is a challenging but valuable task. Forecasts of climate, geopolitical conflict, pandemics and economic indicators help shape policy and decision making. In these domains, the judgment of expert humans contributes to the best forecasts. Given advances in language modeling, can these forecasts be automated? To this end, we introduce Autocast, a dataset containing thousands of forecasting questions and an accompanying news corpus. Questions are taken from forecasting tournaments, ensuring high quality, real-world importance, and diversity. The news corpus is organized by date, allowing us to precisely simulate the conditions under which humans made past forecasts (avoiding leakage from the future). Motivated by the difficulty of forecasting numbers across orders of magnitude (e.g. global cases of COVID-19 in 2022), we also curate IntervalQA, a dataset of numerical questions and metrics for calibration. We test language models on our forecasting task and find that performance is far below a human expert baseline. However, performance improves with increased model size and incorporation of relevant information from the news corpus. In sum, Autocast poses a novel challenge for large language models and improved performance could bring large practical benefits.
Keyword: out of distribution detection
There is no result
Keyword: out-of-distribution detection
Laplacian Autoencoders for Learning Stochastic Representations
Keyword: expected calibration error
There is no result
Keyword: overconfident
There is no result
Keyword: overconfidence
Laplacian Autoencoders for Learning Stochastic Representations
Keyword: confidence
Benchmarking the Robustness of Deep Neural Networks to Common Corruptions in Digital Pathology
Laplacian Autoencoders for Learning Stochastic Representations
Submission to Generic Event Boundary Detection Challenge@CVPR 2022: Local Context Modeling and Global Boundary Decoding Approach
Keyword: scaling
Learnable Model-Driven Performance Prediction and Optimization for Imperfect MIMO System: Framework and Application
A Time Series Forecasting Approach to Minimize Cold Start Time in Cloud-Serverless Platform
The Body Scaling Effect and Its Impact on Physics Plausibility
On-Device Training Under 256KB Memory
Keyword: calibration
Laplacian Autoencoders for Learning Stochastic Representations
Hierarchical Mask Calibration for Unified Domain Adaptive Panoptic Segmentation
Forecasting Future World Events with Neural Networks