Weight decay (WD) is a traditional regularization technique in deep learning, but despite its ubiquity, its behavior is still an area of active research. Golatkar et al. have recently shown that WD only matters at the start of the training in computer vision, upending traditional wisdom. Loshchilov et al. show that for adaptive optimizers, manually decaying weights can outperform adding an $l_2$ penalty to the loss. This technique has become increasingly popular and is referred to as decoupled WD. The goal of this paper is to investigate these two recent empirical observations. We demonstrate that by applying WD only at the start, the network norm stays small throughout training. This has a regularizing effect as the effective gradient updates become larger. However, traditional generalizations metrics fail to capture this effect of WD, and we show how a simple scale-invariant metric can. We also show how the growth of network weights is heavily influenced by the dataset and its generalization properties. For decoupled WD, we perform experiments in NLP and RL where adaptive optimizers are the norm. We demonstrate that the primary issue that decoupled WD alleviates is the mixing of gradients from the objective function and the $l_2$ penalty in the buffers of Adam (which stores the estimates of the first-order moment). Adaptivity itself is not problematic and decoupled WD ensures that the gradients from the $l_2$ term cannot "drown out" the true objective, facilitating easier hyperparameter tuning.

We study the idea of variance reduction applied to adaptive stochastic mirror descent algorithms in nonsmooth nonconvex finite-sum optimization problems. We propose a simple yet generalized adaptive mirror descent algorithm with variance reduction named SVRAMD and provide its convergence analysis in different settings. We prove that variance reduction reduces the gradient complexity of most adaptive mirror descent algorithms and boost their convergence. In particular, our general theory implies variance reduction can be applied to algorithms using time-varying step sizes and self-adaptive algorithms such as AdaGrad and RMSProp. Moreover, our convergence rates recover the best existing rates of non-adaptive algorithms. We check the validity of our claims using experiments in deep learning.

Prediction #6: The U.S. federal government will adopt a more proactive policy approach to AI in 2021 ... [ ] under President Biden. Below are 10 bold predictions about what will unfold in the world of artificial intelligence in 2021, from academic research to startups to capital markets to regulation. To keep ourselves honest, we will revisit these predictions in December 2021 to grade how we did. Autonomous vehicle developers like Waymo and Cruise have massive ongoing cash needs. Public market investors are thirsty for IPOs.

The increasing bandwidth requirement of new wireless applications has lead to standardization of the millimeter wave spectrum for high-speed wireless communication. The millimeter wave spectrum is part of 5G and covers frequencies between 30 and 300 GHz corresponding to wavelengths ranging from 10 to 1 mm. Although millimeter wave is often considered as a communication medium, it has also proved to be an excellent 'sensor', thanks to its narrow beams, operation across a wide bandwidth, and interaction with atmospheric constituents. In this paper, which is to the best of our knowledge the first review that completely covers millimeter wave sensing application pipelines, we provide a comprehensive overview and analysis of different basic application pipeline building blocks, including hardware, algorithms, analytical models, and model evaluation techniques. The review also provides a taxonomy that highlights different millimeter wave sensing application domains. By performing a thorough analysis, complying with the systematic literature review methodology and reviewing 165 papers, we not only extend previous investigations focused only on communication aspects of the millimeter wave technology and using millimeter wave technology for active imaging, but also highlight scientific and technological challenges and trends, and provide a future perspective for applications of millimeter wave as a sensing technology.

Event detection has been an important task in transportation, whose task is to detect points in time when large events disrupts a large portion of the urban traffic network. Travel information {Origin-Destination} (OD) matrix data by map service vendors has large potential to give us insights to discover historic patterns and distinguish anomalies. However, to fully capture the spatial and temporal traffic patterns remains a challenge, yet serves a crucial role for effective anomaly detection. Meanwhile, existing anomaly detection methods have not well-addressed the extreme data sparsity and high-dimension challenges, which are common in OD matrix datasets. To tackle these challenges, we formulate the problem in a novel way, as detecting anomalies in a set of directed weighted graphs representing the traffic conditions at each time interval. We further propose \textit{Context augmented Graph Autoencoder} (\textbf{Con-GAE }), that leverages graph embedding and context embedding techniques to capture the spatial traffic network patterns while working around the data sparsity and high-dimensionality issue. Con-GAE adopts an autoencoder framework and detect anomalies via semi-supervised learning. Extensive experiments show that our method can achieve up can achieve a 0.1-0.4 improvements of the area under the curve (AUC) score over state-of-art anomaly detection baselines, when applied on several real-world large scale OD matrix datasets.

In artificial intelligence, knowledge is the information required by an intelligent system to accomplish tasks. While traditional knowledge bases use discrete, symbolic representations, detecting knowledge encoded in the continuous representations learned from data has received increasing attention recently. In this work, we propose a method for building a continuous knowledge base that can store knowledge imported from multiple, diverse neural networks. The key idea of our approach is to define an interface for each neural network and cast knowledge transferring as a function simulation problem. Preliminary experiments on text classification show promising results: we first import the knowledge encoded in an RNN model and a CNN model to the knowledge base, from which the fused knowledge is exported back to the RNN model, achieving a higher classification accuracy than the original RNN model. With the continuous knowledge base, it is also easy to achieve knowledge distillation and transfer learning. Our work opens the door to building a universal continuous knowledge base to collect, store, and organize all continuous knowledge encoded in different neural networks trained for different AI tasks.

Traditional deep neural networks (NNs) have significantly contributed to the state-of-the-art performance in the task of classification under various application domains. However, NNs have not considered inherent uncertainty in data associated with the class probabilities where misclassification under uncertainty may easily introduce high risk in decision making in real-world contexts (e.g., misclassification of objects in roads leads to serious accidents). Unlike Bayesian NN that indirectly infer uncertainty through weight uncertainties, evidential NNs (ENNs) have been recently proposed to explicitly model the uncertainty of class probabilities and use them for classification tasks. An ENN offers the formulation of the predictions of NNs as subjective opinions and learns the function by collecting an amount of evidence that can form the subjective opinions by a deterministic NN from data. However, the ENN is trained as a black box without explicitly considering inherent uncertainty in data with their different root causes, such as vacuity (i.e., uncertainty due to a lack of evidence) or dissonance (i.e., uncertainty due to conflicting evidence). By considering the multidimensional uncertainty, we proposed a novel uncertainty-aware evidential NN called WGAN-ENN (WENN) for solving an out-of-distribution (OOD) detection problem. We took a hybrid approach that combines Wasserstein Generative Adversarial Network (WGAN) with ENNs to jointly train a model with prior knowledge of a certain class, which has high vacuity for OOD samples. Via extensive empirical experiments based on both synthetic and real-world datasets, we demonstrated that the estimation of uncertainty by WENN can significantly help distinguish OOD samples from boundary samples. WENN outperformed in OOD detection when compared with other competitive counterparts.

Dental caries is one of the most chronic diseases involving the majority of the population during their lifetime. Caries lesions are typically diagnosed by radiologists relying only on their visual inspection to detect via dental x-rays. In many cases, dental caries is hard to identify using x-rays and can be misinterpreted as shadows due to different reasons such as low image quality. Hence, developing a decision support system for caries detection has been a topic of interest in recent years. Here, we propose an automatic diagnosis system to detect dental caries in Panoramic images for the first time, to the best of authors' knowledge. The proposed model benefits from various pretrained deep learning models through transfer learning to extract relevant features from x-rays and uses a capsule network to draw prediction results. On a dataset of 470 Panoramic images used for features extraction, including 240 labeled images for classification, our model achieved an accuracy score of 86.05% on the test set. The obtained score demonstrates acceptable detection performance and an increase in caries detection speed, as long as the challenges of using Panoramic x-rays of real patients are taken into account. Among images with caries lesions in the test set, our model acquired recall scores of 69.44% and 90.52% for mild and severe ones, confirming the fact that severe caries spots are more straightforward to detect and efficient mild caries detection needs a more robust and larger dataset. Considering the novelty of current research study as using Panoramic images, this work is a step towards developing a fully automated efficient decision support system to assist domain experts. Dental caries, also known as tooth decay, is one of the most prevalent infectious chronic dental diseases in humans, affecting individuals throughout their lifetime [1]. According to the National Health and Nutrition Examination Survey, dental caries involves approximately 90% of adults in the United States [2], [3]. Dental caries is a dynamic disease procedure resulting from dental biofilm's metabolic activity, which gradually demineralizes enamel and dentine [4]. Tooth decay is a preventable disease, and if detected, can be stopped and potentially reversed in its early stages [5]. A standard tool for radiologists to distinguish dental diseases, such as caries, is x-ray radiography.

Adversarial Training (AT) with Projected Gradient Descent (PGD) is an effective approach for improving the robustness of the deep neural networks. However, PGD AT has been shown to suffer from two main limitations: i) high computational cost, and ii) extreme overfitting during training that leads to reduction in model generalization. While the effect of factors such as model capacity and scale of training data on adversarial robustness have been extensively studied, little attention has been paid to the effect of a very important parameter in every network optimization on adversarial robustness: the learning rate. In particular, we hypothesize that effective learning rate scheduling during adversarial training can significantly reduce the overfitting issue, to a degree where one does not even need to adversarially train a model from scratch but can instead simply adversarially fine-tune a pre-trained model. Motivated by this hypothesis, we propose a simple yet very effective adversarial fine-tuning approach based on a $\textit{slow start, fast decay}$ learning rate scheduling strategy which not only significantly decreases computational cost required, but also greatly improves the accuracy and robustness of a deep neural network. Experimental results show that the proposed adversarial fine-tuning approach outperforms the state-of-the-art methods on CIFAR-10, CIFAR-100 and ImageNet datasets in both test accuracy and the robustness, while reducing the computational cost by 8-10$\times$. Furthermore, a very important benefit of the proposed adversarial fine-tuning approach is that it enables the ability to improve the robustness of any pre-trained deep neural network without needing to train the model from scratch, which to the best of the authors' knowledge has not been previously demonstrated in research literature.

The ever-growing popularity and rapid improving of artificial intelligence (AI) have raised rethinking on the evolution of wireless networks. Mobile edge computing (MEC) provides a natural platform for AI applications since it is with rich computation resources to train machine learning (ML) models, as well as low-latency access to the data generated by mobile and internet of things (IoT) devices. In this paper, we present an infrastructure to perform ML tasks at an MEC server with the assistance of a reconfigurable intelligent surface (RIS). In contrast to conventional communication systems where the principal criterions are to maximize the throughput, we aim at maximizing the learning performance. Specifically, we minimize the maximum learning error of all participating users by jointly optimizing transmit power of mobile users, beamforming vectors of the base station (BS), and the phase-shift matrix of the RIS. An alternating optimization (AO)-based framework is proposed to optimize the three terms iteratively, where a successive convex approximation (SCA)-based algorithm is developed to solve the power allocation problem, closed-form expressions of the beamforming vectors are derived, and an alternating direction method of multipliers (ADMM)-based algorithm is designed together with an error level searching (ELS) framework to effectively solve the challenging nonconvex optimization problem of the phase-shift matrix. Simulation results demonstrate significant gains of deploying an RIS and validate the advantages of our proposed algorithms over various benchmarks. Lastly, a unified communication-training-inference platform is developed based on the CARLA platform and the SECOND network, and a use case (3D object detection in autonomous driving) for the proposed scheme is demonstrated on the developed platform.