Revisiting Estimation Bias in Policy Gradients for Deep Reinforcement Learning

We revisit the estimation bias in policy gradients for the discounted episodic Markov decision process (MDP) from Deep Reinforcement Learning (DRL) perspective. The objective is formulated theoretically as the expected returns discounted over the time horizon. One of the major policy gradient biases is the state distribution shift: the state distribution used to estimate the gradients differs from the theoretical formulation in that it does not take into account the discount factor. Existing discussion of the influence of this bias was limited to the tabular and softmax cases in the literature. Therefore, in this paper, we extend it to the DRL setting where the policy is parameterized and demonstrate how this bias can lead to suboptimal policies theoretically. We then discuss why the empirically inaccurate implementations with shifted state distribution can still be effective. We show that, despite such state distribution shift, the policy gradient estimation bias can be reduced in the following three ways: 1) a small learning rate; 2) an adaptive-learning-rate-based optimizer; and 3) KL regularization. Specifically, we show that a smaller learning rate, or, an adaptive learning rate, such as that used by Adam and RSMProp optimizers, makes the policy optimization robust to the bias. We further draw connections between optimizers and the optimization regularization to show that both the KL and the reverse KL regularization can significantly rectify this bias. Moreover, we provide extensive experiments on continuous control tasks to support our analysis. Our paper sheds light on how successful PG algorithms optimize policies in the DRL setting, and contributes insights into the practical issues in DRL.

Artificial intelligence for diagnosing and predicting survival of patients with renal cell carcinoma: Retrospective multi-center study

Background: Clear cell renal cell carcinoma (ccRCC) is the most common renal-related tumor with high heterogeneity. There is still an urgent need for novel diagnostic and prognostic biomarkers for ccRCC. Methods: We proposed a weakly-supervised deep learning strategy using conventional histology of 1752 whole slide images from multiple centers. Our study was demonstrated through internal cross-validation and external validations for the deep learning-based models. Results: Automatic diagnosis for ccRCC through intelligent subtyping of renal cell carcinoma was proved in this study. Our graderisk achieved aera the curve (AUC) of 0.840 (95% confidence interval: 0.805-0.871) in the TCGA cohort, 0.840 (0.805-0.871) in the General cohort, and 0.840 (0.805-0.871) in the CPTAC cohort for the recognition of high-grade tumor. The OSrisk for the prediction of 5-year survival status achieved AUC of 0.784 (0.746-0.819) in the TCGA cohort, which was further verified in the independent General cohort and the CPTAC cohort, with AUC of 0.774 (0.723-0.820) and 0.702 (0.632-0.765), respectively. Cox regression analysis indicated that graderisk, OSrisk, tumor grade, and tumor stage were found to be independent prognostic factors, which were further incorporated into the competing-risk nomogram (CRN). Kaplan-Meier survival analyses further illustrated that our CRN could significantly distinguish patients with high survival risk, with hazard ratio of 5.664 (3.893-8.239, p < 0.0001) in the TCGA cohort, 35.740 (5.889-216.900, p < 0.0001) in the General cohort and 6.107 (1.815 to 20.540, p < 0.0001) in the CPTAC cohort. Comparison analyses conformed that our CRN outperformed current prognosis indicators in the prediction of survival status, with higher concordance index for clinical prognosis.

RLogist: Fast Observation Strategy on Whole-slide Images with Deep Reinforcement Learning

Whole-slide images (WSI) in computational pathology have high resolution with gigapixel size, but are generally with sparse regions of interest, which leads to weak diagnostic relevance and data inefficiency for each area in the slide. Most of the existing methods rely on a multiple instance learning framework that requires densely sampling local patches at high magnification. The limitation is evident in the application stage as the heavy computation for extracting patch-level features is inevitable. In this paper, we develop RLogist, a benchmarking deep reinforcement learning (DRL) method for fast observation strategy on WSIs. Imitating the diagnostic logic of human pathologists, our RL agent learns how to find regions of observation value and obtain representative features across multiple resolution levels, without having to analyze each part of the WSI at the high magnification. We benchmark our method on two whole-slide level classification tasks, including detection of metastases in WSIs of lymph node sections, and subtyping of lung cancer. Experimental results demonstrate that RLogist achieves competitive classification performance compared to typical multiple instance learning algorithms, while having a significantly short observation path. In addition, the observation path given by RLogist provides good decision-making interpretability, and its ability of reading path navigation can potentially be used by pathologists for educational/assistive purposes. Our code is available at: \url{https://github.com/tencent-ailab/RLogist}.

Dynamic Feature Pruning and Consolidation for Occluded Person Re-Identification

Occluded person re-identification (ReID) is a challenging problem due to contamination from occluders, and existing approaches address the issue with prior knowledge cues, eg human body key points, semantic segmentations and etc, which easily fails in the presents of heavy occlusion and other humans as occluders. In this paper, we propose a feature pruning and consolidation (FPC) framework to circumvent explicit human structure parse, which mainly consists of a sparse encoder, a global and local feature ranking module, and a feature consolidation decoder. Specifically, the sparse encoder drops less important image tokens (mostly related to background noise and occluders) solely according to correlation within the class token attention instead of relying on prior human shape information. Subsequently, the ranking stage relies on the preserved tokens produced by the sparse encoder to identify k-nearest neighbors from a pre-trained gallery memory by measuring the image and patch-level combined similarity. Finally, we use the feature consolidation module to compensate pruned features using identified neighbors for recovering essential information while disregarding disturbance from noise and occlusion. Experimental results demonstrate the effectiveness of our proposed framework on occluded, partial and holistic Re-ID datasets. In particular, our method outperforms state-of-the-art results by at least 8.6% mAP and 6.0% Rank-1 accuracy on the challenging Occluded-Duke dataset.

Joint Beamforming Design and 3D DoA Estimation for RIS-aided Communication System

In this paper, we consider a reconfigurable intelligent surface (RIS)-assisted 3D direction-of-arrival (DoA) estimation system, in which a uniform planar array (UPA) RIS is deployed to provide virtual line-of-sight (LOS) links and reflect the uplink pilot signal to sensors. To overcome the mutually coupled problem between the beamforming design at the RIS and DoA estimation, we explore the separable sparse representation structure and propose an alternating optimization algorithm. The grid-based DoA estimation is modeled as a joint-sparse recovery problem considering the grid bias, and the Joint-2D-OMP method is used to estimate both on-grid and off-grid parts. The corresponding Cram\'er-Rao lower bound (CRLB) is derived to evaluate the estimation. Then, the beampattern at the RIS is optimized to maximize the signal-to-noise (SNR) at sensors according to the estimated angles. Numerical results show that the proposed alternating optimization algorithm can achieve lower estimation error compared to benchmarks of random beamforming design.

Learning Robust Real-World Dexterous Grasping Policies via Implicit Shape Augmentation

Dexterous robotic hands have the capability to interact with a wide variety of household objects to perform tasks like grasping. However, learning robust real world grasping policies for arbitrary objects has proven challenging due to the difficulty of generating high quality training data. In this work, we propose a learning system (ISAGrasp) for leveraging a small number of human demonstrations to bootstrap the generation of a much larger dataset containing successful grasps on a variety of novel objects. Our key insight is to use a correspondence-aware implicit generative model to deform object meshes and demonstrated human grasps in order to generate a diverse dataset of novel objects and successful grasps for supervised learning, while maintaining semantic realism. We use this dataset to train a robust grasping policy in simulation which can be deployed in the real world. We demonstrate grasping performance with a four-fingered Allegro hand in both simulation and the real world, and show this method can handle entirely new semantic classes and achieve a 79% success rate on grasping unseen objects in the real world.

TestAug: A Framework for Augmenting Capability-based NLP Tests

The recently proposed capability-based NLP testing allows model developers to test the functional capabilities of NLP models, revealing functional failures that cannot be detected by the traditional heldout mechanism. However, existing work on capability-based testing requires extensive manual efforts and domain expertise in creating the test cases. In this paper, we investigate a low-cost approach for the test case generation by leveraging the GPT-3 engine. We further propose to use a classifier to remove the invalid outputs from GPT-3 and expand the outputs into templates to generate more test cases. Our experiments show that TestAug has three advantages over the existing work on behavioral testing: (1) TestAug can find more bugs than existing work; (2) The test cases in TestAug are more diverse; and (3) TestAug largely saves the manual efforts in creating the test suites. The code and data for TestAug can be found at our project website (https://guanqun-yang.github.io/testaug/) and GitHub (https://github.com/guanqun-yang/testaug).

DeepPerform: An Efficient Approach for Performance Testing of Resource-Constrained Neural Networks

Today, an increasing number of Adaptive Deep Neural Networks (AdNNs) are being used on resource-constrained embedded devices. We observe that, similar to traditional software, redundant computation exists in AdNNs, resulting in considerable performance degradation. The performance degradation is dependent on the input and is referred to as input-dependent performance bottlenecks (IDPBs). To ensure an AdNN satisfies the performance requirements of resource-constrained applications, it is essential to conduct performance testing to detect IDPBs in the AdNN. Existing neural network testing methods are primarily concerned with correctness testing, which does not involve performance testing. To fill this gap, we propose DeepPerform, a scalable approach to generate test samples to detect the IDPBs in AdNNs. We first demonstrate how the problem of generating performance test samples detecting IDPBs can be formulated as an optimization problem. Following that, we demonstrate how DeepPerform efficiently handles the optimization problem by learning and estimating the distribution of AdNNs' computational consumption. We evaluate DeepPerform on three widely used datasets against five popular AdNN models. The results show that DeepPerform generates test samples that cause more severe performance degradation (FLOPs: increase up to 552\%). Furthermore, DeepPerform is substantially more efficient than the baseline methods in generating test inputs(runtime overhead: only 6-10 milliseconds).

Honor of Kings Arena: an Environment for Generalization in Competitive Reinforcement Learning

This paper introduces Honor of Kings Arena, a reinforcement learning (RL) environment based on Honor of Kings, one of the world's most popular games at present. Compared to other environments studied in most previous work, ours presents new generalization challenges for competitive reinforcement learning. It is a multi-agent problem with one agent competing against its opponent; and it requires the generalization ability as it has diverse targets to control and diverse opponents to compete with. We describe the observation, action, and reward specifications for the Honor of Kings domain and provide an open-source Python-based interface for communicating with the game engine. We provide twenty target heroes with a variety of tasks in Honor of Kings Arena and present initial baseline results for RL-based methods with feasible computing resources. Finally, we showcase the generalization challenges imposed by Honor of Kings Arena and possible remedies to the challenges. All of the software, including the environment-class, are publicly available at https://github.com/tencent-ailab/hok_env . The documentation is available at https://aiarena.tencent.com/hok/doc/ .