Jun Yang

Junjie Zhang, Jiafei Lyu, Xiaoteng Ma et al.

Equipped with the trained environmental dynamics, model-based offline reinforcement learning (RL) algorithms can often successfully learn good policies from fixed-sized datasets, even some datasets with poor quality. Unfortunately, however, it can not be guaranteed that the generated samples from the trained dynamics model are reliable (e.g., some synthetic samples may lie outside of the support region of the static dataset). To address this issue, we propose Trajectory Truncation with Uncertainty (TATU), which adaptively truncates the synthetic trajectory if the accumulated uncertainty along the trajectory is too large. We theoretically show the performance bound of TATU to justify its benefits. To empirically show the advantages of TATU, we first combine it with two classical model-based offline RL algorithms, MOPO and COMBO. Furthermore, we integrate TATU with several off-the-shelf model-free offline RL algorithms, e.g., BCQ. Experimental results on the D4RL benchmark show that TATU significantly improves their performance, often by a large margin. Code is available here.

Yiqin Yang, Hao Hu, Wenzhe Li et al.

Offline reinforcement learning (RL) enables the agent to effectively learn from logged data, which significantly extends the applicability of RL algorithms in real-world scenarios where exploration can be expensive or unsafe. Previous works have shown that extracting primitive skills from the recurring and temporally extended structures in the logged data yields better learning. However, these methods suffer greatly when the primitives have limited representation ability to recover the original policy space, especially in offline settings. In this paper, we give a quantitative characterization of the performance of offline hierarchical learning and highlight the importance of learning lossless primitives. To this end, we propose to use a \emph{flow}-based structure as the representation for low-level policies. This allows us to represent the behaviors in the dataset faithfully while keeping the expression ability to recover the whole policy space. We show that such lossless primitives can drastically improve the performance of hierarchical policies. The experimental results and extensive ablation studies on the standard D4RL benchmark show that our method has a good representation ability for policies and achieves superior performance in most tasks.

Qiyue Yin, Tongtong Yu, Shengqi Shen et al.

With the breakthrough of AlphaGo, deep reinforcement learning becomes a recognized technique for solving sequential decision-making problems. Despite its reputation, data inefficiency caused by its trial and error learning mechanism makes deep reinforcement learning hard to be practical in a wide range of areas. Plenty of methods have been developed for sample efficient deep reinforcement learning, such as environment modeling, experience transfer, and distributed modifications, amongst which, distributed deep reinforcement learning has shown its potential in various applications, such as human-computer gaming, and intelligent transportation. In this paper, we conclude the state of this exciting field, by comparing the classical distributed deep reinforcement learning methods, and studying important components to achieve efficient distributed learning, covering single player single agent distributed deep reinforcement learning to the most complex multiple players multiple agents distributed deep reinforcement learning. Furthermore, we review recently released toolboxes that help to realize distributed deep reinforcement learning without many modifications of their non-distributed versions. By analyzing their strengths and weaknesses, a multi-player multi-agent distributed deep reinforcement learning toolbox is developed and released, which is further validated on Wargame, a complex environment, showing usability of the proposed toolbox for multiple players and multiple agents distributed deep reinforcement learning under complex games. Finally, we try to point out challenges and future trends, hoping this brief review can provide a guide or a spark for researchers who are interested in distributed deep reinforcement learning.

Guorun Wang, Jun Yang, Yaoru Sun

The Outstanding performance and growing size of Large Language Models has led to increased attention in parameter efficient learning. The two predominant approaches are Adapters and Pruning. Adapters are to freeze the model and give it a new weight matrix on the side, which can significantly reduce the time and memory of training, but the cost is that the evaluation and testing will increase the time and memory consumption. Pruning is to cut off some weight and re-distribute the remaining weight, which sacrifices the complexity of training at the cost of extremely high memory and training time, making the cost of evaluation and testing relatively low. So efficiency of training and inference can't be obtained in the same time. In this work, we propose a task-oriented Pruning-Adapter method that achieve a high memory efficiency of training and memory, and speeds up training time and ensures no significant decrease in accuracy in GLUE tasks, achieving training and inference efficiency at the same time.

Siyu Zhang, Yeming Chen, Yaoru Sun et al.

The key to integrating visual language tasks is to establish a good alignment strategy. Recently, visual semantic representation has achieved fine-grained visual understanding by dividing grids or image patches. However, the coarse-grained semantic interactions in image space should not be ignored, which hinders the extraction of complex contextual semantic relations at the scene boundaries. This paper proposes superpixels as comprehensive and robust visual primitives, which mine coarse-grained semantic interactions by clustering perceptually similar pixels, speeding up the subsequent processing of primitives. To capture superpixel-level semantic features, we propose a Multiscale Difference Graph Convolutional Network (MDGCN). It allows parsing the entire image as a fine-to-coarse visual hierarchy. To reason actual semantic relations, we reduce potential noise interference by aggregating difference information between adjacent graph nodes. Finally, we propose a multi-level fusion rule in a bottom-up manner to avoid understanding deviation by mining complementary spatial information at different levels. Experiments show that the proposed method can effectively promote the learning of multiple downstream tasks. Encouragingly, our method outperforms previous methods on all metrics. Our code will be released upon publication.

Yuhua Jiang, Qihan Liu, Xiaoteng Ma et al.

Among the great successes of Reinforcement Learning (RL), self-play algorithms play an essential role in solving competitive games. Current self-play algorithms optimize the agent to maximize expected win-rates against its current or historical copies, making it often stuck in the local optimum and its strategy style simple and homogeneous. A possible solution is to improve the diversity of policies, which helps the agent break the stalemate and enhances its robustness when facing different opponents. However, enhancing diversity in the self-play algorithms is not trivial. In this paper, we aim to introduce diversity from the perspective that agents could have diverse risk preferences in the face of uncertainty. Specifically, we design a novel reinforcement learning algorithm called Risk-sensitive Proximal Policy Optimization (RPPO), which smoothly interpolates between worst-case and best-case policy learning and allows for policy learning with desired risk preferences. Seamlessly integrating RPPO with population-based self-play, agents in the population optimize dynamic risk-sensitive objectives with experiences from playing against diverse opponents. Empirical results show that our method achieves comparable or superior performance in competitive games and that diverse modes of behaviors emerge. Our code is public online at \url{https://github.com/Jackory/RPBT}.

Siyuan Jing, Guangxue Wang, Haoyang Zhai et al.

Due to the emergence of many sign language datasets, isolated sign language recognition (ISLR) has made significant progress in recent years. In addition, the development of various advanced deep neural networks is another reason for this breakthrough. However, challenges remain in applying the technique in the real world. First, existing sign language datasets do not cover the whole sign vocabulary. Second, most of the sign language datasets provide only single view RGB videos, which makes it difficult to handle hand occlusions when performing ISLR. To fill this gap, this paper presents a dual-view sign language dataset for ISLR named NationalCSL-DP, which fully covers the Chinese national sign language vocabulary. The dataset consists of 134140 sign videos recorded by ten signers with respect to two vertical views, namely, the front side and the left side. Furthermore, a CNN transformer network is also proposed as a strong baseline and an extremely simple but effective fusion strategy for prediction. Extensive experiments were conducted to prove the effectiveness of the datasets as well as the baseline. The results show that the proposed fusion strategy can significantly increase the performance of the ISLR, but it is not easy for the sequence-to-sequence model, regardless of whether the early-fusion or late-fusion strategy is applied, to learn the complementary features from the sign videos of two vertical views.

Qiyue Yin, Jun Yang, Kaiqi Huang et al.

With breakthrough of the AlphaGo, human-computer gaming AI has ushered in a big explosion, attracting more and more researchers all around the world. As a recognized standard for testing artificial intelligence, various human-computer gaming AI systems (AIs) have been developed such as the Libratus, OpenAI Five and AlphaStar, beating professional human players. The rapid development of human-computer gaming AIs indicate a big step of decision making intelligence, and it seems that current techniques can handle very complex human-computer games. So, one natural question raises: what are the possible challenges of current techniques in human-computer gaming, and what are the future trends? To answer the above question, in this paper, we survey recent successful game AIs, covering board game AIs, card game AIs, first-person shooting game AIs and real time strategy game AIs. Through this survey, we 1) compare the main difficulties among different kinds of games and the corresponding techniques utilized for achieving professional human level AIs; 2) summarize the mainstream frameworks and techniques that can be properly relied on for developing AIs for complex human-computer gaming; 3) raise the challenges or drawbacks of current techniques in the successful AIs; and 4) try to point out future trends in human-computer gaming AIs. Finally, we hope this brief review can provide an introduction for beginners, and inspire insights for researchers in the field of AI in human-computer gaming.

Yanan Guo, Andrew Zigerelli, Youtao Zhang et al.

Modern x86 processors have many prefetch instructions that can be used by programmers to boost performance. However, these instructions may also cause security problems. In particular, we found that on Intel processors, there are two security flaws in the implementation of PREFETCHW, an instruction for accelerating future writes. First, this instruction can execute on data with read-only permission. Second, the execution time of this instruction leaks the current coherence state of the target data. Based on these two design issues, we build two cross-core private cache attacks that work with both inclusive and non-inclusive LLCs, named Prefetch+Reload and Prefetch+Prefetch. We demonstrate the significance of our attacks in different scenarios. First, in the covert channel case, Prefetch+Reload and Prefetch+Prefetch achieve 782 KB/s and 822 KB/s channel capacities, when using only one shared cache line between the sender and receiver, the largest-to-date single-line capacities for CPU cache covert channels. Further, in the side channel case, our attacks can monitor the access pattern of the victim on the same processor, with almost zero error rate. We show that they can be used to leak private information of real-world applications such as cryptographic keys. Finally, our attacks can be used in transient execution attacks in order to leak more secrets within the transient window than prior work. From the experimental results, our attacks allow leaking about 2 times as many secret bytes, compared to Flush+Reload, which is widely used in transient execution attacks.

Yu-Jun Zheng, Si-Lan Yu, Jun-Chao Yang et al.

Traditional Chinese medicine (TCM) has played an important role in the prevention and control of the novel coronavirus pneumonia (COVID-19), and community prevention has become the most essential part in reducing the spread risk and protecting populations. However, most communities use a uniform TCM prevention program for all residents, which violates the "treatment based on syndrome differentiation" principle of TCM and limits the effectiveness of prevention. In this paper, we propose an intelligent optimization method to develop diversified TCM prevention programs for community residents. First, we use a fuzzy clustering method to divide the population based on both modern medicine and TCM health characteristics; we then use an interactive optimization method, in which TCM experts develop different TCM prevention programs for different clusters, and a heuristic algorithm is used to optimize the programs under the resource constraints. We demonstrate the computational efficiency of the proposed method and report its successful application to TCM-based prevention of COVID-19 in 12 communities in Zhejiang province, China, during the peak of the pandemic.

Xiaoteng Ma, Junyao Chen, Li Xia et al.

We present Distributional Soft Actor-Critic (DSAC), a distributional reinforcement learning (RL) algorithm that combines the strengths of distributional information of accumulated rewards and entropy-driven exploration from Soft Actor-Critic (SAC) algorithm. DSAC models the randomness in both action and rewards, surpassing baseline performances on various continuous control tasks. Unlike standard approaches that solely maximize expected rewards, we propose a unified framework for risk-sensitive learning, one that optimizes the risk-related objective while balancing entropy to encourage exploration. Extensive experiments demonstrate DSAC's effectiveness in enhancing agent performances for both risk-neutral and risk-sensitive control tasks.