鉴于它在提取功能表示方面的力量，对比性的自我监督学习已成功整合到（深）强化学习（RL）的实践中，从而在各种应用程序中提供了有效的政策学习。尽管取得了巨大的经验成功，但对RL的对比学习的理解仍然难以捉摸。为了缩小这样的差距，我们研究了Markov决策过程（MDP）和Markov Games（MGS）的对比度学习如何赋予RL的能力。对于这两种模型，我们建议通过最大程度地减少对比度损失来提取低级别模型的正确特征表示。此外，在在线环境下，我们提出了新颖的上限置信界（UCB）型算法，该算法将这种对比度损失与MDP或MGS的在线RL算法结合在一起。从理论上讲，我们进一步证明了我们的算法恢复了真实表示形式，并同时在学习MDP和MGS中学习最佳策略和NASH平衡方面同时实现了样本效率。我们还提供实证研究，以证明基于UCB的RL的对比度学习方法的功效。据我们所知，我们提供了第一种可证明有效的在线RL算法，该算法结合了代表学习的对比学习。我们的代码可从https://github.com/baichenjia/contrastive-ucb获得。

离线增强学习（RL）提供了一个有希望的方向，可以利用大量离线数据来实现复杂的决策任务。由于分配转移问题，当前的离线RL算法通常被设计为在价值估计和行动选择方面是保守的。但是，这种保守主义在现实情况下遇到观察偏差时，例如传感器错误和对抗性攻击时会损害学习政策的鲁棒性。为了权衡鲁棒性和保守主义，我们通过一种新颖的保守平滑技术提出了强大的离线增强学习（RORL）。在RORL中，我们明确地介绍了数据集附近国家的策略和价值函数的正则化，以及对这些OOD状态的其他保守价值估计。从理论上讲，我们表明RORL比线性MDP中的最新理论结果更紧密地构成。我们证明RORL可以在一般离线RL基准上实现最新性能，并且对对抗性观察的扰动非常强大。

深度强化学习（DRL）和深度多机构的强化学习（MARL）在包括游戏AI，自动驾驶汽车，机器人技术等各种领域取得了巨大的成功。但是，众所周知，DRL和Deep MARL代理的样本效率低下，即使对于相对简单的问题设置，通常也需要数百万个相互作用，从而阻止了在实地场景中的广泛应用和部署。背后的一个瓶颈挑战是众所周知的探索问题，即如何有效地探索环境和收集信息丰富的经验，从而使政策学习受益于最佳研究。在稀疏的奖励，吵闹的干扰，长距离和非平稳的共同学习者的复杂环境中，这个问题变得更加具有挑战性。在本文中，我们对单格和多代理RL的现有勘探方法进行了全面的调查。我们通过确定有效探索的几个关键挑战开始调查。除了上述两个主要分支外，我们还包括其他具有不同思想和技术的著名探索方法。除了算法分析外，我们还对一组常用基准的DRL进行了全面和统一的经验比较。根据我们的算法和实证研究，我们终于总结了DRL和Deep Marl中探索的公开问题，并指出了一些未来的方向。

有效的探索仍然是强化学习中有挑战性的问题，特别是对于来自环境的外在奖励稀疏甚至完全忽视的任务。基于内在动机的重要进展显示了在简单环境中的有希望的结果，但通常会在具有多式联运和随机动力学的环境中陷入困境。在这项工作中，我们提出了一种基于条件变分推理的变分动力模型来模拟多模和随机性。通过在当前状态，动作和潜在变量的条件下产生下一个状态预测，我们考虑作为条件生成过程的环境状态动作转换，这提供了更好地了解动态并在勘探中引发更好的性能。我们派生了环境过渡的负面日志可能性的上限，并使用这样一个上限作为勘探的内在奖励，这使得代理通过自我监督的探索来学习技能，而无需观察外在奖励。我们在基于图像的仿真任务和真正的机器人操纵任务中评估所提出的方法。我们的方法优于若干基于最先进的环境模型的勘探方法。

Dynamic Graph Neural Networks (DGNNs) have been broadly applied in various real-life applications, such as link prediction and pandemic forecast, to capture both static structural information and temporal characteristics from dynamic graphs. Combining both time-dependent and -independent components, DGNNs manifest substantial parallel computation and data reuse potentials, but suffer from severe memory access inefficiency and data transfer overhead under the canonical one-graph-at-a-time training pattern. To tackle the challenges, we propose PiPAD, a $\underline{\textbf{Pi}}pelined$ and $\underline{\textbf{PA}}rallel$ $\underline{\textbf{D}}GNN$ training framework for the end-to-end performance optimization on GPUs. From both the algorithm and runtime level, PiPAD holistically reconstructs the overall training paradigm from the data organization to computation manner. Capable of processing multiple graph snapshots in parallel, PiPAD eliminates the unnecessary data transmission and alleviates memory access inefficiency to improve the overall performance. Our evaluation across various datasets shows PiPAD achieves $1.22\times$-$9.57\times$ speedup over the state-of-the-art DGNN frameworks on three representative models.

We propose a novel teacher-student model for semi-supervised multi-organ segmentation. In teacher-student model, data augmentation is usually adopted on unlabeled data to regularize the consistent training between teacher and student. We start from a key perspective that fixed relative locations and variable sizes of different organs can provide distribution information where a multi-organ CT scan is drawn. Thus, we treat the prior anatomy as a strong tool to guide the data augmentation and reduce the mismatch between labeled and unlabeled images for semi-supervised learning. More specifically, we propose a data augmentation strategy based on partition-and-recovery N$^3$ cubes cross- and within- labeled and unlabeled images. Our strategy encourages unlabeled images to learn organ semantics in relative locations from the labeled images (cross-branch) and enhances the learning ability for small organs (within-branch). For within-branch, we further propose to refine the quality of pseudo labels by blending the learned representations from small cubes to incorporate local attributes. Our method is termed as MagicNet, since it treats the CT volume as a magic-cube and $N^3$-cube partition-and-recovery process matches with the rule of playing a magic-cube. Extensive experiments on two public CT multi-organ datasets demonstrate the effectiveness of MagicNet, and noticeably outperforms state-of-the-art semi-supervised medical image segmentation approaches, with +7% DSC improvement on MACT dataset with 10% labeled images.

The task of referring video object segmentation aims to segment the object in the frames of a given video to which the referring expressions refer. Previous methods adopt multi-stage approach and design complex pipelines to obtain promising results. Recently, the end-to-end method based on Transformer has proved its superiority. In this work, we draw on the advantages of the above methods to provide a simple and effective pipeline for RVOS. Firstly, We improve the state-of-the-art one-stage method ReferFormer to obtain mask sequences that are strongly correlated with language descriptions. Secondly, based on a reliable and high-quality keyframe, we leverage the superior performance of video object segmentation model to further enhance the quality and temporal consistency of the mask results. Our single model reaches 70.3 J &F on the Referring Youtube-VOS validation set and 63.0 on the test set. After ensemble, we achieve 64.1 on the final leaderboard, ranking 1st place on CVPR2022 Referring Youtube-VOS challenge. Code will be available at https://github.com/Zhiweihhh/cvpr2022-rvos-challenge.git.

Referring image segmentation aims to segment the target object described by a given natural language expression. Typically, referring expressions contain complex relationships between the target and its surrounding objects. The main challenge of this task is to understand the visual and linguistic content simultaneously and to find the referred object accurately among all instances in the image. Currently, the most effective way to solve the above problem is to obtain aligned multi-modal features by computing the correlation between visual and linguistic feature modalities under the supervision of the ground-truth mask. However, existing paradigms have difficulty in thoroughly understanding visual and linguistic content due to the inability to perceive information directly about surrounding objects that refer to the target. This prevents them from learning aligned multi-modal features, which leads to inaccurate segmentation. To address this issue, we present a position-aware contrastive alignment network (PCAN) to enhance the alignment of multi-modal features by guiding the interaction between vision and language through prior position information. Our PCAN consists of two modules: 1) Position Aware Module (PAM), which provides position information of all objects related to natural language descriptions, and 2) Contrastive Language Understanding Module (CLUM), which enhances multi-modal alignment by comparing the features of the referred object with those of related objects. Extensive experiments on three benchmarks demonstrate our PCAN performs favorably against the state-of-the-art methods. Our code will be made publicly available.

Continual Learning is considered a key step toward next-generation Artificial Intelligence. Among various methods, replay-based approaches that maintain and replay a small episodic memory of previous samples are one of the most successful strategies against catastrophic forgetting. However, since forgetting is inevitable given bounded memory and unbounded tasks, how to forget is a problem continual learning must address. Therefore, beyond simply avoiding catastrophic forgetting, an under-explored issue is how to reasonably forget while ensuring the merits of human memory, including 1. storage efficiency, 2. generalizability, and 3. some interpretability. To achieve these simultaneously, our paper proposes a new saliency-augmented memory completion framework for continual learning, inspired by recent discoveries in memory completion separation in cognitive neuroscience. Specifically, we innovatively propose to store the part of the image most important to the tasks in episodic memory by saliency map extraction and memory encoding. When learning new tasks, previous data from memory are inpainted by an adaptive data generation module, which is inspired by how humans complete episodic memory. The module's parameters are shared across all tasks and it can be jointly trained with a continual learning classifier as bilevel optimization. Extensive experiments on several continual learning and image classification benchmarks demonstrate the proposed method's effectiveness and efficiency.

Diffractive optical networks provide rich opportunities for visual computing tasks since the spatial information of a scene can be directly accessed by a diffractive processor without requiring any digital pre-processing steps. Here we present data class-specific transformations all-optically performed between the input and output fields-of-view (FOVs) of a diffractive network. The visual information of the objects is encoded into the amplitude (A), phase (P), or intensity (I) of the optical field at the input, which is all-optically processed by a data class-specific diffractive network. At the output, an image sensor-array directly measures the transformed patterns, all-optically encrypted using the transformation matrices pre-assigned to different data classes, i.e., a separate matrix for each data class. The original input images can be recovered by applying the correct decryption key (the inverse transformation) corresponding to the matching data class, while applying any other key will lead to loss of information. The class-specificity of these all-optical diffractive transformations creates opportunities where different keys can be distributed to different users; each user can only decode the acquired images of only one data class, serving multiple users in an all-optically encrypted manner. We numerically demonstrated all-optical class-specific transformations covering A-->A, I-->I, and P-->I transformations using various image datasets. We also experimentally validated the feasibility of this framework by fabricating a class-specific I-->I transformation diffractive network using two-photon polymerization and successfully tested it at 1550 nm wavelength. Data class-specific all-optical transformations provide a fast and energy-efficient method for image and data encryption, enhancing data security and privacy.