In many control problems that include vision, optimal controls can be inferred from the location of the objects in the scene. This information can be represented using feature points, which is a list of spatial locations in learned feature maps of an input image. Previous works show that feature points learned using unsupervised pre-training or human supervision can provide good features for control tasks. In this paper, we show that it is possible to learn efficient feature point representations end-to-end, without the need for unsupervised pre-training, decoders, or additional losses. Our proposed architecture consists of a differentiable feature point extractor that feeds the coordinates of the estimated feature points directly to a soft actor-critic agent. The proposed algorithm yields performance competitive to the state-of-the art on DeepMind Control Suite tasks.
Reinforcement learning optimizes policies for expected cumulative reward. Need the supervision be so narrow? Reward is delayed and sparse for many tasks, making it a difficult and impoverished signal for end-to-end optimization. To augment reward, we consider a range of self-supervised tasks that incorporate states, actions, and successors to provide auxiliary losses. These losses offer ubiquitous and instantaneous supervision for representation learning even in the absence of reward. While current results show that learning from reward alone is feasible, pure reinforcement learning methods ar…
It is a long-standing problem to find effective representations for training reinforcement learning (RL) agents. This paper demonstrates that learning state representations with supervision from Neural Radiance Fields (NeRFs) can improve the performance of RL compared to other learned representations or even low-dimensional, hand-engineered state information. Specifically, we propose to train an encoder that maps multiple image observations to a latent space describing the objects in the scene. The decoder built from a latent-conditioned NeRF serves as the supervision signal to learn the laten…
Supervised regression to demonstrations has been demonstrated to be a stable way to train deep policy networks. We are motivated to study how we can take full advantage of supervised loss functions for stably training deep reinforcement learning agents. This is a challenging task because it is unclear how the training data could be collected to enable policy improvement. In this work, we propose Self-Supervised Reinforcement Learning (SSRL), a simple algorithm that optimizes policies with purely supervised losses. We demonstrate that, without policy gradient or value estimation, an iterative p…
In this work, we investigate the properties of data that cause popular representation learning approaches to fail. In particular, we find that in environments where states do not significantly overlap, variational autoencoders (VAEs) fail to learn useful features. We demonstrate this failure in a simple gridworld domain, and then provide a solution in the form of metric learning. However, metric learning requires supervision in the form of a distance function, which is absent in reinforcement learning. To overcome this, we leverage the sequential nature of states in a replay buffer to approxim…
Nanjing University of Aeronautics and Astronautics（南京航空航天大学）
Reinforcement learning has achieved great success in various applications. To learn an effective policy for the agent, it usually requires a huge amount of data by interacting with the environment, which could be computational costly and time consuming. To overcome this challenge, the framework called Reinforcement Learning with Expert Demonstrations (RLED) was proposed to exploit the supervision from expert demonstrations. Although the RLED methods can reduce the number of learning iterations, they usually assume the demonstrations are perfect, and thus may be seriously misled by the noisy de…
University of Toronto（多伦多大学）；Vector Institute（向量研究所）；Stanford University（斯坦福大学）；Nvidia（英伟达）
Offline reinforcement learning proposes to learn policies from large collected datasets without interacting with the physical environment. These algorithms have made it possible to learn useful skills from data that can then be deployed in the environment in real-world settings where interactions may be costly or dangerous, such as autonomous driving or factories. However, current algorithms overfit to the dataset they are trained on and exhibit poor out-of-distribution generalization to the environment when deployed. In this paper, we study the effectiveness of performing data augmentations o…
Training automated agents to complete complex tasks in interactive environments is challenging: reinforcement learning requires careful hand-engineering of reward functions, imitation learning requires specialized infrastructure and access to a human expert, and learning from intermediate forms of supervision (like binary preferences) is time-consuming and extracts little information from each human intervention. Can we overcome these challenges by building agents that learn from rich, interactive feedback instead? We propose a new supervision paradigm for interactive learning based on "teacha…
Using a model of the environment, reinforcement learning agents can plan their future moves and achieve superhuman performance in board games like Chess, Shogi, and Go, while remaining relatively sample-efficient. As demonstrated by the MuZero Algorithm, the environment model can even be learned dynamically, generalizing the agent to many more tasks while at the same time achieving state-of-the-art performance. Notably, MuZero uses internal state representations derived from real environment states for its predictions. In this paper, we bind the model's predicted internal state representation …
This paper addresses the question of how a previously available control policy $π_s$ can be used as a supervisor to more quickly and safely train a new learned control policy $π_L$ for a robot. A weighted average of the supervisor and learned policies is used during trials, with a heavier weight initially on the supervisor, in order to allow safe and useful physical trials while the learned policy is still ineffective. During the process, the weight is adjusted to favor the learned policy. As weights are adjusted, the learned network must compensate so as to give safe and reasonable outputs un…
Mila - Quebec Institute of Artificial Intelligence（米拉 - 魁北克人工智能研究所）；Université de Montréal（蒙特利尔大学）
In model-free deep reinforcement learning (RL) algorithms, using noisy value estimates to supervise policy evaluation and optimization is detrimental to the sample efficiency. As this noise is heteroscedastic, its effects can be mitigated using uncertainty-based weights in the optimization process. Previous methods rely on sampled ensembles, which do not capture all aspects of uncertainty. We provide a systematic analysis of the sources of uncertainty in the noisy supervision that occurs in RL, and introduce inverse-variance RL, a Bayesian framework which combines probabilistic ensembles and B…
The use of neural networks and reinforcement learning has become increasingly popular in autonomous vehicle control. However, the opaqueness of the resulting control policies presents a significant barrier to deploying neural network-based control in autonomous vehicles. In this paper, we present a reinforcement learning based approach to autonomous vehicle longitudinal control, where the rule-based safety cages provide enhanced safety for the vehicle as well as weak supervision to the reinforcement learning agent. By guiding the agent to meaningful states and actions, this weak supervision im…
Inducing Process Supervision from Outcome-Only Reinforcement Learning
Renmin University of China（中国人民大学）；Tsinghua University（清华大学）；University of Electronic Science and Technology of China（电子科技大学）；Shanghai Jiao Tong University（上海交通大学）
Hasso Plattner Institute, University of Potsdam（波茨坦大学哈索·普拉特纳学院）
The challenging field of scene text detection requires complex data annotation, which is time-consuming and expensive. Techniques, such as weak supervision, can reduce the amount of data needed. In this paper we propose a weak supervision method for scene text detection, which makes use of reinforcement learning (RL). The reward received by the RL agent is estimated by a neural network, instead of being inferred from ground-truth labels. First, we enhance an existing supervised RL approach to text detection with several training optimizations, allowing us to close the performance gap to regres…
Good Reasoning Makes Good Demonstrations: Implicit Reasoning Quality Supervision via In-Context Reinforcement Learning
School of Data Science, Fudan University（复旦大学数据科学学院）；University of Chinese Academy of Sciences（中国科学院大学）；College of Intelligence and Computing, Tianjin University（天津大学智能与计算学院）
UC Berkeley（加州大学伯克利分校）；University of California, Berkeley（加州大学伯克利分校）
Meta-reinforcement learning (RL) methods can meta-train policies that adapt to new tasks with orders of magnitude less data than standard RL, but meta-training itself is costly and time-consuming. If we can meta-train on offline data, then we can reuse the same static dataset, labeled once with rewards for different tasks, to meta-train policies that adapt to a variety of new tasks at meta-test time. Although this capability would make meta-RL a practical tool for real-world use, offline meta-RL presents additional challenges beyond online meta-RL or standard offline RL settings. Meta-RL learn…
Training Large Language Models for Reasoning through Reverse Curriculum Reinforcement Learning
School of Computer Science, Fudan University（复旦大学计算机科学学院）；ByteDance Research（字节跳动研究部）；Institute of Modern Languages and Linguistics, Fudan University（复旦大学现代语言与语言学研究所）
Self-Harmony: Learning to Harmonize Self-Supervision and Self-Play in Test-Time Reinforcement Learning
The University of Tokyo（东京大学）；RIKEN Center for Advanced Intelligence Project（理化学研究所先进情报项目中心）；The Institute of Statistical Mathematics（统计数学研究所）；Google Research Australia（澳大利亚谷歌研究）
Technion - Israel Institute of Technology（以色列理工学院）
We propose a scheme for training a computerized agent to perform complex human tasks such as highway steering. The scheme is designed to follow a natural learning process whereby a human instructor teaches a computerized trainee. The learning process consists of five elements: (i) unsupervised feature learning; (ii) supervised imitation learning; (iii) supervised reward induction; (iv) supervised safety module construction; and (v) reinforcement learning. We implemented the last four elements of the scheme using deep convolutional networks and applied it to successfully create a computerized a…
Achieving robust performance is crucial when applying deep reinforcement learning (RL) in safety critical systems. Some of the state of the art approaches try to address the problem with adversarial agents, but these agents often require expert supervision to fine tune and prevent the adversary from becoming too challenging to the trainee agent. While other approaches involve automatically adjusting environment setups during training, they have been limited to simple environments where low-dimensional encodings can be used. Inspired by these approaches, we propose genetic curriculum, an algori…
Beyond Token-level Supervision: Unlocking the Potential of Decoding-based Regression via Reinforcement Learning
National Key Laboratory for Novel Software Technology, Nanjing University（新型软件技术国家实验室，南京大学）；School of Artificial Intelligence, Nanjing University（人工智能学院，南京大学）；School of Data Science, The Chinese University of Hong Kong, Shenzhen（数据科学学院，香港中文大学（深圳））；Department of Computer Science and Engineering, The Chinese University of Hong Kong（计算机科学与工程系，香港中文大学）
In imitation and reinforcement learning, the cost of human supervision limits the amount of data that robots can be trained on. An aspirational goal is to construct self-improving robots: robots that can learn and improve on their own, from autonomous interaction with minimal human supervision or oversight. Such robots could collect and train on much larger datasets, and thus learn more robust and performant policies. While reinforcement learning offers a framework for such autonomous learning via trial-and-error, practical realizations end up requiring extensive human supervision for reward f…
State University of New York at Buffalo（纽约州立大学布法罗分校）；Tencent Inc.（腾讯公司）；Pennsylvania State University（宾夕法尼亚州立大学）
Today social media has become the primary source for news. Via social media platforms, fake news travel at unprecedented speeds, reach global audiences and put users and communities at great risk. Therefore, it is extremely important to detect fake news as early as possible. Recently, deep learning based approaches have shown improved performance in fake news detection. However, the training of such models requires a large amount of labeled data, but manual annotation is time-consuming and expensive. Moreover, due to the dynamic nature of news, annotated samples may become outdated quickly and…
Technical University of Darmstadt（达姆施塔特工业大学）；Merck KGaA（默克集团）
Text style transfer has gained increasing attention from the research community over the recent years. However, the proposed approaches vary in many ways, which makes it hard to assess the individual contribution of the model components. In style transfer, the most important component is the optimization technique used to guide the learning in the absence of parallel training data. In this work we empirically compare the dominant optimization paradigms which provide supervision signals during training: backtranslation, adversarial training and reinforcement learning. We find that backtranslati…
MIT CSAIL（麻省理工学院计算机科学与人工智能实验室）；Chung Shan Medical University（中山医学大学）
Clinical finding summaries from an orthopantomogram, or a dental panoramic radiograph, have significant potential to improve patient communication and speed up clinical judgments. While orthopantomogram is a first-line tool for dental examinations, no existing work has explored the summarization of findings from it. A finding summary has to find teeth in the imaging study and label the teeth with several types of past treatments. To tackle the problem, we developDeepOPG that breaks the summarization process into functional segmentation and tooth localization, the latter of which is further ref…
What If Consensus Lies? Selective-Complementary Reinforcement Learning at Test Time
School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）；NLPR & MAIS, Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；Nanjing University（南京大学）
From Supervision to Exploration: What Does Protein Language Model Learn During Reinforcement Learning?
The Chinese University of Hong Kong（香港中文大学）；Peking University（北京大学）；Stanford University（斯坦福大学）；Nanjing University（南京大学）；University of Pennsylvania（宾夕法尼亚大学）；National University of Singapore（新加坡国立大学）；University of Illinois Urbana-Champaign（伊利诺伊大学香槟分校）；Toyota Technological Institute at Chicago（芝加哥丰田技术研究所）
University of Maryland, Baltimore County（马里兰大学巴尔的摩县分校）；US Army Research Lab（美国陆军研究实验室）
Learning to solve long horizon temporally extended tasks with reinforcement learning has been a challenge for several years now. We believe that it is important to leverage both the hierarchical structure of complex tasks and to use expert supervision whenever possible to solve such tasks. This work introduces an interpretable hierarchical agent framework by combining planning and semantic goal directed reinforcement learning. We assume access to certain spatial and haptic predicates and construct a simple and powerful semantic goal space. These semantic goal representations are more interpret…
The models of n-ary cross sentence relation extraction based on distant supervision assume that consecutive sentences mentioning n entities describe the relation of these n entities. However, on one hand, this assumption introduces noisy labeled data and harms the models' performance. On the other hand, some non-consecutive sentences also describe one relation and these sentences cannot be labeled under this assumption. In this paper, we relax this strong assumption by a weaker distant supervision assumption to address the second issue and propose a novel sentence distribution estimator model …
Zhejiang University（浙江大学）；University of Southern California（南加州大学）；Hikvision（海康威视）
This paper provides a new way to improve the efficiency of the REINFORCE training process. We apply it to the task of instance selection in distant supervision. Modeling the instance selection in one bag as a sequential decision process, a reinforcement learning agent is trained to determine whether an instance is valuable or not and construct a new bag with less noisy instances. However unbiased methods, such as REINFORCE, could usually take much time to train. This paper adopts posterior regularization (PR) to integrate some domain-specific rules in instance selection using REINFORCE. As the…
The goal of inverse reinforcement learning (IRL) is to infer a reward function that explains the behavior of an agent performing a task. The assumption that most approaches make is that the demonstrated behavior is near-optimal. In many real-world scenarios, however, examples of truly optimal behavior are scarce, and it is desirable to effectively leverage sets of demonstrations of suboptimal or heterogeneous performance, which are easier to obtain. We propose an algorithm that learns a reward function from such demonstrations together with a weak supervision signal in the form of a distributi…
University of Toronto（多伦多大学）；Vector Institute（矢量研究所）；Northwestern University（西北大学）；UC Santa Cruz（加州大学圣克鲁兹分校）
Most existing policy learning solutions require the learning agents to receive high-quality supervision signals such as well-designed rewards in reinforcement learning (RL) or high-quality expert demonstrations in behavioral cloning (BC). These quality supervisions are usually infeasible or prohibitively expensive to obtain in practice. We aim for a unified framework that leverages the available cheap weak supervisions to perform policy learning efficiently. To handle this problem, we treat the "weak supervision" as imperfect information coming from a peer agent, and evaluate the learning agen…
