Deep Reinforcement Learning (RL) techniques can benefit greatly from leveraging prior experience, which can be either self-generated or acquired from other entities. Action advising is a framework that provides a flexible way to transfer such knowledge in the form of actions between teacher-student peers. However, due to the realistic concerns, the number of these interactions is limited with a budget; therefore, it is crucial to perform these in the most appropriate moments. There have been several promising studies recently that address this problem setting especially from the student's perspective. Despite their success, they have some shortcomings when it comes to the practical applicability and integrity as an overall solution to the learning from advice challenge. In this paper, we extend the idea of advice reusing via teacher imitation to construct a unified approach that addresses both advice collection and advice utilisation problems. We also propose a method to automatically tune the relevant hyperparameters of these components on-the-fly to make it able to adapt to any task with minimal human intervention. The experiments we performed in 5 different Atari games verify that our algorithm either surpasses or performs on-par with its top competitors while being far simpler to be employed. Furthermore, its individual components are also found to be providing significant advantages alone.
Recent work has explored the problem of autonomous navigation by imitating a teacher and learning an end-to-end policy, which directly predicts controls from raw images. However, these approaches tend to be sensitive to mistakes by the teacher and do not scale well to other environments or vehicles. To this end, we propose Observational Imitation Learning (OIL), a novel imitation learning variant that supports online training and automatic selection of optimal behavior by observing multiple imperfect teachers. We apply our proposed methodology to the challenging problems of autonomous driving …
Action advising is a peer-to-peer knowledge exchange technique built on the teacher-student paradigm to alleviate the sample inefficiency problem in deep reinforcement learning. Recently proposed student-initiated approaches have obtained promising results. However, due to being in the early stages of development, these also have some substantial shortcomings. One of the abilities that are absent in the current methods is further utilising advice by reusing, which is especially crucial in the practical settings considering the budget and cost constraints in peer-to-peer. In this study, we pres…
Deep reinforcement learning (DRL) policies are vulnerable to unauthorized replication attacks, where an adversary exploits imitation learning to reproduce target policies from observed behavior. In this paper, we propose Constrained Randomization of Policy (CRoP) as a mitigation technique against such attacks. CRoP induces the execution of sub-optimal actions at random under performance loss constraints. We present a parametric analysis of CRoP, address the optimality of CRoP, and establish theoretical bounds on the adversarial budget and the expectation of loss. Furthermore, we report the exp…
Reinforcement learning (RL) makes it possible to train agents capable of achieving sophisticated goals in complex and uncertain environments. A key difficulty in reinforcement learning is specifying a reward function for the agent to optimize. Traditionally, imitation learning in RL has been used to overcome this problem. Unfortunately, hitherto imitation learning methods tend to require that demonstrations are supplied in the first-person: the agent is provided with a sequence of states and a specification of the actions that it should have taken. While powerful, this kind of imitation learni…
In this work we describe a novel deep reinforcement learning architecture that allows multiple actions to be selected at every time-step in an efficient manner. Multi-action policies allow complex behaviours to be learnt that would otherwise be hard to achieve when using single action selection techniques. We use both imitation learning and temporal difference (TD) reinforcement learning (RL) to provide a 4x improvement in training time and 2.5x improvement in performance over single action selection TD RL. We demonstrate the capabilities of this network using a complex in-house 3D game. Mimic…
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…
University of Texas at Austin（德克萨斯大学奥斯汀分校）；University of Florida（佛罗里达大学）
Kickstarting deep reinforcement learning algorithms facilitate a teacher-student relationship among the agents and allow for a well-performing teacher to share demonstrations with a student to expedite the student's training. However, despite the known benefits, the demonstrations may contain sensitive information about the teacher's training data and existing kickstarting methods do not take any measures to protect it. Therefore, we use the framework of differential privacy to develop a mechanism that securely shares the teacher's demonstrations with the student. The mechanism allows for the …
In the search for more sample-efficient reinforcement-learning (RL) algorithms, a promising direction is to leverage as much external off-policy data as possible. For instance, expert demonstrations. In the past, multiple ideas have been proposed to make good use of the demonstrations added to the replay buffer, such as pretraining on demonstrations only or minimizing additional cost functions. We present a new method, able to leverage both demonstrations and episodes collected online in any sparse-reward environment with any off-policy algorithm. Our method is based on a reward bonus given to…
University of Tokyo（东京大学）；Technische Universität Darmstadt（达姆施塔特工业大学）；University of Lincoln（林肯大学）；Carnegie Mellon University（卡内基梅隆大学）；University of California, Berkeley（加州大学伯克利分校）
As robots and other intelligent agents move from simple environments and problems to more complex, unstructured settings, manually programming their behavior has become increasingly challenging and expensive. Often, it is easier for a teacher to demonstrate a desired behavior rather than attempt to manually engineer it. This process of learning from demonstrations, and the study of algorithms to do so, is called imitation learning. This work provides an introduction to imitation learning. It covers the underlying assumptions, approaches, and how they relate; the rich set of algorithms develope…
Reinforcement learning (RL) has been shown to be effective at learning control from experience. However, RL typically requires a large amount of online interaction with the environment. This limits its applicability to real-world settings, such as in robotics, where such interaction is expensive. In this work we investigate ways to minimize online interactions in a target task, by reusing a suboptimal policy we might have access to, for example from training on related prior tasks, or in simulation. To this end, we develop two RL algorithms that can speed up training by using not only the acti…
Distilled Reinforcement Learning for LLM Post-training
College of Elite Engineers, Nankai University（南开大学精英工程师学院）；Zhongguancun Academy（中关村学院）；Beijing Institute of Technology（北京理工大学）；Zhejiang University（浙江大学）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；Harbin Institute of Technology（哈尔滨工业大学）；College of Software, Nankai University（南开大学软件学院）
Open-MOPD: Diagnosing and Fixing Capability Imbalance in Multi-Teacher On-Policy Distillation
SIA-Lab of Tsinghua AIR and ByteDance Seed（清华大学AIR与字节跳动种子SIA实验室）；Institute for AI Industry Research (AIR), Tsinghua University（清华大学人工智能产业研究院）；Department of Computer Science and Technology, Tsinghua University（清华大学计算机科学与技术系）
Information Sciences Institute（信息科学研究所）；University of Southern California（南加州大学）
We consider problems of making sequences of decisions to accomplish tasks, interacting via the medium of language. These problems are often tackled with reinforcement learning approaches. We find that these models do not generalize well when applied to novel task domains. However, the large amount of computation necessary to adequately train and explore the search space of sequential decision making, under a reinforcement learning paradigm, precludes the inclusion of large contextualized language models, which might otherwise enable the desired generalization ability. We introduce a teacher-st…
University of Maryland, College Park（马里兰大学帕克分校）；Microsoft Research（微软研究院）
We formulate the problem of learning to imitate multiple, non-deterministic teachers with minimal interaction cost. Rather than learning a specific policy as in standard imitation learning, the goal in this problem is to learn a distribution over a policy space. We first present a general framework that efficiently models and estimates such a distribution by learning continuous representations of the teacher policies. Next, we develop Active Performance-Based Imitation Learning (APIL), an active learning algorithm for reducing the learner-teacher interaction cost in this framework. By making q…
Deep Reinforcement Learning (RL) algorithms can solve complex sequential decision tasks successfully. However, they have a major drawback of having poor sample efficiency which can often be tackled by knowledge reuse. In Multi-Agent Reinforcement Learning (MARL) this drawback becomes worse, but at the same time, a new set of opportunities to leverage knowledge are also presented through agent interactions. One promising approach among these is peer-to-peer action advising through a teacher-student framework. Despite being introduced for single-agent RL originally, recent studies show that it c…
A classical learning setting typically concerns an agent/student who collects data, or observations, from a system in order to estimate a certain property of interest. Correctional learning is a type of cooperative teacher-student framework where a teacher, who has partial knowledge about the system, has the ability to observe and alter (correct) the observations received by the student in order to improve the accuracy of its estimate. In this paper, we show how the variance of the estimate of the student can be reduced with the help of the teacher. We formulate the corresponding online proble…
Animation and machine learning research have shown great advancements in the past decade, leading to robust and powerful methods for learning complex physically-based animations. However, learning can take hours or days, especially if no reference movement data is available. In this paper, we propose and evaluate a novel combination of techniques for accelerating the learning of stable locomotion movements through self-imitation learning of synthetic animations. First, we produce synthetic and cyclic reference movement using a recent online tree search approach that can discover stable walking…
We study the problem of learning efficient algorithms that strongly generalize in the framework of neural program induction. By carefully designing the input / output interfaces of the neural model and through imitation, we are able to learn models that produce correct results for arbitrary input sizes, achieving strong generalization. Moreover, by using reinforcement learning, we optimize for program efficiency metrics, and discover new algorithms that surpass the teacher used in imitation. With this, our approach can learn to outperform custom-written solutions for a variety of problems, as …
Institute of Industrial Science, The University of Tokyo（东京大学工业科学研究所）；LINE Corporation（LINE公司）；Faculty of Engineering, Kyoto University of Advanced Science（京都先进科学大学工学部）
As the number of the robot's degrees of freedom increases, the implementation of robot motion becomes more complex and difficult. In this study, we focus on learning 6DOF-grasping motion and consider dividing the grasping motion into multiple tasks. We propose to combine imitation and reinforcement learning in order to facilitate a more efficient learning of the desired motion. In order to collect demonstration data as teacher data for the imitation learning, we created a virtual reality (VR) interface that allows humans to operate the robot intuitively. Moreover, by dividing the motion into s…
The University of Melbourne（墨尔本大学）；Monash University（莫纳什大学）；Sony AI（索尼人工智能）；ANU（澳大利亚国立大学）；Data61 CSIRO（联邦科学与工业研究组织数据61研究所）
Machine-learning-as-a-service (MLaaS) has attracted millions of users to their splendid large-scale models. Although published as black-box APIs, the valuable models behind these services are still vulnerable to imitation attacks. Recently, a series of works have demonstrated that attackers manage to steal or extract the victim models. Nonetheless, none of the previous stolen models can outperform the original black-box APIs. In this work, we conduct unsupervised domain adaptation and multi-victim ensemble to showing that attackers could potentially surpass victims, which is beyond previous un…
National University of Singapore（新加坡国立大学）；Huawei Noah’s Ark Lab（华为诺亚方舟实验室）
State-of-the-art CNN based recognition models are often computationally prohibitive to deploy on low-end devices. A promising high level approach tackling this limitation is knowledge distillation, which let small student model mimic cumbersome teacher model's output to get improved generalization. However, related methods mainly focus on simple task of classification while do not consider complex tasks like object detection. We show applying the vanilla knowledge distillation to detection model gets minor gain. To address the challenge of distilling knowledge in detection model, we propose a …
Speculative Rollback Correction for Quality-Diverse Web Agent Imitation
Beihang University（北京航空航天大学）；Institute of Software, Chinese Academy of Sciences（中国科学院软件研究所）；The Hong Kong University of Science and Technology（香港科技大学）；Northwestern Polytechnical University（西北工业大学）；Tsinghua University（清华大学）；The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））；Peking University（北京大学）
Rethinking Reverse KL as Adaptive Entropy Distillation
School of Computer Science and Engineering, Sun Yat-sen University（中山大学计算机科学与工程学院）；School of Science and Technology, Hong Kong Metropolitan University（香港都会大学科技学院）
