Mixed-precision quantization mostly predetermines the model bit-width settings before actual training due to the non-differential bit-width sampling process, obtaining sub-optimal performance. Worse still, the conventional static quality-consistent training setting, i.e., all data is assumed to be of the same quality across training and inference, overlooks data quality changes in real-world applications which may lead to poor robustness of the quantized models. In this paper, we propose a novel Data Quality-aware Mixed-precision Quantization framework, dubbed DQMQ, to dynamically adapt quantization bit-widths to different data qualities. The adaption is based on a bit-width decision policy that can be learned jointly with the quantization training. Concretely, DQMQ is modeled as a hybrid reinforcement learning (RL) task that combines model-based policy optimization with supervised quantization training. By relaxing the discrete bit-width sampling to a continuous probability distribution that is encoded with few learnable parameters, DQMQ is differentiable and can be directly optimized end-to-end with a hybrid optimization target considering both task performance and quantization benefits. Trained on mixed-quality image datasets, DQMQ can implicitly select the most proper bit-width for each layer when facing uneven input qualities. Extensive experiments on various benchmark datasets and networks demonstrate the superiority of DQMQ against existing fixed/mixed-precision quantization methods.

Why should moral philosophers, moral psychologists, and machine ethicists care about computational complexity? Debates on whether artificial intelligence (AI) can or should be used to solve problems in ethical domains have mainly been driven by what AI can or cannot do in terms of human capacities. In this paper, we tackle the problem from the other end by exploring what kind of moral machines are possible based on what computational systems can or cannot do. To do so, we analyze normative ethics through the lens of computational complexity. First, we introduce computational complexity for the uninitiated reader and discuss how the complexity of ethical problems can be framed within Marr's three levels of analysis. We then study a range of ethical problems based on consequentialism, deontology, and virtue ethics, with the aim of elucidating the complexity associated with the problems themselves (e.g., due to combinatorics, uncertainty, strategic dynamics), the computational methods employed (e.g., probability, logic, learning), and the available resources (e.g., time, knowledge, learning). The results indicate that most problems the normative frameworks pose lead to tractability issues in every category analyzed. Our investigation also provides several insights about the computational nature of normative ethics, including the differences between rule- and outcome-based moral strategies, and the implementation-variance with regard to moral resources. We then discuss the consequences complexity results have for the prospect of moral machines in virtue of the trade-off between optimality and efficiency. Finally, we elucidate how computational complexity can be used to inform both philosophical and cognitive-psychological research on human morality by advancing the Moral Tractability Thesis (MTT).

A practical challenge in reinforcement learning are combinatorial action spaces that make planning computationally demanding. For example, in cooperative multi-agent reinforcement learning, a potentially large number of agents jointly optimize a global reward function, which leads to a combinatorial blow-up in the action space by the number of agents. As a minimal requirement, we assume access to an argmax oracle that allows to efficiently compute the greedy policy for any Q-function in the model class. Building on recent work in planning with local access to a simulator and linear function approximation, we propose efficient algorithms for this setting that lead to polynomial compute and query complexity in all relevant problem parameters. For the special case where the feature decomposition is additive, we further improve the bounds and extend the results to the kernelized setting with an efficient algorithm.

State of the art reinforcement learning has enabled training agents on tasks of ever increasing complexity. However, the current paradigm tends to favor training agents from scratch on every new task or on collections of tasks with a view towards generalizing to novel task configurations. The former suffers from poor data efficiency while the latter is difficult when test tasks are out-of-distribution. Agents that can effectively transfer their knowledge about the world pose a potential solution to these issues. In this paper, we investigate transfer learning in the context of model-based agents. Specifically, we aim to understand when exactly environment models have an advantage and why. We find that a model-based approach outperforms controlled model-free baselines for transfer learning. Through ablations, we show that both the policy and dynamics model learnt through exploration matter for successful transfer. We demonstrate our results across three domains which vary in their requirements for transfer: in-distribution procedural (Crafter), in-distribution identical (RoboDesk), and out-of-distribution (Meta-World). Our results show that intrinsic exploration combined with environment models present a viable direction towards agents that are self-supervised and able to generalize to novel reward functions.

Recent progress in end-to-end Imitation Learning approaches has shown promising results and generalization capabilities on mobile manipulation tasks. Such models are seeing increasing deployment in real-world settings, where scaling up requires robots to be able to operate with high autonomy, i.e. requiring as little human supervision as possible. In order to avoid the need for one-on-one human supervision, robots need to be able to detect and prevent policy failures ahead of time, and ask for help, allowing a remote operator to supervise multiple robots and help when needed. However, the black-box nature of end-to-end Imitation Learning models such as Behavioral Cloning, as well as the lack of an explicit state-value representation, make it difficult to predict failures. To this end, we introduce Behavioral Cloning Value Approximation (BCVA), an approach to learning a state value function based on and trained jointly with a Behavioral Cloning policy that can be used to predict failures. We demonstrate the effectiveness of BCVA by applying it to the challenging mobile manipulation task of latched-door opening, showing that we can identify failure scenarios with with 86% precision and 81% recall, evaluated on over 2000 real world runs, improving upon the baseline of simple failure classification by 10 percentage-points.

Abstract-- Designing appropriate reward functions for Reinforcement Learning (RL) approaches has been a significant problem, especially for complex environments such as Atari games. Utilizing natural language instructions to provide intermediate rewards to RL agents in a process known as reward shaping can help the agent in reaching the goal state faster. In this work, we propose a natural language-based reward shaping approach that maps trajectories from the Montezuma's Revenge game environment to corresponding natural language instructions using an extension of the LanguagE-Action Reward Network (LEARN) framework. These trajectory-language mappings are further used to generate intermediate rewards which are integrated into reward functions that can be utilized to learn an optimal policy for any standard RL algorithms. For a set of 15 tasks from Atari's Montezuma's Revenge game, the Ext-LEARN approach leads to the successful completion of tasks more often on average than the reward shaping approach that uses the LEARN framework and performs even better than the reward shaping framework without natural languagebased rewards. Reinforcement Learning (RL) has been used extensively in games with promising results in terms of performance. RL algorithms utilize the concepts of rewards (positive or negative reinforcements) to help agents in learning to optimize their behaviors in an environment with which they interact. Subsequently, RL has been successfully implemented in complex environments such as Atari games [1]. However, defining reward functions has been the most significant challenge in RL, especially while scaling RL algorithms for real-world applications with large state spaces.

In many applications of Reinforcement Learning (RL), it is critically important that the algorithm performs safely, such that instantaneous hard constraints are satisfied at each step, and unsafe states and actions are avoided. However, existing algorithms for ''safe'' RL are often designed under constraints that either require expected cumulative costs to be bounded or assume all states are safe. Thus, such algorithms could violate instantaneous hard constraints and traverse unsafe states (and actions) in practice. Therefore, in this paper, we develop the first near-optimal safe RL algorithm for episodic Markov Decision Processes with unsafe states and actions under instantaneous hard constraints and the linear mixture model. It not only achieves a regret $\tilde{O}(\frac{d H^3 \sqrt{dK}}{\Delta_c})$ that tightly matches the state-of-the-art regret in the setting with only unsafe actions and nearly matches that in the unconstrained setting, but is also safe at each step, where $d$ is the feature-mapping dimension, $K$ is the number of episodes, $H$ is the number of steps in each episode, and $\Delta_c$ is a safety-related parameter. We also provide a lower bound $\tilde{\Omega}(\max\{dH \sqrt{K}, \frac{H}{\Delta_c^2}\})$, which indicates that the dependency on $\Delta_c$ is necessary. Further, both our algorithm design and regret analysis involve several novel ideas, which may be of independent interest.

In offline reinforcement learning (RL), one detrimental issue to policy learning is the error accumulation of deep Q function in out-of-distribution (OOD) areas. Unfortunately, existing offline RL methods are often over-conservative, inevitably hurting generalization performance outside data distribution. In our study, one interesting observation is that deep Q functions approximate well inside the convex hull of training data. Inspired by this, we propose a new method, DOGE (Distance-sensitive Offline RL with better GEneralization). DOGE marries dataset geometry with deep function approximators in offline RL, and enables exploitation in generalizable OOD areas rather than strictly constraining policy within data distribution. Specifically, DOGE trains a state-conditioned distance function that can be readily plugged into standard actor-critic methods as a policy constraint. Simple yet elegant, our algorithm enjoys better generalization compared to state-of-the-art methods on D4RL benchmarks. Theoretical analysis demonstrates the superiority of our approach to existing methods that are solely based on data distribution or support constraints. Offline reinforcement learning (RL) provides a new possibility to learn optimized policies from large, pre-collected datasets without any environment interaction (Levine et al., 2020). This holds great promise to solve many real-world problems when online interaction is costly or dangerous yet historical data is easily accessible (Zhan et al., 2022). However, the optimization nature of RL, as well as the need for counterfactual reasoning on unseen data under offline setting, have caused great technical challenges for designing effective offline RL algorithms. Evaluating value function outside data coverage areas can produce falsely optimistic values; without corrective information from online interaction, such estimation errors can accumulate quickly and misguide policy learning process (Van Hasselt et al., 2018; Fujimoto et al., 2018; Kumar et al., 2019). Recent model-free offline RL methods investigate this error accumulation challenge in several ways: 1) Policy Constraint: directly constraining learned policy to stay inside distribution, or with the support of dataset (Kumar et al., 2019); 2) Value Regularization: regularizing value function to assign low values at out-of-distribution (OOD) actions (Kumar et al., 2020b); 3) In-sample Learning: learning value function within data samples (Kostrikov et al., 2021b) or simply treating it as the value function of behavioral policy (Brandfonbrener et al., 2021). All three schools of methods share similar traits of being conservative and omitting evaluation on OOD data, which brings benefits of minimizing model exploitation error, but at the expense of poor generalization of learned policy in OOD regions.

Predicting the behaviors of other road users is crucial to safe and intelligent decision-making for autonomous vehicles (AVs). However, most motion prediction models ignore the influence of the AV's actions and the planning module has to treat other agents as unalterable moving obstacles. To address this problem, this paper proposes an interaction-aware motion prediction model that is able to predict other agents' future trajectories according to the ego agent's future plan, i.e., their reactions to the ego's actions. Specifically, we employ Transformers to effectively encode the driving scene and incorporate the AV's plan in decoding the predicted trajectories. To train the model to accurately predict the reactions of other agents, we develop an online learning framework, where the ego agent explores the environment and collects other agents' reactions to itself. We validate the decision-making and learning framework in three highly interactive simulated driving scenarios. The results reveal that our decision-making method significantly outperforms the reinforcement learning methods in terms of data efficiency and performance. We also find that using the interaction-aware model can bring better performance than the non-interaction-aware model and the exploration process helps improve the success rate in testing.

Simulation based learning often provides a cost-efficient recourse to reinforcement learning applications in robotics. However, simulators are generally incapable of accurately replicating real-world dynamics, and thus bridging the sim2real gap is an important problem in simulation based learning. Current solutions to bridge the sim2real gap involve hybrid simulators that are augmented with neural residual models. Unfortunately, they require a separate residual model for each individual environment configuration (i.e., a fixed setting of environment variables such as mass, friction etc.), and thus are not transferable to new environments quickly. To address this issue, we propose a Reverse Action Transformation (RAT) policy which learns to imitate simulated policies in the real-world. Once learnt from a single environment, RAT can then be deployed on top of a Universal Policy Network to achieve zero-shot adaptation to new environments. We empirically evaluate our approach in a set of continuous control tasks and observe its advantage as a few-shot and zero-shot learner over competing baselines.