Despite the impressive feats demonstrated by Reinforcement Learning (RL), these algorithms have seen little adoption in high-risk, real-world applications due to current difficulties in explaining RL agent actions and building user trust. We present Counterfactual Demonstrations for Explanation (CODEX), a method that incorporates semantic clustering, which can effectively summarize RL agent behavior in the state-action space. Experimentation on the MiniGrid and StarCraft II gaming environments reveals the semantic clusters retain temporal as well as entity information, which is reflected in the constructed summary of agent behavior. Furthermore, clustering the discrete+continuous game-state latent representations identifies the most crucial episodic events, demonstrating a relationship between the latent and semantic spaces. This work contributes to the growing body of work that strives to unlock the power of RL for widespread use by leveraging and extending techniques from Natural Language Processing.

Calibrating agent-based models (ABMs) in economics and finance typically involves a derivative-free search in a very large parameter space. In this work, we benchmark a number of search methods in the calibration of a well-known macroeconomic ABM on real data, and further assess the performance of "mixed strategies" made by combining different methods. We find that methods based on random-forest surrogates are particularly efficient, and that combining search methods generally increases performance since the biases of any single method are mitigated. Moving from these observations, we propose a reinforcement learning (RL) scheme to automatically select and combine search methods on-the-fly during a calibration run. The RL agent keeps exploiting a specific method only as long as this keeps performing well, but explores new strategies when the specific method reaches a performance plateau. The resulting RL search scheme outperforms any other method or method combination tested, and does not rely on any prior information or trial and error procedure.

High-quality traffic flow generation is the core module in building simulators for autonomous driving. However, the majority of available simulators are incapable of replicating traffic patterns that accurately reflect the various features of real-world data while also simulating human-like reactive responses to the tested autopilot driving strategies. Taking one step forward to addressing such a problem, we propose Realistic Interactive TrAffic flow (RITA) as an integrated component of existing driving simulators to provide high-quality traffic flow for the evaluation and optimization of the tested driving strategies. RITA is developed with consideration of three key features, i.e., fidelity, diversity, and controllability, and consists of two core modules called RITABackend and RITAKit. RITABackend is built to support vehicle-wise control and provide traffic generation models from real-world datasets, while RITAKit is developed with easy-to-use interfaces for controllable traffic generation via RITABackend. We demonstrate RITA's capacity to create diversified and high-fidelity traffic simulations in several highly interactive highway scenarios. The experimental findings demonstrate that our produced RITA traffic flows exhibit all three key features, hence enhancing the completeness of driving strategy evaluation. Moreover, we showcase the possibility for further improvement of baseline strategies through online fine-tuning with RITA traffic flows.

Sim-to-real is a mainstream method to cope with the large number of trials needed by typical deep reinforcement learning methods. However, transferring a policy trained in simulation to actual hardware remains an open challenge due to the reality gap. In particular, the characteristics of actuators in legged robots have a considerable influence on sim-to-real transfer. There are two challenges: 1) High reduction ratio gears are widely used in actuators, and the reality gap issue becomes especially pronounced when backdrivability is considered in controlling joints compliantly. 2) The difficulty in achieving stable bipedal locomotion causes typical system identification methods to fail to sufficiently transfer the policy. For these two challenges, we propose 1) a new simulation model of gears and 2) a method for system identification that can utilize failed attempts. The method's effectiveness is verified using a biped robot, the ROBOTIS-OP3, and the sim-to-real transferred policy can stabilize the robot under severe disturbances and walk on uneven surfaces without using force and torque sensors.

To tackle long planning horizon problems in reinforcement learning with general function approximation, we propose the first algorithm, termed as UCRL-WVTR, that achieves both \emph{horizon-free} and \emph{instance-dependent}, since it eliminates the polynomial dependency on the planning horizon. The derived regret bound is deemed \emph{sharp}, as it matches the minimax lower bound when specialized to linear mixture MDPs up to logarithmic factors. Furthermore, UCRL-WVTR is \emph{computationally efficient} with access to a regression oracle. The achievement of such a horizon-free, instance-dependent, and sharp regret bound hinges upon (i) novel algorithm designs: weighted value-targeted regression and a high-order moment estimator in the context of general function approximation; and (ii) fine-grained analyses: a novel concentration bound of weighted non-linear least squares and a refined analysis which leads to the tight instance-dependent bound. We also conduct comprehensive experiments to corroborate our theoretical findings.

Machine learning (ML) methods generally assume the ready availability of the complete set of input features at deployment, typically incurring little to no cost. However, this assumption does not hold universally, especially in scenarios where feature acquisitions are associated with substantial costs. In contexts like medical diagnostics, the cost of acquiring certain features, such as X-rays, biopsies, etc. encompasses not only financial costs but also poses potential risks to patient well-being. In such cases, the cost or harm of the feature acquisition should be balanced against the predictive value of the feature. Active feature acquisition (AFA) addresses this problem by training two AI components: i) the "AFA agent," an AI system tasked with determining which features should be observed, and ii) an ML prediction model that undertakes the prediction task based on the acquired feature set. While missingness was effectively determined by, for example, a physician during the acquisition of the retrospective dataset, the missingness at the deployment of the AFA agent is determined by the AFA agent, thereby leading to a missingness distribution shift. In our companion paper [1], we formulate the problem of active feature acquisition performance evaluation (AFAPE) which addresses the task of estimating the performance an AFA agent would have at deployment, from the retrospective dataset. Consequently, upon completing the AFAPE problem, the physician will be well-informed about expected rates of incorrect diagnoses and the average costs associated with feature acquisitions when the AFA system is put into operation.

Machine learning methods often assume input features are available at no cost. However, in domains like healthcare, where acquiring features could be expensive or harmful, it is necessary to balance a feature's acquisition cost against its predictive value. The task of training an AI agent to decide which features to acquire is called active feature acquisition (AFA). By deploying an AFA agent, we effectively alter the acquisition strategy and trigger a distribution shift. To safely deploy AFA agents under this distribution shift, we present the problem of active feature acquisition performance evaluation (AFAPE). We examine AFAPE under i) a no direct effect (NDE) assumption, stating that acquisitions don't affect the underlying feature values; and ii) a no unobserved confounding (NUC) assumption, stating that retrospective feature acquisition decisions were only based on observed features. We show that one can apply offline reinforcement learning under the NUC assumption and missing data methods under the NDE assumption. When NUC and NDE hold, we propose a novel semi-offline reinforcement learning framework, which requires a weaker positivity assumption and yields more data-efficient estimators. We introduce three novel estimators: a direct method (DM), an inverse probability weighting (IPW), and a double reinforcement learning (DRL) estimator.

Classifier-free guidance is a key component for enhancing the performance of conditional generative models across diverse tasks. While it has previously demonstrated remarkable improvements for the sample quality, it has only been exclusively employed for diffusion models. In this paper, we integrate classifier-free guidance into Flow Matching (FM) models, an alternative simulation-free approach that trains Continuous Normalizing Flows (CNFs) based on regressing vector fields. We explore the usage of \emph{Guided Flows} for a variety of downstream applications. We show that Guided Flows significantly improves the sample quality in conditional image generation and zero-shot text-to-speech synthesis, boasting state-of-the-art performance. Notably, we are the first to apply flow models for plan generation in the offline reinforcement learning setting, showcasing a 10x speedup in computation compared to diffusion models while maintaining comparable performance.

Reinforcement Learning (RL) offers a versatile framework for achieving long-term goals. Its generality allows us to formalize a wide range of problems that real-world intelligent systems encounter, such as dealing with delayed rewards, handling partial observability, addressing the exploration and exploitation dilemma, utilizing offline data to improve online performance, and ensuring safety constraints are met. Despite considerable progress made by the RL research community in addressing these issues, existing open-source RL libraries tend to focus on a narrow portion of the RL solution pipeline, leaving other aspects largely unattended. This paper introduces Pearl, a Production-ready RL agent software package explicitly designed to embrace these challenges in a modular fashion. In addition to presenting preliminary benchmark results, this paper highlights Pearl's industry adoptions to demonstrate its readiness for production usage.

Offline reinforcement learning (RL) faces a significant challenge of distribution shift. Model-free offline RL penalizes the Q value for out-of-distribution (OOD) data or constrains the policy closed to the behavior policy to tackle this problem, but this inhibits the exploration of the OOD region. Model-based offline RL, which uses the trained environment model to generate more OOD data and performs conservative policy optimization within that model, has become an effective method for this problem. However, the current model-based algorithms rarely consider agent robustness when incorporating conservatism into policy. Therefore, the new model-based offline algorithm with a conservative Bellman operator (MICRO) is proposed. This method trades off performance and robustness via introducing the robust Bellman operator into the algorithm. Compared with previous model-based algorithms with robust adversarial models, MICRO can significantly reduce the computation cost by only choosing the minimal Q value in the state uncertainty set. Extensive experiments demonstrate that MICRO outperforms prior RL algorithms in offline RL benchmark and is considerably robust to adversarial perturbations.