In this interview series, we're meeting some of the AAAI/SIGAI Doctoral Consortium participants to find out more about their research. Tanmay Ambadkar is researching the reward structure in reinforcement learning, with the goal of providing generalizable solutions that can provide robust guarantees and are easily deployable. We caught up with Tanmay to find out more about his research, and in particular, the constrained reinforcement learning framework he has been working on. Tell us a bit about your PhD - where are you studying, and what is the topic of your research? I am a 4th year PhD candidate at The Pennsylvania State University, PA, USA.

Designing reward functions for efficiently guiding reinforcement learning (RL) agents toward specific behaviors is a complex task.This is challenging since it requires the identification of reward structures that are not sparse and that avoid inadvertently inducing undesirable behaviors. Naively modifying the reward structure to offer denser and more frequent feedback can lead to unintended outcomes and promote behaviors that are not aligned with the designer's intended goal. Although potential-based reward shaping is often suggested as a remedy, we systematically investigate settings where deploying it often significantly impairs performance. To address these issues, we introduce a new framework that uses a bi-level objective to learn \emph{behavior alignment reward functions}. These functions integrate auxiliary rewards reflecting a designer's heuristics and domain knowledge with the environment's primary rewards.

Since the 1990s, considerable empirical work has been carried out to train statistical models, such as neural networks (NNs), as learned heuristics for combinatorial optimization (CO) problems. When successful, such an approach eliminates the need for experts to design heuristics per problem type. Due to their structure, many hard CO problems are amenable to treatment through reinforcement learning (RL). Indeed, we find a wealth of literature training NNs using value-based, policy gradient, or actor-critic approaches, with promising results, both in terms of empirical optimality gaps and inference runtimes. Nevertheless, there has been a paucity of theoretical work undergirding the use of RL for CO problems. To this end, we introduce a unified framework to model CO problems through Markov decision processes (MDPs) and solve them using RL techniques. We provide easy-to-test assumptions under which CO problems can be formulated as equivalent undiscounted MDPs that provide optimal solutions to the original CO problems. Moreover, we establish conditions under which value-based RL techniques converge to approximate solutions of the CO problem with a guarantee on the associated optimality gap. Our convergence analysis provides: (1) a sufficient rate of increase in batch size and projected gradient descent steps at each RL iteration; (2) the resulting optimality gap in terms of problem parameters and targeted RL accuracy; and (3) the importance of a choice of state-space embedding. Together, our analysis illuminates the success (and limitations) of the celebrated deep Q-learning algorithm in this problem context.

We present a novel, regulator-driven approach for post hoc safety evaluation of learning-based, black-box autonomous mobile robots, ensuring ongoing compliance with evolving, human-defined safety rules. In our iterative workflow, human safety requirements are translated by regulators into Signal Temporal Logic (STL) specifications. Rollout traces from the black-box model are externally verified for compliance, yielding quantitative safety metrics, Total Robustness Value (TRV) and Largest Robustness Value (LRV), which measure average and worst-case specification adherence. These metrics inform targeted retraining and iterative improvement by model designers. We apply our method across two different applications: a virtual driving scenario and an autonomous mobile robot navigating a complex environment, and observe statistically significant improvements across both scenarios. In the virtual driving scenario, we see a 177% increase in traces adhering to the simulation speed limit, a 1138% increase in traces minimizing off-road driving, and a 16% increase in traces successfully reaching the goal within the time limit. In the autonomous navigation scenario, there is a 300% increase in traces avoiding sharp turns, a 200% increase in traces reaching the goal within the time limit, and a 49% increase in traces minimizing time spent near obstacles. Finally, we validate our approach on a TurtleBot3 robot in the real world, and demonstrate improved obstacle navigation with safety buffers.

Designing rewards for autonomous cyber attack and defense learning agents in a complex, dynamic environment is a challenging task for subject matter experts. We propose a large language model (LLM)-based reward design approach to generate autonomous cyber defense policies in a deep reinforcement learning (DRL)-driven experimental simulation environment. Multiple attack and defense agent personas were crafted, reflecting heterogeneity in agent actions, to generate LLM-guided reward designs where the LLM was first provided with contextual cyber simulation environment information. These reward structures were then utilized within a DRL-driven attack-defense simulation environment to learn an ensemble of cyber defense policies. Our results suggest that LLM-guided reward designs can lead to effective defense strategies against diverse adversarial behaviors.

The field of Offline Reinforcement Learning (RL) aims to derive effective policies from pre-collected datasets without active environment interaction. While traditional offline RL algorithms like Conservative Q-Learning (CQL) and Implicit Q-Learning (IQL) have shown promise, they often face challenges in balancing exploration and exploitation, especially in environments with varying reward densities. The recently proposed Decision Transformer (DT) approach, which reframes offline RL as a sequence modelling problem, has demonstrated impressive results across various benchmarks. This paper presents a comparative study evaluating the performance of DT against traditional offline RL algorithms in dense and sparse reward settings for the ANT con-tinous control environment. Our research investigates how these algorithms perform when faced with different reward structures, examining their ability to learn effective policies and generalize across varying levels of feedback. Through empirical analysis in the ANT environment, we found that DTs showed less sensitivity to varying reward density compared to other methods and particularly excelled with medium-expert datasets in sparse reward scenarios. In contrast, traditional value-based methods like IQL showed improved performance in dense reward settings with high-quality data, while CQL offered balanced performance across different data qualities. Additionally, DTs exhibited lower variance in performance but required significantly more computational resources compared to traditional approaches. These findings suggest that sequence modelling approaches may be more suitable for scenarios with uncertain reward structures or mixed-quality data, while value-based methods remain competitive in settings with dense rewards and high-quality demonstrations.

The action cost differs by row, where the corresponding optimal policy parameterization is highlighted in green.

Reward machines (RMs) inform reinforcement learning agents about the reward structure of the environment. This is particularly advantageous for complex non-Markovian tasks because agents with access to RMs can learn more efficiently from fewer samples. However, learning with RMs is ill-suited for long-horizon problems in which a set of subtasks can be executed in any order. In such cases, the amount of information to learn increases exponentially with the number of unordered subtasks. In this work, we address this limitation by introducing three generalisations of RMs: (1) Numeric RMs allow users to express complex tasks in a compact form. (2) In Agenda RMs, states are associated with an agenda that tracks the remaining subtasks to complete. (3) Coupled RMs have coupled states associated with each subtask in the agenda. Furthermore, we introduce a new compositional learning algorithm that leverages coupled RMs: Q-learning with coupled RMs (CoRM). Our experiments show that CoRM scales better than state-of-the-art RM algorithms for long-horizon problems with unordered subtasks.

Animals achieve energy-efficient locomotion by their implicit passive dynamics, a marvel that has captivated roboticists for decades.Recently, methods incorporated Adversarial Motion Prior (AMP) and Reinforcement learning (RL) shows promising progress to replicate Animals' naturalistic motion. However, such imitation learning approaches predominantly capture explicit kinematic patterns, so-called gaits, while overlooking the implicit passive dynamics. This work bridges this gap by incorporating a reward term guided by Impact Mitigation Factor (IMF), a physics-informed metric that quantifies a robot's ability to passively mitigate impacts. By integrating IMF with AMP, our approach enables RL policies to learn both explicit motion trajectories from animal reference motion and the implicit passive dynamic. We demonstrate energy efficiency improvements of up to 32%, as measured by the Cost of Transport (CoT), across both AMP and handcrafted reward structure.