IRL for Restless Multi-Armed Bandits with Applications in Maternal and Child Health

Public health practitioners often have the goal of monitoring patients and maximizing patients' time spent in "favorable" or healthy states while being constrained to using limited resources. Restless multi-armed bandits (RMAB) are an effective model to solve this problem as they are helpful to allocate limited resources among many agents under resource constraints, where patients behave differently depending on whether they are intervened on or not. However, RMABs assume the reward function is known. This is unrealistic in many public health settings because patients face unique challenges and it is impossible for a human to know who is most deserving of any intervention at such a large scale. To address this shortcoming, this paper is the first to present the use of inverse reinforcement learning (IRL) to learn desired rewards for RMABs, and we demonstrate improved outcomes in a maternal and child health telehealth program. First we allow public health experts to specify their goals at an aggregate or population level and propose an algorithm to design expert trajectories at scale based on those goals. Second, our algorithm WHIRL uses gradient updates to optimize the objective, allowing for efficient and accurate learning of RMAB rewards. Third, we compare with existing baselines and outperform those in terms of run-time and accuracy. Finally, we evaluate and show the usefulness of WHIRL on thousands on beneficiaries from a real-world maternal and child health setting in India. We publicly release our code here: https://github.com/Gjain234/WHIRL.

Semantic loss guided data efficient supervised fine tuning for Safe Responses in LLMs

Large Language Models (LLMs) generating unsafe responses to toxic prompts is a significant issue in their applications. While various efforts aim to address this safety concern, previous approaches often demand substantial human data collection or rely on the less dependable option of using another LLM to generate corrective data. In this paper, we aim to take this problem and overcome limitations of requiring significant high-quality human data. Our method requires only a small set of unsafe responses to toxic prompts, easily obtained from the unsafe LLM itself. By employing a semantic cost combined with a negative Earth Mover Distance (EMD) loss, we guide the LLM away from generating unsafe responses. Additionally, we propose a novel lower bound for EMD loss, enabling more efficient optimization. Our results demonstrate superior performance and data efficiency compared to baselines, and we further examine the nuanced effects of over-alignment and potential degradation of language capabilities when using contrastive data.

UNIQ: Offline Inverse Q-learning for Avoiding Undesirable Demonstrations

We address the problem of offline learning a policy that avoids undesirable demonstrations. Unlike conventional offline imitation learning approaches that aim to imitate expert or near-optimal demonstrations, our setting involves avoiding undesirable behavior (specified using undesirable demonstrations). To tackle this problem, unlike standard imitation learning where the aim is to minimize the distance between learning policy and expert demonstrations, we formulate the learning task as maximizing a statistical distance, in the space of state-action stationary distributions, between the learning policy and the undesirable policy. This significantly different approach results in a novel training objective that necessitates a new algorithm to address it. Our algorithm, UNIQ, tackles these challenges by building on the inverse Q-learning framework, framing the learning problem as a cooperative (non-adversarial) task. We then demonstrate how to efficiently leverage unlabeled data for practical training. Our method is evaluated on standard benchmark environments, where it consistently outperforms state-of-the-art baselines. The code implementation can be accessed at: https://github.com/hmhuy0/UNIQ.

Towards Neural Network based Cognitive Models of Dynamic Decision-Making by Humans

Modelling human cognitive processes in dynamic decision-making tasks has been an endeavor in AI for a long time. Some initial works have attempted to utilize neural networks (and large language models) but often assume one common model for all humans and aim to emulate human behavior in aggregate. However, behavior of each human is distinct, heterogeneous and relies on specific past experiences in specific tasks. To that end, we build on a well known model of cognition, namely Instance Based Learning (IBL), that posits that decisions are made based on similar situations encountered in the past. We propose two new attention based neural network models to model human decision-making in dynamic settings. We experiment with two distinct datasets gathered from human subject experiment data, one focusing on detection of phishing email by humans and another where humans act as attackers in a cybersecurity setting and decide on an attack option. We conduct extensive experiments with our two neural network models, IBL, and GPT3.5, and demonstrate that one of our neural network models achieves the best performance in representing human decision-making. We find an interesting trend that all models predict a human's decision better if that human is better at the task. We also explore explanation of human decisions based on what our model considers important in prediction. Overall, our work yields promising results for further use of neural networks in cognitive modelling of human decision making. Our code is available at https://github.com/shshnkreddy/NCM-HDM.

Preserving the Privacy of Reward Functions in MDPs through Deception

Preserving the privacy of preferences (or rewards) of a sequential decision-making agent when decisions are observable is crucial in many physical and cybersecurity domains. For instance, in wildlife monitoring, agents must allocate patrolling resources without revealing animal locations to poachers. This paper addresses privacy preservation in planning over a sequence of actions in MDPs, where the reward function represents the preference structure to be protected. Observers can use Inverse RL (IRL) to learn these preferences, making this a challenging task. Current research on differential privacy in reward functions fails to ensure guarantee on the minimum expected reward and offers theoretical guarantees that are inadequate against IRL-based observers. To bridge this gap, we propose a novel approach rooted in the theory of deception. Deception includes two models: dissimulation (hiding the truth) and simulation (showing the wrong). Our first contribution theoretically demonstrates significant privacy leaks in existing dissimulation-based methods. Our second contribution is a novel RL-based planning algorithm that uses simulation to effectively address these privacy concerns while ensuring a guarantee on the expected reward. Experiments on multiple benchmark problems show that our approach outperforms previous methods in preserving reward function privacy.

Safety through feedback in Constrained RL

In safety-critical RL settings, the inclusion of an additional cost function is often favoured over the arduous task of modifying the reward function to ensure the agent's safe behaviour. However, designing or evaluating such a cost function can be prohibitively expensive. For instance, in the domain of self-driving, designing a cost function that encompasses all unsafe behaviours (e.g. aggressive lane changes) is inherently complex. In such scenarios, the cost function can be learned from feedback collected offline in between training rounds. This feedback can be system generated or elicited from a human observing the training process. Previous approaches have not been able to scale to complex environments and are constrained to receiving feedback at the state level which can be expensive to collect. To this end, we introduce an approach that scales to more complex domains and extends to beyond state-level feedback, thus, reducing the burden on the evaluator. Inferring the cost function in such settings poses challenges, particularly in assigning credit to individual states based on trajectory-level feedback. To address this, we propose a surrogate objective that transforms the problem into a state-level supervised classification task with noisy labels, which can be solved efficiently. Additionally, it is often infeasible to collect feedback on every trajectory generated by the agent, hence, two fundamental questions arise: (1) Which trajectories should be presented to the human? and (2) How many trajectories are necessary for effective learning? To address these questions, we introduce \textit{novelty-based sampling} that selectively involves the evaluator only when the the agent encounters a \textit{novel} trajectory. We showcase the efficiency of our method through experimentation on several benchmark Safety Gymnasium environments and realistic self-driving scenarios.

EduQate: Generating Adaptive Curricula through RMABs in Education Settings

There has been significant interest in the development of personalized and adaptive educational tools that cater to a student's individual learning progress. A crucial aspect in developing such tools is in exploring how mastery can be achieved across a diverse yet related range of content in an efficient manner. While Reinforcement Learning and Multi-armed Bandits have shown promise in educational settings, existing works often assume the independence of learning content, neglecting the prevalent interdependencies between such content. In response, we introduce Education Network Restless Multi-armed Bandits (EdNetRMABs), utilizing a network to represent the relationships between interdependent arms. Subsequently, we propose EduQate, a method employing interdependency-aware Q-learning to make informed decisions on arm selection at each time step. We establish the optimality guarantee of EduQate and demonstrate its efficacy compared to baseline policies, using students modeled from both synthetic and real-world data.

Unlocking Large Language Model's Planning Capabilities with Maximum Diversity Fine-tuning

Large language models (LLMs) have demonstrated impressive task-solving capabilities, achieved through either prompting techniques or system designs. However, concerns have arisen regarding their proficiency in planning tasks, as they often struggle to generate valid plans. This paper investigates the impact of fine-tuning on LLMs' planning capabilities. Our findings indicate that LLMs can achieve good performance in planning through substantial (thousands of specific examples) fine-tuning. However, fine-tuning is associated with significant economic and computational costs. To address this challenge, we propose the Maximum Diversity Fine-Tuning (MDFT) strategy to improve the sample efficiency of fine-tuning in the planning domain. Specifically, our algorithm, referred to as MDFT-g, encodes the planning task instances with their graph representations and selects a subset of samples in the vector space that maximizes data diversity. We empirically demonstrate that MDFT-g consistently outperforms existing baselines at various scales across multiple benchmark domains.

Bootstrapping Language Models with DPO Implicit Rewards

Human alignment in large language models (LLMs) is an active area of research. A recent groundbreaking work, direct preference optimization (DPO), has greatly simplified the process from past work in reinforcement learning from human feedback (RLHF) by bypassing the reward learning stage in RLHF. DPO, after training, provides an implicit reward model. In this work, we make a novel observation that this implicit reward model can by itself be used in a bootstrapping fashion to further align the LLM. Our approach is to use the rewards from a current LLM model to construct a preference dataset, which is then used in subsequent DPO rounds. We incorporate refinements that debias the length of the responses and improve the quality of the preference dataset to further improve our approach. Our approach, named self-alignment with DPO ImpliCit rEwards (DICE), shows great improvements in alignment and achieves superior performance than Gemini Pro on AlpacaEval 2, reaching 27.55% length-controlled win rate against GPT-4 Turbo, but with only 8B parameters and no external feedback. Our code is available at https://github.com/sail-sg/dice.

Probabilistic Perspectives on Error Minimization in Adversarial Reinforcement Learning

Deep Reinforcement Learning (DRL) policies are critically vulnerable to adversarial noise in observations, posing severe risks in safety-critical scenarios. For example, a self-driving car receiving manipulated sensory inputs about traffic signs could lead to catastrophic outcomes. Existing strategies to fortify RL algorithms against such adversarial perturbations generally fall into two categories: (a) using regularization methods that enhance robustness by incorporating adversarial loss terms into the value objectives, and (b) adopting "maximin" principles, which focus on maximizing the minimum value to ensure robustness. While regularization methods reduce the likelihood of successful attacks, their effectiveness drops significantly if an attack does succeed. On the other hand, maximin objectives, although robust, tend to be overly conservative. To address this challenge, we introduce a novel objective called Adversarial Counterfactual Error (ACoE), which naturally balances optimizing value and robustness against adversarial attacks. To optimize ACoE in a scalable manner in model-free settings, we propose a theoretically justified surrogate objective known as Cumulative-ACoE (C-ACoE). The core idea of optimizing C-ACoE is utilizing the belief about the underlying true state given the adversarially perturbed observation. Our empirical evaluations demonstrate that our method outperforms current state-of-the-art approaches for addressing adversarial RL problems across all established benchmarks (MuJoCo, Atari, and Highway) used in the literature.