Dana Hughes

Yue Guo, Joseph Campbell, Simon Stepputtis et al. · cmu

Action advising is a knowledge transfer technique for reinforcement learning based on the teacher-student paradigm. An expert teacher provides advice to a student during training in order to improve the student's sample efficiency and policy performance. Such advice is commonly given in the form of state-action pairs. However, it makes it difficult for the student to reason with and apply to novel states. We introduce Explainable Action Advising, in which the teacher provides action advice as well as associated explanations indicating why the action was chosen. This allows the student to self-reflect on what it has learned, enabling advice generalization and leading to improved sample efficiency and learning performance - even in environments where the teacher is sub-optimal. We empirically show that our framework is effective in both single-agent and multi-agent scenarios, yielding improved policy returns and convergence rates when compared to state-of-the-art methods

Renos Zabounidis, Joseph Campbell, Simon Stepputtis et al.

Multi-agent robotic systems are increasingly operating in real-world environments in close proximity to humans, yet are largely controlled by policy models with inscrutable deep neural network representations. We introduce a method for incorporating interpretable concepts from a domain expert into models trained through multi-agent reinforcement learning, by requiring the model to first predict such concepts then utilize them for decision making. This allows an expert to both reason about the resulting concept policy models in terms of these high-level concepts at run-time, as well as intervene and correct mispredictions to improve performance. We show that this yields improved interpretability and training stability, with benefits to policy performance and sample efficiency in a simulated and real-world cooperative-competitive multi-agent game.

Huao Li, Yu Quan Chong, Simon Stepputtis et al.

While Large Language Models (LLMs) have demonstrated impressive accomplishments in both reasoning and planning, their abilities in multi-agent collaborations remains largely unexplored. This study evaluates LLM-based agents in a multi-agent cooperative text game with Theory of Mind (ToM) inference tasks, comparing their performance with Multi-Agent Reinforcement Learning (MARL) and planning-based baselines. We observed evidence of emergent collaborative behaviors and high-order Theory of Mind capabilities among LLM-based agents. Our results reveal limitations in LLM-based agents' planning optimization due to systematic failures in managing long-horizon contexts and hallucination about the task state. We explore the use of explicit belief state representations to mitigate these issues, finding that it enhances task performance and the accuracy of ToM inferences for LLM-based agents.

Owen Burns, Dana Hughes, Katia Sycara

Automated planning using a symbolic planning language, such as PDDL, is a general approach to producing optimal plans to achieve a stated goal. However, creating suitable machine understandable descriptions of the planning domain, problem, and goal requires expertise in the planning language, limiting the utility of these tools for non-expert humans. Recent efforts have explored utilizing a symbolic planner in conjunction with a large language model to generate plans from natural language descriptions given by a non-expert human (LLM+PDDL). Our approach performs initial translation of goal specifications to a set of PDDL goal constraints using an LLM; such translations often result in imprecise symbolic specifications, which are difficult to validate directly. We account for this using an evolutionary approach to generate a population of symbolic goal specifications with slight differences from the initial translation, and utilize a trained LSTM-based validation model to assess whether each induced plan in the population adheres to the natural language specifications. We evaluate our approach on a collection of prototypical specifications in a notional naval disaster recovery task, and demonstrate that our evolutionary approach improve adherence of generated plans to natural language specifications when compared to plans generated using only LLM translations. The code for our method can be found at https://github.com/owenonline/PlanCritic.

Sarah Aguasvivas Manzano, Dana Hughes, Cooper Simpson et al.

We present a library to automatically embed signal processing and neural network predictions into the material robots are made of. Deep and shallow neural network models are first trained offline using state-of-the-art machine learning tools and then transferred onto general purpose microcontrollers that are co-located with a robot's sensors and actuators. We validate this approach using multiple examples: a smart robotic tire for terrain classification, a robotic finger sensor for load classification and a smart composite capable of regressing impact source localization. In each example, sensing and computation are embedded inside the material, creating artifacts that serve as stand-in replacement for otherwise inert conventional parts. The open source software library takes as inputs trained model files from higher level learning software, such as Tensorflow/Keras, and outputs code that is readable in a microcontroller that supports C. We compare the performance of this approach for various embedded platforms. In particular, we show that low-cost off-the-shelf microcontrollers can match the accuracy of a desktop computer, while being fast enough for real-time applications at different neural network configurations. We provide means to estimate the maximum number of parameters that the hardware will support based on the microcontroller's specifications.

Shreya Sharma, Dana Hughes, Katia Sycara

This paper describes CBGT-Net, a neural network model inspired by the cortico-basal ganglia-thalamic (CBGT) circuits found in mammalian brains. Unlike traditional neural network models, which either generate an output for each provided input, or an output after a fixed sequence of inputs, the CBGT-Net learns to produce an output after a sufficient criteria for evidence is achieved from a stream of observed data. For each observation, the CBGT-Net generates a vector that explicitly represents the amount of evidence the observation provides for each potential decision, accumulates the evidence over time, and generates a decision when the accumulated evidence exceeds a pre-defined threshold. We evaluate the proposed model on two image classification tasks, where models need to predict image categories based on a stream of small patches extracted from the image. We show that the CBGT-Net provides improved accuracy and robustness compared to models trained to classify from a single patch, and models leveraging an LSTM layer to classify from a fixed sequence length of patches.

Mycal Tucker, Huao Li, Siddharth Agrawal et al.

Neural agents trained in reinforcement learning settings can learn to communicate among themselves via discrete tokens, accomplishing as a team what agents would be unable to do alone. However, the current standard of using one-hot vectors as discrete communication tokens prevents agents from acquiring more desirable aspects of communication such as zero-shot understanding. Inspired by word embedding techniques from natural language processing, we propose neural agent architectures that enables them to communicate via discrete tokens derived from a learned, continuous space. We show in a decision theoretic framework that our technique optimizes communication over a wide range of scenarios, whereas one-hot tokens are only optimal under restrictive assumptions. In self-play experiments, we validate that our trained agents learn to cluster tokens in semantically-meaningful ways, allowing them communicate in noisy environments where other techniques fail. Lastly, we demonstrate both that agents using our method can effectively respond to novel human communication and that humans can understand unlabeled emergent agent communication, outperforming the use of one-hot communication.

Ini Oguntola, Dana Hughes, Katia Sycara

When developing AI systems that interact with humans, it is essential to design both a system that can understand humans, and a system that humans can understand. Most deep network based agent-modeling approaches are 1) not interpretable and 2) only model external behavior, ignoring internal mental states, which potentially limits their capability for assistance, interventions, discovering false beliefs, etc. To this end, we develop an interpretable modular neural framework for modeling the intentions of other observed entities. We demonstrate the efficacy of our approach with experiments on data from human participants on a search and rescue task in Minecraft, and show that incorporating interpretability can significantly increase predictive performance under the right conditions.

Tianwei Ni, Huao Li, Siddharth Agrawal et al.

Teamwork is a set of interrelated reasoning, actions and behaviors of team members that facilitate common objectives. Teamwork theory and experiments have resulted in a set of states and processes for team effectiveness in both human-human and agent-agent teams. However, human-agent teaming is less well studied because it is so new and involves asymmetry in policy and intent not present in human teams. To optimize team performance in human-agent teaming, it is critical that agents infer human intent and adapt their polices for smooth coordination. Most literature in human-agent teaming builds agents referencing a learned human model. Though these agents are guaranteed to perform well with the learned model, they lay heavy assumptions on human policy such as optimality and consistency, which is unlikely in many real-world scenarios. In this paper, we propose a novel adaptive agent architecture in human-model-free setting on a two-player cooperative game, namely Team Space Fortress (TSF). Previous human-human team research have shown complementary policies in TSF game and diversity in human players' skill, which encourages us to relax the assumptions on human policy. Therefore, we discard learning human models from human data, and instead use an adaptation strategy on a pre-trained library of exemplar policies composed of RL algorithms or rule-based methods with minimal assumptions of human behavior. The adaptation strategy relies on a novel similarity metric to infer human policy and then selects the most complementary policy in our library to maximize the team performance. The adaptive agent architecture can be deployed in real-time and generalize to any off-the-shelf static agents. We conducted human-agent experiments to evaluate the proposed adaptive agent framework, and demonstrated the suboptimality, diversity, and adaptability of human policies in human-agent teams.

Vidhi Jain, Rohit Jena, Huao Li et al.

In a search and rescue scenario, rescuers may have different knowledge of the environment and strategies for exploration. Understanding what is inside a rescuer's mind will enable an observer agent to proactively assist them with critical information that can help them perform their task efficiently. To this end, we propose to build models of the rescuers based on their trajectory observations to predict their strategies. In our efforts to model the rescuer's mind, we begin with a simple simulated search and rescue task in Minecraft with human participants. We formulate neural sequence models to predict the triage strategy and the next location of the rescuer. As the neural networks are data-driven, we design a diverse set of artificial "faux human" agents for training, to test them with limited human rescuer trajectory data. To evaluate the agents, we compare it to an evidence accumulation method that explicitly incorporates all available background knowledge and provides an intended upper bound for the expected performance. Further, we perform experiments where the observer/predictor is human. We show results in terms of prediction accuracy of our computational approaches as compared with that of human observers.

Dana Hughes, Nikolaus Correll

This paper reviews machine learning applications and approaches to detection, classification and control of intelligent materials and structures with embedded distributed computation elements. The purpose of this survey is to identify desired tasks to be performed in each type of material or structure (e.g., damage detection in composites), identify and compare common approaches to learning such tasks, and investigate models and training paradigms used. Machine learning approaches and common temporal features used in the domains of structural health monitoring, morphable aircraft, wearable computing and robotic skins are explored. As the ultimate goal of this research is to incorporate the approaches described in this survey into a robotic material paradigm, the potential for adapting the computational models used in these applications, and corresponding training algorithms, to an amorphous network of computing nodes is considered. Distributed versions of support vector machines, graphical models and mixture models developed in the field of wireless sensor networks are reviewed. Potential areas of investigation, including possible architectures for incorporating machine learning into robotic nodes, training approaches, and the possibility of using deep learning approaches for automatic feature extraction, are discussed.