David V. Pynadath

JAAMAS Journal 2026 Journal Article

An Automated Teamwork Infrastructure for Heterogeneous Software Agents and Humans

• David V. Pynadath
• Milind Tambe

Abstract Agent integration architectures enable a heterogeneous, distributed set of agents to work together to address problems of greater complexity than those addressed by the individual agents themselves. Unfortunately, integrating software agents and humans to perform real-world tasks in a large-scale system remains difficult, especially due to three main challenges: ensuring robust execution in the face of a dynamic environment, providing abstract task specifications without all the low-level coordination details, and finding appropriate agents for inclusion in the overall system. To address these challenges, our Teamcore project provides the integration architecture with general-purpose teamwork coordination capabilities. We make each agent team-ready by providing it with a proxy capable of general teamwork reasoning. Thus, a key novelty and strength of our framework is that powerful teamwork capabilities are built into its foundations by providing the proxies themselves with a teamwork model. Given this teamwork model, the Teamcore proxies addresses the first agent integration challenge, robust execution, by automatically generating the required coordination actions for the agents they represent. We can also exploit the proxies' reusable general teamwork knowledge to address the second agent integration challenge. Through team - oriented programming, a developer specifies a hierarchical organization and its goals and plans, abstracting away from coordination details. Finally, KARMA, our Knowledgeable Agent Resources Manager Assistant, can aid the developer in conquering the third agent integration challenge by locating agents that match the specified organization's requirements. Our integration architecture enables teamwork among agents with no coordination capabilities, and it establishes and automates consistent teamwork among agents with some coordination capabilities. Thus, team-oriented programming provides a level of abstraction that can be used on top of previous approaches to agent-oriented programming. We illustrate how the Teamcore architecture successfully addressed the challenges of agent integration in two application domains: simulated rehearsal of a military evacuation mission and facilitation of human collaboration.

AAMAS Conference 2025 Conference Paper

Predicting Team Performance from Communications in Simulated Search-and-Rescue

• Ali Jalal-Kamali
• Nikolos M. Gurney
• David V. Pynadath

Understanding how individual traits influence team performance is valuable, but these traits are not always directly observable. Prior research has inferred traits like trust from behavioral data. We analyze conversational data to identify team traits and their correlation with teaming outcomes. Using transcripts from a Minecraft-based search-and-rescue experiment, we apply topic modeling and clustering to uncover key interaction patterns. Our findings show that variations in teaming outcomes can be explained through these inferences, with different levels of predictive power derived from individual traits and team dynamics.

AAMAS Conference 2023 Conference Paper

Effectiveness of Teamwork-Level Interventions through Decision-Theoretic Reasoning in a Minecraft Search-and-Rescue Task

• David V. Pynadath
• Nikolos Gurney
• Sarah Kenny
• Rajay Kumar
• Stacy C. Marsella
• Haley Matuszak
• Hala Mostafa
• Pedro Sequeira

AAMAS Conference 2023 Conference Paper

Multiagent Inverse Reinforcement Learning via Theory of Mind Reasoning

• Haochen Wu
• Pedro Sequeira
• David V. Pynadath

We approach the problem of understanding how people interact with each other in collaborative settings, especially when individuals know little about their teammates, via Multiagent Inverse Reinforcement Learning (MIRL), where the goal is to infer the reward functions guiding the behavior of each individual given trajectories of a team’s behavior during some task. Unlike current MIRL approaches, we do not assume that team members know each other’s goals a priori; rather, that they collaborate by adapting to the goals of others perceived by observing their behavior, all while jointly performing a task. To address this problem, we propose a novel approach to MIRL via Theory of Mind (MIRL-ToM). For each agent, we first use ToM reasoning to estimate a posterior distribution over baseline reward profiles given their demonstrated behavior. We then perform MIRL via decentralized equilibrium by employing single-agent Maximum Entropy IRL to infer a reward function for each agent, where we simulate the behavior of other teammates according to the time-varying distribution over profiles. We evaluate our approach in a simulated 2-player search-and-rescue operation where the goal of the agents, playing different roles, is to search for and evacuate victims in the environment. Our results show that the choice of baseline profiles is paramount to the recovery of the ground-truth rewards, and that MIRL-ToM is able to recover the rewards used by agents interacting both with known and unknown teammates.

AAAI Conference 2023 Conference Paper

The Role of Heuristics and Biases during Complex Choices with an AI Teammate

• Nikolos Gurney
• John H. Miller
• David V. Pynadath

Behavioral scientists have classically documented aversion to algorithmic decision aids, from simple linear models to AI. Sentiment, however, is changing and possibly accelerating AI helper usage. AI assistance is, arguably, most valuable when humans must make complex choices. We argue that classic experimental methods used to study heuristics and biases are insufficient for studying complex choices made with AI helpers. We adapted an experimental paradigm designed for studying complex choices in such contexts. We show that framing and anchoring effects impact how people work with an AI helper and are predictive of choice outcomes. The evidence suggests that some participants, particularly those in a loss frame, put too much faith in the AI helper and experienced worse choice outcomes by doing so. The paradigm also generates computational modeling-friendly data allowing future studies of human-AI decision making.

AAMAS Conference 2018 Conference Paper

Clustering Behavior to Recognize Subjective Beliefs in Human-Agent Teams

• David V. Pynadath
• Ning Wang
• Ericka Rovira
• Michael J. Barnes

Trust is critical to the success of human-agent teams, and a critical antecedents to trust is transparency. To best interact with human teammates, an agent explain itself so that they understand its decision-making process. However, individual differences among human teammates require that the agent dynamically adjust its explanation strategy based on their unobservable subjective beliefs. The agent must therefore recognize its teammates’ subjective beliefs relevant to trust-building (e. g. , their understanding of the agent’s capabilities and process). We leverage a nonparametric method to enable an agent to use its history of prior interactions as a means for recognizing and predicting a new teammate’s subjective beliefs. We first gather data combining observable behavior sequences with survey-based observations of typically unobservable perceptions. We then use a nearest-neighbor approach to identify the prior teammates most similar to the new one. We use these neighbors’ responses to infer the likelihood of possible beliefs, as in collaborative filtering. The results provide insights into the types of beliefs that are easy (and hard) to infer from purely behavioral observations.

AAMAS Conference 2016 Conference Paper

Semi-Automated Construction of Decision-Theoretic Models of Human Behavior

• David V. Pynadath
• Heather Rosoff
• Richard S. John

Multiagent social simulation provides a powerful mechanism for policy makers to understand the potential outcomes of their decisions before implementing them. However, the value of such simulations depends on the accuracy of their underlying agent models. In this work, we present a method for automatically exploring a space of decision-theoretic models to arrive at a multiagent social simulation that is consistent with human behavior data. We start with a factored Partially Observable Markov Decision Process (POMDP) whose states, actions, and reward capture the questions asked in a survey from a disaster response scenario. Using input from domain experts, we construct a set of hypothesized dependencies that may or may not exist in the transition probability function. We present an algorithm to search through each of these hypotheses, evaluate their accuracy with respect to the data, and choose the models that best reflect the observed behavior, including individual differences. The result is a mechanism for constructing agent models that are grounded in human behavior data, while still being able to support hypothetical reasoning that is the main advantage of multiagent social simulation. General Terms Algorithms

AAMAS Conference 2016 Conference Paper

The Impact of POMDP-Generated Explanations on Trust and Performance in Human-Robot Teams

• Ning Wang
• David V. Pynadath
• Susan G. Hill

Researchers have observed that people will more accurately trust an autonomous system, such as a robot, if they have a more accurate understanding of its decision-making process. Studies have shown that hand-crafted explanations can help maintain effective team performance even when the system is less than 100% reliable. However, current explanation algorithms are not sufficient for making a robot’s quantitative reasoning (in terms of both uncertainty and conflicting goals) transparent to human teammates. In this work, we develop a novel mechanism for robots to automatically generate explanations of reasoning based on Partially Observable Markov Decision Problems (POMDPs). Within this mechanism, we implement alternate natural-language templates and then measure their differential impact on trust and team performance within an agent-based online testbed that simulates a human-robot team task. The results demonstrate that the added explanation capability leads to improvement in transparency, trust, and team performance. Furthermore, by observing the different outcomes due to variations in the robot’s explanation content, we gain valuable insight that can help lead to refinement of explanation algorithms to further improve human-robot interaction. General Terms Algorithms

JAAMAS Journal 2009 Journal Article

Modeling appraisal in theory of mind reasoning

• Mei Si
• Stacy C. Marsella
• David V. Pynadath

Abstract Cognitive appraisal theories, which link human emotional experience to their interpretations of events happening in the environment, are leading approaches to model emotions. Cognitive appraisal theories have often been used both for simulating “real emotions” in virtual characters and for predicting the human user’s emotional experience to facilitate human–computer interaction. In this work, we investigate the computational modeling of appraisal in a multi-agent decision-theoretic framework using Partially Observable Markov Decision Process-based (POMDP) agents. Domain-independent approaches are developed for five key appraisal dimensions (motivational relevance, motivation congruence, accountability, control and novelty). We also discuss how the modeling of theory of mind (recursive beliefs about self and others) is realized in the agents and is critical for simulating social emotions. Our model of appraisal is applied to three different scenarios to illustrate its usages. This work not only provides a solution for computationally modeling emotion in POMDP-based agents, but also illustrates the tight relationship between emotion and cognition—the appraisal dimensions are derived from the processes and information required for the agent’s decision-making and belief maintenance processes, which suggests a uniform cognitive structure for emotion and cognition.

JAAMAS Journal 2009 Journal Article

Modeling self-deception within a decision-theoretic framework

• Jonathan Y. Ito
• David V. Pynadath
• Stacy C. Marsella

Abstract Computational modeling of human belief maintenance and decision-making processes has become increasingly important for a wide range of applications. In this paper, we present a framework for modeling the human capacity for self-deception from a decision-theoretic perspective in which we describe an integrated process of wishful thinking which includes the determination of a desired belief state, the biasing of internal beliefs towards or away from this desired belief state, and the final decision-making process. Finally, we show that in certain situations self-deception can be beneficial.

AAMAS Conference 2009 Conference Paper

Self-Deceptive Decision Making: Normative and Descriptive Insights

• Jonathan Y. Ito
• David V. Pynadath
• Stacy C. Marsella

Computational modeling of human belief maintenance and decision-making processes has become increasingly important for a wide range of applications. We present a framework for modeling the psychological phenomenon of selfdeception in a decision-theoretic framework. Specifically, we model the self-deceptive behavior of wishful thinking as a psychological bias towards the belief in a particularly desirable situation or state. By leveraging the structures and axioms of Expected Utility (EU) Theory we are able to operationalize both the determination and the application of the desired belief state with respect to the decision-making process of expected utility maximization. While we categorize our framework as a descriptive model of human decision making, we show that in certain circumstances the realized expected utility of an action biased by wishful thinking can exceed that of an action motivated purely by the maximization of perceived expected utility. Finally, we show that our framework of self-deception and wishful thinking has the descriptive flexibility to account for the inconsistencies highlighted by the Common Ratio Effect and the Allais Paradox.

AAAI Conference 2007 Conference Paper

Minimal Mental Models

• David V. Pynadath

Agents must form and update mental models about each other in a wide range of domains: team coordination, plan recognition, social simulation, user modeling, games of incomplete information, etc. Existing research typically treats the problem of forming beliefs about other agents as an isolated subproblem, where the modeling agent starts from an initial set of possible models for another agent and then maintains a belief about which of those models applies. This initial set of models is typically a full specification of possible agent types. Although such a rich space gives the modeling agent high accuracy in its beliefs, it will also incur high cost in maintaining those beliefs. In this paper, we demonstrate that by taking this modeling problem out of its isolation and placing it back within the overall decision-making context, the modeling agent can drastically reduce this rich model space without sacrificing any performance. Our approach comprises three methods. The first method clusters models that lead to the same behaviors in the modeling agent’s decision-making context. The second method clusters models that may produce different behaviors, but produce equally preferred outcomes with respect to the utility of the modeling agent. The third technique sacrifices a fixed amount of accuracy by clustering models that lead to performance losses that are below a certain threshold. We illustrate our framework using a social simulation domain and demonstrate its value by showing the minimal mental model spaces that it generates.

IJCAI Conference 2005 Conference Paper

PsychSim: Modeling Theory of Mind with Decision-Theoretic Agents

• David V. Pynadath
• Stacy C

Agent-based modeling of human social behavior is an increasingly important research area. A key factor in human social interaction is our beliefs about others, a theory of mind. Whether we believe a message depends not only on its content but also on our model of the communicator. How we act depends not only on the immediate effect but also on how we believe others will react. In this paper, we discuss PsychSim, an implemented multiagent-based simulation tool for modeling interactions and influence. While typical approaches to such modeling have used first-order logic, Psych- Sim agents have their own decision-theoretic model of the world, including beliefs about its environment and recursive models of other agents. Using these quantitative models of uncertainty and preferences, we have translated existing psychological theories into a decision-theoretic semantics that allow the agents to reason about degrees of believability in a novel way. We discuss PsychSim’s underlying architecture and describe its application to a school violence scenario for illustration.

AAMAS Conference 2003 Conference Paper

A Prototype Infrastructure for Distributed Robot-Agent-Person Teams

• Paul Scerri
• David V. Pynadath
• Lewis Johnson
• Paul Rosenbloom
• Nathan Schurr
• Milind Tambe

AAMAS Conference 2002 Conference Paper

Multiagent Teamwork: Analyzing the Optimality and Complexity of Key Theories and Models

• David V. Pynadath
• Milind Tambe

AAMAS Conference 2002 Conference Paper

Why the Elf Acted Autonomously: Towards a Theory of Adjustable Autonomy

• Paul Scerri
• David V. Pynadath
• Milind Tambe

UAI Conference 2000 Conference Paper

Probabilistic State-Dependent Grammars for Plan Recognition

• David V. Pynadath
• Michael P. Wellman

Techniques for plan recognition under uncertainty require a stochastic model of the plan-generation process. We introduce Probabilistic State-Dependent Grammars (PSDGs) to represent an agent's plan-generation process. The PSDG language model extends probabilistic context-free grammars (PCFGs) by allowing production probabilities to depend on an explicit model of the planning agent's internal and external state. Given a PSDG description of the plan-generation process, we can then use inference algorithms that exploit the particular independence properties of the PSDG language to efficiently answer plan-recognition queries. The combination of the PSDG language model and inference algorithms extends the range of plan-recognition domains for which practical probabilistic inference is possible, as illustrated by applications in traffic monitoring and air combat.

AAAI Conference 1996 Conference Paper

Generalized Queries on Probabilistic Context-Free Grammars

• David V. Pynadath

Probabilistic context-free grammars (PCFGs) provide a simple way to represent a particular class of distributions over sentences in a context-free language. Efficient parsing algorithms for answering particular queries about a PCFG (i. e. , calculating the probability of a given sentence, or finding the most likely parse) have been applied to a variety of pattern-recognition problems. We extend the class of queries that can be answered in several ways: (1) allowing missing tokens in a sentence or sentence fragment, (2) supporting queries about intermediate structure, such as the presence of particular nonterminals, and (3) flexible conditioning on a variety of types of evidence. Our method works by constructing a Bayesian network to represent the distribution of parse trees induced by a given PCFG. The network structure mirrors that of the chart in a standard parser, and is generated using a similar dynamic-programming approach. We present an algorithm for constructing Bayesian networks from PCFGs, and show how queries or patterns of queries on the network correspond to interesting queries on PCFGs.

UAI Conference 1995 Conference Paper

Accounting for Context in Plan Recognition, with Application to Traffic Monitoring

• David V. Pynadath
• Michael P. Wellman

Typical approaches to plan recognition start from a representation of an agent's possible plans, and reason evidentially from observations of the agent's actions to assess the plausibility of the various candidates. A more expansive view of the task (consistent with some prior work) accounts for the context in which the plan was generated, the mental state and planning process of the agent, and consequences of the agent's actions in the world. We present a general Bayesian framework encompassing this view, and focus on how context can be exploited in plan recognition. We demonstrate the approach on a problem in traffic monitoring, where the objective is to induce the plan of the driver from observation of vehicle movements. Starting from a model of how the driver generates plans, we show how the highway context can appropriately influence the recognizer's interpretation of observed driver behavior.