Understanding and Enhancing Human Decision Making in a Dynamic World
In today’s world, decisions are made in environments that are complex, fast-changing, and uncertain. Options do not appear all at once—they unfold over time, often under pressure, requiring people to explore, adapt, and act with limited information. Whether navigating information overload, operating under time constraints, or responding to unpredictable change, effective decision-making is both challenging and critical.
At the Dynamic Decision Making Laboratory (DDMLab), we study how people make decisions in these dynamic environments. Our research develops and tests cognitive theories that explain how individuals learn from experience, adapt over time, and make sequential choices under uncertainty. We formalize these processes through computational models that capture the mechanisms underlying human decision-making.
By linking theory to application, we use these models to inform the design of tools and systems that improve decision-making in high-stakes domains, including healthcare, emergency response, cybersecurity, and beyond.
Our Research Approach: Bridging Human Behavior and Cognitive Modeling
At DDMLab, we integrate controlled behavioral experiments with cognitive computational modeling to understand, and ultimately improve, human decision-making in dynamic environments.
|
In our experiments, individuals and teams engage in tasks that evolve over time and space, requiring continuous adaptation under uncertainty. These studies reveal the strategies people use, how they learn from experience, and where systematic biases or limitations emerge. We complement this work with cognitive models that simulate decision-making processes through formal algorithms. These models operate in the same tasks as humans, enabling direct comparisons across key dimensions such as learning, risk sensitivity, adaptation, and performance relative to optimal benchmarks. |
 |
By aligning human behavior with model predictions, we identify the cognitive mechanisms that drive decisions in dynamic settings. These insights guide the development of practical solutions in domains including cybersecurity, climate resilience, phishing prevention, and human–AI interaction.
Instance-Based Learning: How Experience Shapes Decisions
At DDMLab, we use Instance-Based Learning Theory (IBLT) () to explain how people make decisions from experience. In dynamic situations individuals rely on memories of past situations—instances—to guide their choices.
When facing a new decision, people retrieve similar past experiences, evaluate the outcomes associated with those experiences, and select the option that appears most promising. After acting, the outcome is stored as a new instance, continuously updating the knowledge base that supports future decisions. This cycle of retrieval, choice, and feedback enables learning and adaptation in dynamic environments.
IBLT formalizes this process through a mathematical framework grounded in the The model specifies how memory is shaped by recency, frequency, and similarity—factors that determine which experiences are retrieved and how strongly they influence decisions.
To make these ideas operational, we developed PyIBL, a Python-based platform for building and testing IBL models. PyIBL allows researchers and practitioners to simulate human decision-making, evaluate competing hypotheses, and apply IBLT to real-world problems. Visit our Cognitive Modeling page to access PyIBL, documentation, and example models.
We evaluate IBLT—and alternative theories of decisions from experience—through systematic model comparison. Models are assessed based on how well they reproduce observed human behavior across tasks, providing a rigorous foundation for theory development and application.
Latest Projects: Behavioral Cybersecurity
Impact of Cognition on Cyber Behavior
This research program aims to improve cybersecurity by understanding how human cognition impacts cyber behavior and could affect cyber actors' success in network attack activities. Specifically, well-established cognitive patterns, such as loss aversion and the representativeness bias, are be investigated as potentially mitigating factors in the efficacy of cyber attack behavior. This research contributes to the broader goals of improving cyber defense practices by delaying and thwarting attacks.
Phishing Training and Detection
Phishing relies largely on social engineering tactics, where attackers take advantage of human weaknesses such as: reacting to familiar senders, to immediate requests, and to emotional requests. IBLT demonstrates that phishing classification decisions are influenced by the type of experiences people have. We build to emulate end-used classification decisions and compared classification decisions from humans in email processing tasks with the goal to build better training scenarios.
Human-Machine Collaborations in Autonomous Cyber Operations
IBLT is in a theory of individual decision making, and groups learn through individual group members. We have demonstrated that group effects and dynamics can be captured by the aggregation of individual members of a group and their interdependencies, and have constructed an architecture in which IBL models develop by observing other agents.
Defense Strategies in a Repeated Binary Choice Task
Adapting to dynamic environments poses significant challenges for humans, even in seemingly simple scenarios, such as repeated binary choice tasks. This research investigates the effectiveness of interventions based on Instance-Based Learning (IBL) cognitive models and what role they could play in facilitating adaptation to changing conditions, with the goal to design defense strategies to influence human choices in real time.
Latest Projects: Human-Machine Collaborations
Disaster Relief Management Decision Making
This research program aims to test how real professional disaster relief managers make decisions about resource allocation and information gathering during natural disasters. We examine questions of perceptual aggregation, allocation trade-offs, and temporal drift. Our goal is to support the development of learned social welfare functions that could inform decision support tools for future disaster response.
Theory of Mind Capabilities in AI Partners
As autonomous agents become more ubiquitous, it is important that we understand how best to design AI agents for effective human-AI collaboration. In line with that goal, much work has been devoted to developing AI agents that are adaptive to a variety of situations and to human partners. Theory of Mind (ToM) has been suggested as a solution to achieve implicit coordination between team members.
Cognitive Models of Behavior in Sequential Decision Tasks
To understand how people make sequential decisions in various tasks involving balancing exploration and exploitation, we develop cognitive models of their behavior in these tasks. In this research prgram we introduce a novel sequential stopping task to shed light on how, when, and why people decide to stop exploring. We then systematically examine some of the factors that may influence stopping behavior and validate the predictions of our cognitive model. And finally we Investigate interventions leveraging wisdom of crowdsaggregation techniques to provide personalized, cognitive AI-driven recommendations for when to stop searching.
Cognitively Aware Reinforcement Learning
In collaborative domains, a desirable attribute of AI partners is the ability to be adaptive to the behaviors and preferences of humans. The goal of this research is to investigate how cognitive models can be used in tandem with reinforcement learning (RL) agents to learn policies that complement human behavior. To achieve this we incorporate cognitive models into the RL training and testing pipelines to see how such models can improve performance in cooperative tasks. We test these models with human proxies and real humans, and analyze their behavior using collaborative fluency metrics, to see how well they learn collaborative policies.
