WHO WE ARE ::

Monjira Biswas
Hajira Choudry
Yasaman Kazerooni
Sooin Lee
Mehgha Shyam

research assistants

COLLABORATORS ::

universität mainz
university of birmingham
university of massachusetts
universität marburg
nasa ames research center
national science foundation
universität
potsdam

Xiaoxiao Guo
Nan Jiang
John Laird

Satinder Singh

michigan computer science
michigan linguistics
Marc Berman
university of toronto
michigan psychology

RECENT PAPERS

Quick links below; for more publications and citation information, click here.

2019
Thirty-Third AAAI Conference on Artificial Intelligence (AAAI)
Learning to Communicate and Solve Visual Blocks-World Tasks

2018
Journal of Psycholinguistic Research
The Quantificational Domain of dou: An Experimental Study

2018
Computational Interaction
Interaction as an emergent property of a partially observable markov decision process

2018
Proceedings of the 8th Workshop on Cognitive Modeling and Computational Linguistics (CMCL 2018)
Dynamic encoding of structural uncertainty in gradient symbols

2018
Behavioral and Brain Sciences}
The role of (bounded) optimization in theory testing and prediction

2017
Proceedings of the Annual Conference of the Cognitive Science Society
Human Visual Search as a Deep Reinforcement Learning Solution to a POMDP

2016
Psychological Review
Why contextual preference reversals maximize expected value

2016
25th International Joint Conference on Artificial Intelligence (IJCAI)
Deep Learning for Reward Design to Improve Monte Carlo Tree Search in ATARI Games

2016
Frontiers in Psychology
Retrieval Interference in Syntactic Processing: The Case of Reflexive Binding in English

2015
Cognitive Science
Predicting Short-Term Remembering as Boundedly Optimal Strategy Choice

2015
14th International Conference on Autonomous Agents and Multiagent Systems (AAMAS2015)
The Dependence of Effective Planning Horizon on Model Accuracy

2015
Advances in Neural Information Processing Systems (NIPS)
Action-Conditional Video Prediction Using Deep Networks in Atari Games

2014
Advances in Neural Information Processing Systems (NIPS)
Deep Learning for Real-Time Atari Game Play Using Offline Monte-Carlo Tree Search Planning

2014
Topics in Cognitive Science
Utility Maximization and Bounds on Human Information Processing

2014
Proceedings of the 13th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2014)
Improving UCT Planning via Approximate Homomorphisms

2014
Topics in Cognitive Science
Computational Rationality: Linking Mechanism and Behavior Through Utility Maximization

2014
IEEE Transactions on Autonomous Mental Development
Optimal Rewards for Cooperative Agents

2014
Proceedings of the 5th Workshop on Cognitive Modeling and Computational Linguistics (CMCL 2014)
Computationally Rational Saccadic Control: An Explanation of Spillover Effects Based on Sampling from Noisy Perception and Memory

Welcome to the Language and Cognitive Architecture Lab at the University of Michigan.

 

about our research

The goal of our research is to develop theories of language, thought, and action—theories capturing the adaptive nature of human behavior, and grounded in integrated architectures that explain how the computational subsystems of the mind and brain work together.

An over-arching theoretical principle guiding much of our work is computational rationality—the idea that behavior is the adaptive response to the joint constraints of the biological processing architecture and the external probabilistic environment.

The specific topics we focus on are the (boundedly optimal) adaptive control of perceptual, motor, cognitive and linguistic processes; language processing (especially the role of working memory in sentence comprehension and production); flexible artificial intelligence (AI) agents and reinforcement learning; and the interaction of cognition and emotion.

The work makes contact with several areas of cognitive science, including psycholinguistics, linguistic theory, cognitive psychology, cognitive modeling, human-computer interaction, cognitive architectures, and reinforcement learning. It is highly collaborative (see left). To learn more about the research, click on the topics at left, or the questions below.