MARA
Modeling, Abstraction, and Reasoning Agents: systems that build and use world models through active experimentation and abstract reasoning.
People
Archana Warrier
Research Trainee, Basis, New York
Archana Warrier is a research trainee at Basis, advised by Zenna Tavares, and an incoming ELLIS PhD student at TU Darmstadt, advised by Angela Yu in the Computational Modelling of Intelligent Systems lab. She is interested in building agents that learn and efficiently update world models of their environment and of other agents' goals, beliefs, and abilities.
About
I am a research trainee at Basis, advised by Zenna Tavares, where I work on MARA, a long-term effort to build AI systems capable of everyday scientific discovery through active experimentation and abstract reasoning.
I am an incoming ELLIS PhD student at TU Darmstadt, advised by Angela Yu in the Computational Modelling of Intelligent Systems lab, where I will work on collaborative multi-agent systems: building agents that model other agents as intentional systems with their own goals, beliefs, and abilities, and update those models efficiently through interaction.
I am broadly interested in understanding intelligence: how humans think, and how thinking can be modeled computationally. Rather than directly imitating cognition, I want to identify the principles behind intelligent behavior.
I completed my Master’s in Computer Science at RPTU Kaiserslautern-Landau and my Bachelor’s in Mathematics and Computing at Birla Institute of Technology, Mesra.
Projects
Current and recent Basis projects.
Recent Publications
Recent papers and preprints.
Benchmarking World-Model Learning via Environment-Level Queries
A. Warrier, D. Nguyen, M. Naim, Moksh Jain, Y. Liang, K. Schroeder, Cambridge Yang, Joshua B. Tenenbaum, Sebastian J. Vollmer, Kevin Ellis, Z. Tavares
Articles
Basis essays and updates this person wrote or contributed to.
AutumnBench: World Model Learning in Humans and AI
We’re releasing a new version of Autumn with human baseline results, AI performance comparisons, and an interactive benchmark for world model discovery. This release includes the MARA protocol and provides a public platform for testing causal reasoning capabilities.