Archana Warrier

I’m interested in identifying and modeling the computational principles that underlie goal-directed collaboration, and in building AI systems grounded in those principles. Rather than copying biological mechanisms directly, I want to draw on the principles behind biological collaboration to build collaborative AI.

I’m currently a Research Trainee at Basis (New York), advised by Zenna Tavares, where I work on Project MARA—building AI systems capable of everyday scientific discovery through active experimentation and abstract reasoning. I’ll soon be starting as an incoming ELLIS PhD student at TU Darmstadt, advised by Angela Yu in the Computational Modelling of Intelligent Systems lab.

Previously, 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.

You can find more about my work on the publications and projects pages.

news

Apr 30, 2026	Our paper Benchmarking World-Model Learning via Environment-Level Queries was accepted at ICML 2026. The paper introduces WorldTest, a protocol for evaluating world-model learning through environment-level queries that go beyond next-frame prediction — testing whether agents can predict unobserved states, plan action sequences, and detect changes in dynamics. We instantiate WorldTest with AutumnBench, a suite of 43 interactive grid-world environments and 129 tasks across three families: masked-frame prediction, planning, and predicting changes to causal dynamics. I’ll be presenting the poster at the conference.
Apr 03, 2026	Excited to share that I’ll be joining the ELLIS PhD Program as an incoming PhD student at TU Darmstadt, advised by Angela Yu in the Computational Modelling of Intelligent Systems lab.
Jul 18, 2025	Presented a poster at ICML 2025 in Vancouver, Canada, as part of Basis on AutumnBench at the World Model Learning Workshop.
Aug 05, 2024	Started as Research Trainee at Basis, New York, advised by Zenna Tavares. I’m contributing to Project MARA (Modeling, Abstraction, and Reasoning Agents), which aims to build AI systems capable of everyday scientific discovery through active experimentation and abstract reasoning.

selected publications

ICML
Benchmarking World-Model Learning with Environment-Level Queries

Archana Warrier, D. Nguyen, M. Naim, and 8 more authors

International Conference on Learning Representations, 2026

Abs Bib Website

Model-learning agents should gather information to learn world models that support many downstream tasks and inferences, such as predicting unobserved states, estimating near- and far-term consequences of actions, planning action sequences, and detecting changes in dynamics. Current methods for learning and evaluating world models diverge from this goal: training and evaluation are anchored to next-frame prediction, and success is scored by reward maximization in the same environment. We propose WorldTest, a protocol to evaluate model-learning agents that separates reward-free interaction from a scored test phase in a different but related environment. WorldTest is open-ended—models should support many different tasks unknown ahead of time—and agnostic to model representation, allowing comparison across approaches. We instantiated WorldTest with AutumnBench, a suite of 43 interactive grid-world environments and 129 tasks across three families: masked-frame prediction, planning, and predicting changes to the causal dynamics. We compared 517 human participants and three frontier models on AutumnBench. We found that humans outperform the models, and scaling compute improves performance only in some environments but not others. WorldTest provides a novel template—reward-free exploration, derived tests, and behavior-based scoring—to evaluate what agents learn about environment dynamics, and AutumnBench exposes significant headroom in world-model learning.
@article{warrier2025benchmarking, title = {Benchmarking World-Model Learning with Environment-Level Queries}, author = {Warrier, Archana and Nguyen, D. and Naim, M. and Jain, M. and Liang, Y. and Schroeder, K. and Yang, C. and Tenenbaum, J. and Vollmer, S. and Ellis, K. and Tavares, Z.}, journal = {International Conference on Learning Representations}, year = {2026}, }
NeurIPS Workshop
Had Enough of Experts? Elicitation and Evaluation of Bayesian Priors from LLMs

D. Selby, K. Spriestersbach, Y. Iwashita, and 5 more authors

NeurIPS 2024 Workshop, 2024

Abs Bib

Large language models (LLMs) have been extensively studied for their abilities to generate convincing natural language sequences, however their utility for quantitative information retrieval is less well understood. Here we explore the feasibility of LLMs as a mechanism for quantitative knowledge retrieval to aid elicitation of expert-informed prior distributions for Bayesian statistical models. We present a prompt engineering framework, treating an LLM as an interface to scholarly literature, evaluating responses in different contexts and domains. We discuss the implications and challenges of treating LLMs as “experts”.
@article{selby2024experts, title = {Had Enough of Experts? Elicitation and Evaluation of Bayesian Priors from LLMs}, author = {Selby, D. and Spriestersbach, K. and Iwashita, Y. and Bappert, D. and Warrier, Archana and Mukherjee, S. and Kise, K. and Vollmer, S.}, journal = {NeurIPS 2024 Workshop}, year = {2024} }