Marta Skreta

Date:

Speaker

Marta is a PhD student in Computer Science at the University of Toronto and Vector Institute working with Alán Aspuru-Guzik under the Canada Graduate Scholarship. Her research focuses on molecular discovery using generative modelling, as well as automating chemistry experiments in self-driving labs. Previously, Marta completed internships at Apple and Mila AI for Humanity, and has also been invited as a speaker at Google Deepmind, CECAM, the Accelerate Conference, LoGG, and FutureHouse. Marta is a founding member of the AI4Materials workshop (NeurIPS 2022-24 and ICLR 2025) and the Frontiers in Probabilistic Learning workshop (ICLR 2025).

Speaker Links: Website - Google Scholar

Abstract

Automating chemistry experiments in self-driving labs is an exciting area but there remain many challenging problems we need to solve. Adding natural language interfaces to autonomous chemistry experiment systems lowers the barrier to using complicated robotics systems and increases utility for non-expert users, but translating natural language experiment descriptions from users into low-level robotics languages is nontrivial. Furthermore, while recent advances have used large language models to generate task plans, reliably executing those plans in the real world by an embodied agent remains challenging. To enable autonomous chemistry experiments, robots must interpret natural language commands, perceive the workspace, autonomously plan multi-step actions and motions, consider safety precautions, and interact with various laboratory equipment. In this talk, I will describe our progress towards creating a robot chemist, including a framework that can plan long-horizon chemistry experiments from natural language inputs and execute them on a real robot.

Papers covered during the talk

  • Davishi, K., Skreta, M., Zhao, Y., Yoshikawa, N., Som, S., Bogdanovic, M., Cao, Y., Hao, H., Xu, H., Aspuru-Guzik, A., & Sikurt, F. (2024). ORGANA: A robotic assistant for automated chemistry experimentation and characterization. Matter, 8(2), 101897. https://doi.org/10.1016/j.matt.2024.10.015 link - video

  • Skreta, M., Yoshikawa, N., Arellano-Rubach, S., Ji, Z., Kristensen, L. B., Davishi, K., Aspuru-Guzik, A., Shkurti, F., & Garg, A. (2023). Errors are useful prompts: Instruction guided task programming with verifier-assisted iterative prompting. arXiv preprint arXiv:2303.14100. link

  • Skreta, M., Zhou, Z., Yuan, J. L., Darvish, K., Aspuru-Guzik, A., & Garg, A. (2024). RePPlan: Robotic replanning with perception and language models. arXiv preprint arXiv:2401.04157. link