[Ssnet_list] "Bringing serendipity back into the scientific process" | Maxim Ziatdinov | Nov. 19, 2025 11 am

[Monika Sengul-Jones via Ssnet_list]

*Apologies for cross-posting.*

Hey everyone,

On behalf of Kelly Olenyik (IPhD, UW Seattle), I'm letting you know about a
talk on Wednesday, Nov. 19 from 11:00 AM - 12:00 PM in Gould Hall, Room
322, on how to set up a "modern autonomous [scientific] laboratory" using
an open-source, multi-agent AI framework that, to a stranger, could sound
like a tutorial on autonomous robot debate matches. Of course, the vantage
of the stranger can be illuminating. You may also find that learning to use
AI to speed up the research process—and to research gaps more
creatively—will spark your own creativity. Details below (you can email
Kelly at olen at uw.edu to learn more and access the paper).

Yours,
Monika Sengul-Jones
www.societyandtechnology.uw.edu

Speaking Session (UW community welcome) AI Core Seminar | UW Molecular
Engineering Materials Center | Dr. Maxim Ziatdinov, Wednesday, Nov. 19,
2025 11:00 AM – 12:00 PM, Gould Hall, GLD 322

In this talk, Maxim will introduceSciLink, an open-source, multi-agent AI
framework designed to bring serendipity back into the scientific process.
SciLink connects experimental data, novelty detection, and theoretical
modeling in a fully automated loop. It uses a hybrid AI approach: machine
learning models handle the quantitative analysis, while large language
models take care of higher-level reasoning. Together, these agents turn raw
microscopy and spectroscopy data into testable scientific claims—and then
score those claims for novelty by comparing them to the existing
literature. Maxim will walk through examples of SciLink in action, from
atomic-resolution imaging to hyperspectral datasets, and show how it can
even incorporate real-time input from human experts. His talk will
highlight how this system doesn’t just make research faster—it makes it
more open-ended, more creative, and more likely to uncover the unexpected.

About the speaker
Dr. Maxim Ziatdinov is a senior research scientist at Pacific Northwest
National Laboratory (PNNL), where he leads cutting-edge efforts at the
intersection of artificial intelligence and materials science. With a Ph.D.
in engineering science from Tokyo Institute of Technology, Dr. Ziatdinov
has spent over a decade advancing autonomous experimentation and machine
learning-driven discovery in microscopy, spectroscopy, and chemical
synthesis. His work focuses on integrating domain-informed AI models into
scientific instrumentation to enable real-time data analysis and
closed-loop control. Dr. Ziatdinov is the creator of several influential
opensource software tools—includingAtomAI,GPax, andpyroVED— that have
accelerated progress in deep learning, Bayesian optimization, and invariant
representation learning for physical sciences. His contributions have been
recognized through numerous high-impact publications and a U.S. patent for
AI-guided experimentation. At the forefront of human-AI collaboration, Dr.
Ziatdinov is passionate about building reproducible, intelligent systems
that transform how science is conducted.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman23.u.washington.edu/pipermail/ssnet_list/attachments/20251117/6adf57f0/attachment.html>