M3RG@IITD

Multiphysics and Multiscale Mechanics Research Group, @iitdelhi. Group leader: N. M. Anoop Krishnan @anoopnm007

[N M Anoop Krishnan @anoopnm007]

Is AI4Science stuck in a local minimum? 🤔

This question has been at the core of the discussions between @kmjablonka and me for months, and the highlights are reflected in our comment “Real AI advances require collaboration” in @NatRevChem 1/n
https://t.co/lhH8GVb3MJ

[Jablonka Lab (Lab for AI for Materials) @jablonkagroup]

🚀Our revised MaCBench paper is now on arxiv! https://t.co/WAjAnTW4di
Key updates!
🌟Robust reproducibility: 5x experiment runs + error bars for statistical confidence
🌟Full dataset & leaderboard: Now live on HuggingFace with model comparisons https://t.co/8hmyGqsadc

[Alexandre Duval @ADuvalinho]

📢@entalpic_ai is opening a new position: Material Scientist, ML for material synthesis prediction - to join our team in Paris ! The role involves bridging the gap between simulations and experimental data!

Please share this thread & reach out if you are interested

[N M Anoop Krishnan @anoopnm007]

Also, if you are excited about these areas, we have several PhD/postdoc positions. Do write to me if you are excited to join our interdisciplinary group!
@iitdelhi

[N M Anoop Krishnan @anoopnm007]

With these advances, we hope to develop tools that address crucial future challenges regarding energy, health care, agriculture, education, and sustainability.

Do reach out to us if you have any comments/feedback!

[N M Anoop Krishnan @anoopnm007]

These works also present several aspects of automating materials research and discovery that we have been pursuing regarding information extraction, in silico evaluation, and high throughput experiments.

[N M Anoop Krishnan @anoopnm007]

5. LLaMaT: Large language model for materials capable of being a research copilot and crystal generation. With Vaibhav Mishra, Somaditya Singh, Dhruv Ahlawat, MOHD ZAKI, Vaibhav Bihani, Hargun Singh Grover, Biswajit Mishra, Santiago Miret, Mausam .
Paper: https://t.co/luZXNvSwp2

[N M Anoop Krishnan @anoopnm007]

4. PeGaMaT: Large-scale composition and property extraction from materials tables. With Kausik Hira, MOHD ZAKI, Mausam .
Workshop paper: https://t.co/vXVw2jdvEK

[N M Anoop Krishnan @anoopnm007]

3. Automated Microscopy with LLM agents: Evaluate the performance of LLM agents to perform real-world experiments. With Indrajeet Mandal, Jitendra Soni, MOHD ZAKI, Morten Mattrup Smedskjaer, Katrin Wondraczek, Lothar Wondraczek, Nitya Nand Gosvami.
Link: https://t.co/1Ob4mZOkmG

[N M Anoop Krishnan @anoopnm007]

2. Differentiable simulation: Optimizing force fields directly towards target property through end-to-end differentiable simulation. With Abhijeet Gangan, Samuel Schoenholz, Ekin Dogus Cubuk, Mathieu Bauchy
Full paper: https://t.co/Q0FwqJXPe4

[N M Anoop Krishnan @anoopnm007]

1. MaCBench: Evaluation of multimodal LLMs' chemistry and materials knowledge such as Gemini, Claude, GPT4V, and LLaVa. With Nawaf Alampara, Mara Schilling-Wilhelmi, Martino Rios-Garcia, Indrajeet Mandal, Pranav Khetarpal, Hargun Singh Grover. Paper: https://t.co/DN5PrMs93t

[N M Anoop Krishnan @anoopnm007]

Are you @NeurIPSConf? Are you excited about AI4Materials?

Do drop by our workshop, AI4Mat, to engage with the community. Check out the exciting schedule!

Although I'll be missing it (thanks to the delay due to my visa), we will present five accepted works as follows.

[N M Anoop Krishnan @anoopnm007]

🔬 Large Language Models for Materials! 🌟

Can Large Language Models (LLMs) help accelerate materials discovery?

Introducing LLaMaT, a family of foundational LLM models for materials research and crystal generation! 🚀

https://t.co/pwD14jCtGM

[Sayan Ranu @SayanRanu]

Excited to announce that FUGAL (Feature-fortified Unrestricted Graph Alignment) will be presented at @NeurIPSConf today!

📍 West Ballroom A-D, Booth #5900
🗓️ Dec 11th, 4:30 p.m. PST — 7:30 p.m.
Let's chat if you're at NeurIPS! 💬

@VNSAditya2002 @DavideMottin
#NeurIPS2024

[Jablonka Lab (Lab for AI for Materials) @jablonkagroup]

Are Vision Language Models ready for scientific research?🔭👩‍🔬

We compared leading VLLMs on the three pillars of chemical and material science discovery: data extraction, lab experimentation and data interpretation.
https://t.co/uNJPTyBOV5

[N M Anoop Krishnan @anoopnm007]

Travelling to Las Vegas for @ACerSNews, GOMD 2024! There are seven presentations from our group given by @zaki_here_, @SajidMannan4, @Vaibhav__Bihani, Sourav Sahoo, and Shweta Rani Keshri! Do drop a message if you are interested to chat!

#gomd24

[N M Anoop Krishnan @anoopnm007]

Which machine learning framework (or neural operator) is better for learning parametric differential equations?

We answer this question through CoDBench: https://t.co/JBtHkaDJvK

@digital_rsc

[N M Anoop Krishnan @anoopnm007]

If you are interested in writing a book (original or edited), do reach out! We are looking for contributions in the areas of LLMs for materials, AI-driven simulations, AI and experiments to name a few.

Looking forward to excellent contributions!

[N M Anoop Krishnan @anoopnm007]

Finally, this book also marks my beginning as the series editor for a springer book series on “Machine intelligence for materials science” with Zach Evenson.

https://t.co/oVxjWbFuD6

[N M Anoop Krishnan @anoopnm007]

Want to learn machine learning for materials domain or other application domains?!

Our (@haribamsuri and @RavinderBhattoo) book titled “Machine learning for materials discovery” is out! Do check it out!
https://t.co/6lDSVajrrW

Special thanks to @ProfBuehlerMIT for the preface!