Acceleration Consortium

2024 has been a year of momentum for us at the Acceleration Consortium, thanks in great part to the $200 million grant we received from the Canada First Research Excellence Fund (TIPS-SPIIE). Since then, we’ve been busy scaling. Take a peek at what we've been up to and what we have planned for 2025. Wishing everyone a happy holiday season and all the best for the year ahead!

#JOB OPPORTUNITIES: We’re hiring across a range of areas at the Acceleration Consortium at the University of Toronto, including lab technician, knowledge mobilization and postdoctoral fellowships. It's a great time to join our team and be part of our historic $200M grant from the Canada First Research Excellence Fund (TIPS-SPIIE). Apply now to support our work in #AI for science: https://lnkd.in/gQP9cQpe Lab Technician (Term)  ➡️ https://lnkd.in/gTbYfxRQ    Knowledge Mobilization Lead (2 Year Term)  ➡️ https://lnkd.in/gSb6qSTY    Postdoctoral Fellowships in Self-Driving Labs for Material Synthesis and Characterization  ➡️ https://lnkd.in/gE6eam4K    Postdoctoral Fellowship in AI-powered mapping of multiscale, multicellular human models  ➡️ https://lnkd.in/gep7yi7C #Careers #ScienceCareers

Save the date for our 2025 Accelerate conference in Toronto, August 11–14! Join a global community pushing the boundaries of #AI for science for 4 days of talks, workshops, and more. To learn about registration and sponsorship opportunities, visit 2025.accelerateconf.ca.

Congratulations to the nearly 50 University of Toronto researchers awarded Canada Research Chairs, including the AC’s scientific leadership team member Helen Tran from the Department of Chemistry at the University of Toronto and Accelerate Moonshot grant holder Yu Zou in the department of Materials Science & Engineering - University of Toronto!   https://lnkd.in/d9_zQQBi

GRANT OPPORTUNITY: AI and automation are changing the way we do science, but what are the ethical, legal, or economic consequences? If you’re a social science or humanities researcher at the University of Toronto interested in exploring these topics, apply for our new grant, supported by the Canada First Research Excellence Fund (TIPS-SPIIE) NOI deadline: February 14, 2025! Learn more: https://lnkd.in/gdU9rQ9K And thanks to #UofT faculty for helping us shape this grant through your feedback on interdisciplinary research. #AIResearch #FundingOpportunity

"At six, my parents gave me an encyclopedia. Not thinking it unusual, I read it cover to cover twice during that year." Congrats to AC director Alán Aspuru-Guzik for being part of Cell Press' 50 scientists that inspire. Learn about what led Alán to his career path, what he's learned along the way, and how he sees science changing in the next 50 years. ↗️ https://bit.ly/3DcVydG

Today, the AC is excited to announce that we're joining The AI Alliance, a global network co-founded by IBM and Meta that is shaping the future of open, safe, and responsible AI. Together we will develop trusted open-source AI tools and foundation models to accelerate scientific discovery. From climate change to cancer, new materials and molecules are needed to quickly but safely address the world’s most pressing problems. Learn more ↗️ https://lnkd.in/gcE2pwCU

📅 When: Dec 9, 1-2PM 📍 Where: 700 University Ave. 10th Floor, Room 10031 Join AC member Tonio Buonassisi for a talk on minimal material representations in #machinelearning for materials science. Buonassisi is a Professor of Mechanical Engineering at the Massachusetts Institute of Technology. He is pioneering the application of artificial intelligence to develop new materials for societally beneficial applications. The AC is delighted to support this talk, as part of Materials Science & Engineering - University of Toronto's distinguished seminar series. About the talk: Machine learning models have become indispensable tools in computational materials science, enabling rapid and accurate predictions of material properties and facilitating the inverse design of new materials. Surprisingly, these models often achieve exceptional performance using minimal or "poor" physical representations, lacking explicit structural details traditionally deemed essential. This counterintuitive success highlights an apparent disconnect between conventional physical understanding and the data-driven approaches of machine learning. In this work, we investigate what information is necessary and sufficient to include in a material's representation to accurately predict its properties. By analyzing whether structural information is implicitly encoded in compositional data and how machine learning models extract and utilize this information, we aim to reconcile this apparent fracture. Our findings shed light on designing more effective material representations for machine learning, potentially simplifying models while maintaining or enhancing predictive capabilities.

AC staff scientist Kangming Li and AC Scientific leadership team member Jason Hattrick-Simpers present their findings in their new paper, Evaluating the Performance and Robustness of LLMs in Materials Science Q&A and Property Predictions. Large Language Models (LLMs) have the potential to revolutionize scientific research, yet their robustness and reliability in materials science applications remain underexplored. This work assesses the performance of #LLMs on the prediction of materials properties, and test the robustness of these predictions against various forms of dataset noise to evaluate their resilience and reliability under real-world conditions. The study uncovers unique phenomena of LLMs during predictive tasks, such as failure to explore the entire distribution of the training data when the prompt examples are altered and performance enhancement from train/test mismatch. The findings aim to provide informed skepticism for the broad use of LLMs in materials science and to inspire advancements that enhance their robustness and reliability for practical applications, e.g. in self-driving materials discovery that AC is focusing on. https://lnkd.in/gtewkym6 #LargeLanguageModels #AI #MaterialsScience #SelfDrivingLabs #ScientificResearch #MachineLearning

Congrats to our friends The AI Alliance on their 1st anniversary! Looking forward to joining the party...stay tuned! 👀