www.science.org/doi/10.1126/... represents a step toward advancing our understanding of complex systems and developing new tools for AI and biological intelligence. Thank you very much to National Science Foundation (NSF) and University of Michigan COMPASS: Center for Complex Particle Systems!

Have been cooking with so many amazing professionals, Seongheon Park, @teknology.bsky.social , Jiatong Li, Wendi Li, Samuel Yeh, Xuefeng Du, Hamed Hassani, @paulbogdan.bsky.social , Dawn Song, and Sharon Li🥰 Truly enjoyed it and learned a lot🙏

Graph theory and multifractal graph analysis can explore the thermodynamic parameters among elements, identify possible favorable (miscible) pairs and their congeners and enable new artificial intelligence frameworks for engineering complexity and material discovery. doi.org/10.1002/advs...

🎉 Excited to share our NeurIPS 2025 paper on circuit representation learning! MIHC bridges topological netlist and geometric layout information for chip design tasks. An important step toward trustworthy AI in Electronic Design Automation. 📄 Read the full paper: openreview.net/pdf?id=YFQ0X...

Thank you COMPASS: compass.engin.umich.edu/spiking-dyna...

Fundamental to our existence, agriculture relies on consistent seed flow for seeding and seed collection. Over-reliance on toxic solid lubricants harms our ecosystems, pollinators and all of us. authors.elsevier.com/c/1lurr_wvIm... shows graph theory enable the design of benign solid lubricant !

Faced with brain sensing limitations, can we + Decipher the networks from neuronal spiking behavior? + Infer neuronal networks that are resilient to variations in circuit inputs? + Distinguish networks of cognitive tasks? Congratulations to Ruochen Yang, Roozbeh Kiani, Heng Ping and Xiongye Xiao !

How can we control non-Markovian dynamical systems? This ICML 2025 paper icml.cc/virtual/2025... or arxiv.org/abs/2502.04649 introduces a novel end-to-end machine learning framework for solving non-Markovian optimal control problems.