Bella (Chih-Hsuan) Yang

Bella (Chih-Hsuan) Yang

Position
  • Masters Thesis
I am a Ph.D. student in Mechanical Engineering and Computer Engineering, as well as an MS student in Computer Science. With over 5 years of experience as a Visiting Scientist, I specialize in Language Models, multi-modality models, Vision-Language Models (VLM), multi-fidelity models, and Generative AI. I led a team of 9 Ph.D. students in a project that earned a highlight acceptance at the 2024 NeurIPS conference. This work produced the world’s largest AI-ready dataset for biodiversity, named Arboretum, and improved OpenAI’s CLIP model accuracy from 7.3% to 91.1% on the benchmark dataset. During my tenure at Google X, I developed integration pipelines for specialized LLM agents, engineered self-improving AI pipelines, and initiated two patents in multi-agent LLM systems. Additionally, I created a Python package for Generative AI in molecular discovery, significantly boosting model accuracy and earning multiple presentation awards. For more details, please visit my LinkedIn profile by searching my name.

Area of Expertise

  • Language Models
  • Generative AI
  • Multi-Modality Models

Education

  • B.S., Bioenvironmental System Engineering, National Taiwan University (NTU)

Publications

  1. Arboretum: A Large Multimodal Dataset Enabling AI for Biodiversity, [Submitted to NeurIPS]: Yang, C. H., Feuer, B., Jubery,... & Baskar, G.
  2. Generative AI for Advanced Prediction and Exploration of Large Organic Molecules, [Proceeding to Royal Society of Chemistry]: Yang, C. H.,.. & Baskar, G
  3. Multi-fidelity machine learning models for structure-property mapping of organic electronics. Computational Materials Science: Yang, C. H., Pokuri, B. S. S., Lee, X. Y..& Baskar, G
  4. Fast inverse design of microstructures via generative invariance networks. Nature Computational Science: Lee, X. Y., Waite, J. R., Yang, C. H.., ... & Sarkar, S.
  5. Neural PDE Solvers for Irregular Domains, Computer-Aided Design : Khara, B.,., ... Yang, C. H., .& Baskar, G. • Physics-Informed Neural Networks for Extracting Phenomenological Models from Molecular Dynamics Simulations. National Congress on Computational Mechanics: Yang, C. H., Pokuri, B.Bazant, ..& Baskar, G.