Scholar
Cheng-Hao Liu
Google Scholar ID: iVJGx0cAAAAJ
Caltech
Citations & Impact
Citations
1,756
H-index
20
i10-index
24
Publications
20
Co-authors
14
list available
Publications
Resume (English only)
Academic Achievements
- Several papers under review in 2025, including 'Navigating Ternary Doping in Li-ion Cathodes with Closed-loop Multi-objective Bayesian Optimization' and 'SynCoGen: Synthesizable 3D Molecule Generation via Joint Reaction and Coordinate Modeling'; Published 'Structural Control of Band Gap and Polaron Delocalization in 2D Azatriangulene Covalent Organic Frameworks' in J. Am. Chem. Soc. in 2025; 'Generating π-Functional Molecules Using STGG+ with Active Learning' was a Spotlight at the ICLR Machine Learning Multiscale Processes Workshop in 2025; Published 'Substrate-aware Zero-Shot Predictors for Non-Native Enzyme Activities' at the ICLR GEM Workshop in 2025; Published '[Invited News and Views] Bilayer Covalent Organic Frameworks Take a Twist' in Nat. Chem. in 2025; Published 'Steering Masked Discrete Diffusion Models via Discrete Denoising Posterior Prediction' at ICLR in 2025; Published 'Flow Factorization for Efficient Generative Flow Networks' at AAAI in 2024 (Oral presentation); Published 'RGFN: Synthesizable Molecular Generation Using GFlowNets' at NeurIPS in 2024.
Research Experience
- Co-founder of the protein design startup Dreamfold (2021-2025); Postdoctoral fellow at Caltech (2025-present); Organizer of the GEM workshop series.
Education
- BSc (Hons) from McGill University (2016-2019), supervised by Prof. Dmytro Perepichka; Ph.D. from McGill University (2019-2024), supervised by Prof. Dmytro Perepichka; concurrently studied at Mila with Prof. Yoshua Bengio; Postdoctoral fellow at Caltech (2025-present), supervised by Prof. Frances Arnold.
Background
- A student of chemistry and computer science. Broadly interested in understanding how uncharted physical properties emerge from molecular and supramolecular structures, and their design from the infinite chemical space. On the machine learning side, interested in generative modelling and active learning.
Miscellany
- Vanier scholar (2021-2024); Spent younger years in Germany, the Netherlands, and China.
