Publications
Publications, preprints, tutorials, and conference abstracts on generative modeling, model steering and reliability, agent evaluation, and AI for science. For citation metrics and the latest indexing record, see Google Scholar.
Persistent Activation Steering Disrupts Language-Model Termination: A Failure Mode Distinct from Quality Degradation
Under review.
Emergent Steering Beyond Endpoint Alignment in Chemical Reaction Models
Under review.
Driving Reaction Trajectories via Latent Flow Matching
KDD 2026.
AgentTrap: Measuring Runtime Trust Failures in Third-Party Agent Skills
arXiv preprint.
SenseMath: Do LLMs Have Number Sense? Evaluating Shortcut Use, Judgment, and Generation
arXiv preprint.
Reward Transport: Property Control in Flow Matching via Noise-Space Alignment
arXiv preprint.
Transferable FB-GNN-MBE Framework for Potential Energy Surfaces: Data-Adaptive Transfer Learning in Deep Learned Many-Body Expansion Theory
The Journal of Chemical Physics, 2026.
Organic Chemistry as a Catalyst for AI Innovation: Challenges, Methods, and Emerging Paradigms
Chemical Reviews, 126(13), 7587–7635, 2026.
From VAEs to Diffusion and LLMs: Modern Generative Models for Molecular Discovery
KDD 2026 Tutorial.
Towards Few-shot Chemical Reaction Outcome Prediction
CIKM 2025, 2599–2609.
Artificial Intelligence in Spectroscopy: Advancing Chemistry from Prediction to Generation and Beyond
IJCAI 2025 Survey Track, 10445–10454.
Accurate Electronic and Optical Properties of Organic Doublet Radicals Using Machine Learned Range-Separated Functionals
The Journal of Physical Chemistry A, 128(12), 2457–2471, 2024.
Integrating Graph Neural Networks and Many-Body Expansion Theory for Potential Energy Surfaces
NeurIPS 2024 AI4Mat Spotlight.
Improving Subgraph Representation Learning via Multi-View Augmentation
ICML 2022 AI4Science.
TCR: A Transformer-Based Deep Network for Predicting Cancer Drug Response
arXiv preprint.
Accelerating Many-Body Expansion Theory Through Graph Convolutional Networks
75th International Symposium on Molecular Spectroscopy, 2022.
Quality Prediction of Discrete Manufacturing Process Based on CGAN & CatBoost Hybrid Model
Journal of Physics: Conference Series, 1757(1), 012072, 2021.