Pheng Ann Heng

JEPA-Anything, a domain-agnostic framework based on orthogonal predictive factorization (OPF), achieves the lowest one-step and 100-step molecular errors among compared methods in all four systems and supports a common factorized predictive principle across heterogeneous worlds.

Orthogonal JEPA achieves superior predictive performance by factorizing latent states, allowing for more efficient learning in complex systems.

Operation laundering in vision encoders can be effectively mitigated, revealing hidden boundaries in learned assignments that traditional methods obscure.

Subsampling MSAs based on energetic frustration can dramatically enhance the recovery of alternative protein conformations, outperforming traditional sequence-based methods.

AlloGen reveals that conformational selectivity in protein binder design can be learned and optimized, not just assumed.

Reward models optimized for single-step generation can fail spectacularly when integrated into multi-stage LLM pipelines, but pipeline-aware training can fix this.

Chemical modifications to proteins, often missed by standard protein language models, can now be accurately captured and leveraged for improved biomolecule design, thanks to a new framework that fuses atomic-level details with broader sequence context.