Search papers, labs, and topics across Lattice.
This paper introduces GeoCoupling, a novel framework that optimizes the temporal coupling of generative processes in multimodal biomolecular co-design, addressing the limitations of existing methods that enforce fixed synchronous coupling. By allowing for flexible temporal relationships between sequence and structure during training and generation, the authors mitigate issues related to high-variance supervision and inconsistent intermediate states. Empirical evaluations in structure-based drug design and protein design show that GeoCoupling significantly enhances the physical validity and diversity of generated biomolecules compared to traditional approaches.
Temporal coupling in multimodal biomolecular design can dramatically improve the quality and diversity of generated compounds, outperforming conventional methods.
Biomolecules such as proteins and small-molecule ligands play a central role in biological systems, arising from the tight interplay between sequence and three-dimensional structure. Recent generative models for biomolecular co-design aim to capture this interplay by jointly modeling coupled modalities. However, existing approaches largely adopt a parallel execution of marginal generative processes, implicitly enforcing fixed synchronous coupling. We argue that a critical but overlooked degree of freedom lies in how these marginal processes are temporally coupled during training and generation, where inappropriate coupling can introduce high-variance supervision and inconsistent intermediate states, affecting modality consistency. To address this, we introduce GeoCoupling, a systematic framework that optimizes for temporal couplings between heterogeneous modalities. Empirical results across structure-based drug design and unconditional protein design demonstrate the learned couplings consistently outperform synchronous and randomly coupled baselines, yielding biomolecules with improved physical validity and diversity.
Tsinghua AI