Selected work

Representative Papers

Training-free Graph-based Imputation of Missing Modalities in Multimodal Recommendation

Feb 19, 2026

This work addresses the challenge in multimodal recommendation systems where items lacking certain modalities—such as images or text—are often discarded, degrading overall performance. The paper formally characterizes this missing-modality problem and introduces a training-free graph propagation approach: it constructs an item co-purchase graph from user–item interactions and leverages graph signal interpolation to propagate available modality features to missing nodes. The method highlights the critical role of feature homophily over the item graph in enabling effective interpolation. It can be seamlessly integrated into existing recommender systems and consistently outperforms conventional imputation strategies across diverse missing-modality scenarios, while preserving—and often amplifying—the performance advantage of multimodal over unimodal recommendation models.

0 citationsRead paper

Flow-Based Fragment Identification via Binding Site-Specific Latent Representations

Sep 16, 2025

In fragment-based drug discovery, identifying initial weak-binding, low-specificity fragments remains a critical bottleneck. This paper introduces LatentFrag—the first protein-fragment joint modeling framework that integrates contrastive learning with conditional generation: it constructs a shared latent space to enable protein surface-guided, chemically valid fragment embedding learning and 3D pose generation. The method significantly improves binding site identification sensitivity and virtual screening (VS) efficiency, achieving state-of-the-art fragment recovery rates on standard benchmarks; it accelerates generation by two orders of magnitude over conventional VS approaches while drastically reducing computational cost. Furthermore, LatentFrag is extended to full ligand generation, demonstrating end-to-end de novo design capability.

0 citationsRead paper

Multi-domain Distribution Learning for De Novo Drug Design

Aug 25, 2025

This work addresses the challenge of 3D protein–ligand co-modeling in structure-based de novo drug design. We propose DrugFlow, a generative framework integrating continuous flow matching with discrete Markov bridges. DrugFlow jointly learns molecular chemical structures, 3D geometries, and physical interaction distributions; introduces uncertainty estimation—first in this domain—to reliably detect out-of-distribution samples; employs a joint preference alignment strategy to bias sampling toward regions with high binding affinity and drug-likeness; and extends to simultaneous ligand generation and protein side-chain conformational sampling, enabling exploration of the protein–ligand co-conformational space. On multiple benchmarks, DrugFlow achieves state-of-the-art performance, significantly improving generated molecules’ binding affinity, validity, and distributional fidelity.

0 citationsRead paper

Recent publications

Latest Papers

Training-free Graph-based Imputation of Missing Modalities in Multimodal Recommendation

Feb 19, 2026

This work addresses the challenge in multimodal recommendation systems where items lacking certain modalities—such as images or text—are often discarded, degrading overall performance. The paper formally characterizes this missing-modality problem and introduces a training-free graph propagation approach: it constructs an item co-purchase graph from user–item interactions and leverages graph signal interpolation to propagate available modality features to missing nodes. The method highlights the critical role of feature homophily over the item graph in enabling effective interpolation. It can be seamlessly integrated into existing recommender systems and consistently outperforms conventional imputation strategies across diverse missing-modality scenarios, while preserving—and often amplifying—the performance advantage of multimodal over unimodal recommendation models.

0 citationsRead paper

Flow-Based Fragment Identification via Binding Site-Specific Latent Representations

Sep 16, 2025

In fragment-based drug discovery, identifying initial weak-binding, low-specificity fragments remains a critical bottleneck. This paper introduces LatentFrag—the first protein-fragment joint modeling framework that integrates contrastive learning with conditional generation: it constructs a shared latent space to enable protein surface-guided, chemically valid fragment embedding learning and 3D pose generation. The method significantly improves binding site identification sensitivity and virtual screening (VS) efficiency, achieving state-of-the-art fragment recovery rates on standard benchmarks; it accelerates generation by two orders of magnitude over conventional VS approaches while drastically reducing computational cost. Furthermore, LatentFrag is extended to full ligand generation, demonstrating end-to-end de novo design capability.

0 citationsRead paper

Multi-domain Distribution Learning for De Novo Drug Design

Aug 25, 2025

This work addresses the challenge of 3D protein–ligand co-modeling in structure-based de novo drug design. We propose DrugFlow, a generative framework integrating continuous flow matching with discrete Markov bridges. DrugFlow jointly learns molecular chemical structures, 3D geometries, and physical interaction distributions; introduces uncertainty estimation—first in this domain—to reliably detect out-of-distribution samples; employs a joint preference alignment strategy to bias sampling toward regions with high binding affinity and drug-likeness; and extends to simultaneous ligand generation and protein side-chain conformational sampling, enabling exploration of the protein–ligand co-conformational space. On multiple benchmarks, DrugFlow achieves state-of-the-art performance, significantly improving generated molecules’ binding affinity, validity, and distributional fidelity.

0 citationsRead paper