Institution profile

Max Planck Institute of Biochemistry

Academic institutioneurope · de
Official website
Research library8linked papers
Opportunities0open roles
Selected work

Representative Papers

From Persistence to Survival: Hypothesis Testing, Effect Sizes and Vectorisation for Topological Features

Jun 10, 2026

Persistent diagrams lack a natural vector space structure, and existing statistical methods struggle to integrate effectively with downstream predictive tasks. This work addresses these limitations by treating persistent diagrams as survival data and introduces a unified framework based on persistence survival functions. For the first time, this approach simultaneously enables hypothesis testing, effect size quantification, and 1-Wasserstein stable vectorization from a single interpretable representation. By integrating survival analysis with nonparametric two-sample testing, the method demonstrates well-calibrated type I error control and high statistical power on synthetic manifolds. It achieves strong performance across 14 benchmark tasks involving graphs and 3D point clouds and is successfully applied to fMRI-based brain functional connectivity analysis.

0 citationsRead paper

End-to-End Subgraph Detection with GraphDETR

Jun 04, 2026

This work addresses the scalability bottleneck in subgraph pattern detection within large-scale graphs, a challenge rooted in the NP-completeness of the problem, by introducing the DETR paradigm to this task for the first time. The proposed method formulates subgraph detection as a set prediction problem, leveraging a graph neural network to encode the target graph, learnable query embeddings, and a Transformer decoder to jointly predict all pattern instances in an end-to-end manner via bipartite matching loss. This framework supports both exact and approximate pattern matching, thereby overcoming the limitation of traditional approaches that are restricted to exact structural matches. Experiments demonstrate that the method efficiently detects diverse patterns of up to 50 nodes in graphs containing 1,000 nodes, achieving an AP₁₀₀ of 91.2 on functional group detection in the ChEMBL molecular dataset.

0 citationsRead paper
Recent publications

Latest Papers

From Persistence to Survival: Hypothesis Testing, Effect Sizes and Vectorisation for Topological Features

Jun 10, 2026

Persistent diagrams lack a natural vector space structure, and existing statistical methods struggle to integrate effectively with downstream predictive tasks. This work addresses these limitations by treating persistent diagrams as survival data and introduces a unified framework based on persistence survival functions. For the first time, this approach simultaneously enables hypothesis testing, effect size quantification, and 1-Wasserstein stable vectorization from a single interpretable representation. By integrating survival analysis with nonparametric two-sample testing, the method demonstrates well-calibrated type I error control and high statistical power on synthetic manifolds. It achieves strong performance across 14 benchmark tasks involving graphs and 3D point clouds and is successfully applied to fMRI-based brain functional connectivity analysis.

0 citationsRead paper

End-to-End Subgraph Detection with GraphDETR

Jun 04, 2026

This work addresses the scalability bottleneck in subgraph pattern detection within large-scale graphs, a challenge rooted in the NP-completeness of the problem, by introducing the DETR paradigm to this task for the first time. The proposed method formulates subgraph detection as a set prediction problem, leveraging a graph neural network to encode the target graph, learnable query embeddings, and a Transformer decoder to jointly predict all pattern instances in an end-to-end manner via bipartite matching loss. This framework supports both exact and approximate pattern matching, thereby overcoming the limitation of traditional approaches that are restricted to exact structural matches. Experiments demonstrate that the method efficiently detects diverse patterns of up to 50 nodes in graphs containing 1,000 nodes, achieving an AP₁₀₀ of 91.2 on functional group detection in the ChEMBL molecular dataset.

0 citationsRead paper