Publications
Research in geometric deep learning, foundation models, and AI for biology. Google Scholar ↗
Better Models, Faster Training: Sigmoid Attention for single-cell Foundation Models
STRAND: Sequence-Conditioned Transport for Single-Cell Perturbations
Geometric self-supervised pretraining on 3D protein structures using subgraphs
TEDDY: A Family Of Foundation Models For Understanding Single Cell Biology
E(n) Equivariant Topological Neural Networks
A Systematic Analysis of AI Trends in Environmental Research
ProCyon: A multimodal foundation model for protein phenotypes
ICML Topological Deep Learning Challenge 2024: Beyond the Graph Domain
TopoBench: A Framework for Benchmarking Topological Deep Learning
Learn to Code Sustainably: An Empirical Study on LLM-based Green Code Generation
Learn2Extend: Extending sequences by retaining their statistical properties with mixture models
Carbon footprint evaluation of code generation through LLM as a service
GNNDelete: A General Unlearning Strategy for Graph Neural Networks
Graph Ordering Attention Networks
Based on connections with the Partial Information Decomposition framework, we introduce a novel GNN layer, namely the Graph Ordering Attention (GOAT) that imposes neighborhood orderings according to the attention coefficients.
Multimodal learning with graphs
Permute Me Softly: Learning Soft Permutations for Graph Representations
We study how we can approximate graph distances by aligning adjacency matrices in a corpus of graphs. In order to allow the differentiable optimization, we suggest the utilization of soft permutation matrices.
Towards Expressive Graph Neural Networks: Theory, Algorithms, and Applications
Modularity-Aware Graph Autoencoders for Joint Community Detection and Link Prediction
Solving simultaneously link prediction and community detection is important in recommendation systems. Here, we show how we can extend the information that Graph Auto-encoders process towards this direction.
Ego-based Entropy Measures for Structural Representations on Graphs
Lipschitz Normalization for Self-Attention Layers with Application to Graph Neural Networks
We derive a theoretical analysis on the Lipschitz continuity of attention and show that enforcing Lipschitz continuity through normalization can significantly improve the performance of deep attention models.
Graph-based Neural Architecture Search with Operation Embeddings
We propose the replacement of fixed operator encoding in NAS problems with learnable representations in the optimization process.
Learning Parametrised Graph Shift Operators
We propose a parametrised graph shift operator (PGSO) to encode graphs, providing a unified view of common GSOs, and improve GNN performance by including the PGSO into the training in an end-to-end manner.
Coloring Graph Neural Networks for Node Disambiguation
Based on topological criteria and, specifically the separability, we introduce a universal approximation scheme of continuous functions on graphs. It is based on the disambiguation of identical node attributes.
Hcore-Init: Neural Network Initialization based on Graph Degeneracy
We propose a graph-based initialization of neural networks extending graph degeneracy observations. Such an initialization can encourage neurons that have structural importance in the neural network.
k-hop Graph Neural Networks
Standard GNNs use a 1-hop aggregation per layer, limiting their ability to capture graph properties. We iteratively extend the aggregation operator of graph neural networks to increase their receptive field.
No publications match your search. Try another keyword or category.