Nikolaos Nakis

§ 04 · Recent updates

News & Updates

A timeline of recent paper acceptances, awards, and research milestones.

2026 Recognized as a Gold Reviewer at ICML 2026 — among the top reviewers for the conference.
2026 Two papers accepted at ICML 2026 (Seoul): Aitchison Embeddings for Learning Compositional Graph Representations, which learns interpretable, role-mixture node embeddings grounded in Aitchison geometry; and InvGNN: Learning Invertible Node Representations on Graphs, which introduces fully invertible GNN layers for density estimation and generative modeling.
2026 Modeling roles and trade-offs in multiplex networks published in Nature Communications — a framework for recovering latent social roles and their trade-offs across multiplex interaction layers.
2026 Archetypal Graph Generative Models accepted at AISTATS 2026 as a Spotlight — explainable and identifiable community structure via anchor-dominant convex hulls.
2025 How Low Can You Go? Searching for the Intrinsic Dimensionality of Complex Networks using Metric Node Embeddings accepted at ICLR 2025 — probing how few dimensions a network embedding really needs.
2025 Signed Graph Autoencoder for Explainable and Polarization-Aware Network Embeddings accepted at AISTATS 2025 — autoencoder embeddings that surface polarization structure in signed networks.
2025 The signed two-space proximity model for learning representations in protein–protein interaction networks published in OUP Bioinformatics — a signed latent-space model for protein–protein interaction graphs.
2024 Continuous-time Graph Representation with Sequential Survival Process accepted at AAAI 2024 — continuous-time node embeddings derived from a sequential survival formulation.
2024 Time to Cite: Modeling Citation Networks using the Dynamic Impact Single-Event Embedding Model accepted at AISTATS 2024 — a dynamic single-event embedding model for citation dynamics.
2023 Best Paper Award at the Temporal Graph Learning Workshop, NeurIPS 2023.
2023 A Hierarchical Block Distance Model for Ultra Low-Dimensional Graph Representations accepted at IEEE TKDE — ultra low-dimensional embeddings via a hierarchical block-distance formulation.
2023 Characterizing Polarization in Social Networks using the Signed Relational Latent Distance Model accepted at AISTATS 2023 — characterizing polarization in social networks via signed latent-distance models.