PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
February 27, 20260 citationsOpen Access

Dynamic Vector Networks: Self-Organizing Knowledge Structures Beyond Transformers

View Full Paper
ACAdrien CrosAAva

Key Points

  • This research aims to introduce Dynamic Vector Networks, a novel architecture for knowledge representation that differs from traditional Transformers.
  • Introduced an architecture with rich embedding vectors instead of scalar neurons.
  • Used Hebbian learning for real-time connection adjustments between nodes.
  • Enabled dynamic creation of new nodes for representing novel concepts.
  • Implemented self-organization based on semantic similarity.
  • Demonstrated enhanced flexibility and adaptability compared to existing systems.
  • Showed potential for improved efficiency in knowledge representation and learning.
  • Established a framework for understanding complex relationships in data.

Abstract

We introduce Dynamic Vector Networks (DVN), an architecture fundamentally different from Transformers: each node is a rich embedding vector (not a scalar neuron), connections form through Hebbian learning in real-time, and new nodes are spawned dynamically for novel concepts. The architecture combines four properties no existing system offers simultaneously: rich vector nodes, real-time weight evolution, dynamic node creation, and self-organization by semantic similarity. Also available in French: Dynamic Vector Networks : Au-delà des Transformers

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Cros et al. (2026) studied this question.

synapsesocial.com/papers/69a1359eed1d949a99abfae6https://doi.org/10.5281/zenodo.18770882
Ask AI
Helpful
Bookmark
Share
View Full Paper

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1DTFormer: A Transformer-Based Method for Discrete-Time Dynamic Graph Representation Learning2024
  2. 2The Dynamic Net Architecture: Learning Robust and Holistic Visual Representations Through Self-Organizing Networks2024
  3. 3VideoGNN: Video Representation Learning via Dynamic Graph Modelling2025
  4. 4Learning Dynamic Graphs via Tensorized and Lightweight Graph Convolutional Networks2025
  5. 5Dynamic Neural Dowker Network: Approximating Persistent Homology in Dynamic Directed Graphs2024