Michael Bronstein – Medium

Michael Bronstein

11.7K Followers

Pinned

Published in
TDS Archive

The Road to Biology 2.0 Will Pass Through Black-Box Data

Future bio-AI breakthroughs will arise from novel high-throughput low-cost AI-specific “black-box” data modalities.

Mar 18, 2024

The Road to Biology 2.0 Will Pass Through Black-Box Data

Mar 18, 2024

Published in
TDS Archive

Co-operative Graph Neural Networks

A new message-passing paradigm where every node can choose to either ‘listen’, ‘broadcast’, ‘listen & broadcast’ or ‘isolate’.

Dec 6, 2023

Co-operative Graph Neural Networks

Dec 6, 2023

Published in
TDS Archive

Topological Generalisation with Advective Diffusion Transformers

A new diffusion-based continuous GNN model offers better generalisation capabilities

Oct 19, 2023

Topological Generalisation with Advective Diffusion Transformers

Oct 19, 2023

Published in
TDS Archive

Dynamically rewired delayed message passing GNNs

Dynamic rewiring and delayed message passing mechanisms offer a tradeoff between standard MPNNs and graph Transformers

Jun 19, 2023

Dynamically rewired delayed message passing GNNs

Jun 19, 2023

Published in
TDS Archive

Direction Improves Graph Learning

How a wise use of direction when doing message passing on heterophilic graphs can result in very significant gains.

Jun 8, 2023

Direction Improves Graph Learning

Jun 8, 2023

Published in
TDS Archive

Hyperbolic Deep Reinforcement Learning

Many RL problems have hierarchical tree-like nature. Hyperbolic geometry offers a powerful prior for such problems.

Apr 30, 2023

Hyperbolic Deep Reinforcement Learning

Apr 30, 2023

Published in
TDS Archive

Learning Network Games

How to learn the network underlying the interactions of players in social applications, economics, and beyond.

Apr 20, 2023

Learning Network Games

Apr 20, 2023

Published in
TDS Archive

Graph Neural Networks as gradient flows

GNNs derived as gradient flows minimising a learnable energy that describes attractive and repulsive forces between graph nodes.

Oct 14, 2022

Graph Neural Networks as gradient flows

Oct 14, 2022

Published in
TDS Archive

Towards Geometric Deep Learning IV: Chemical Precursors of GNNs

In the last post in the “Towards Geometric Deep Learning” series, we look at early prototypes of GNNs in the field of chemistry.

Jul 25, 2022

Towards Geometric Deep Learning IV: Chemical Precursors of GNNs

Jul 25, 2022

Published in
TDS Archive

Towards Geometric Deep Learning III: First Geometric Architectures

In the third post of our series “Towards Geometric Deep Learning” we look at the first “geometric” architectures: Neocognitron and CNNs

Jul 18, 2022

Towards Geometric Deep Learning III: First Geometric Architectures

Jul 18, 2022

Published in
TDS Archive

Towards Geometric Deep Learning II: The Perceptron Affair

In the second post of our series “Towards Geometric Deep Learning” we discuss how the criticism of Perceptrons led to geometric insights

Jul 11, 2022

Towards Geometric Deep Learning II: The Perceptron Affair

Jul 11, 2022

Published in
TDS Archive

Towards Geometric Deep Learning I: On the Shoulders of Giants

In a new series of posts, we discuss how geometric ideas of symmetry underpinning Geometric Deep Learning have emerged through history.

Jul 4, 2022

Towards Geometric Deep Learning I: On the Shoulders of Giants

Jul 4, 2022

Published in
TDS Archive

Accelerating and scaling Temporal Graph Networks on the Graphcore IPU

Is GPU the best hardware choice for GNNs? Together with Graphcore, we explore the advantages of the new IPU architecture for temporal GNNs.

Jun 14, 2022

Accelerating and scaling Temporal Graph Networks on the Graphcore IPU

Jun 14, 2022

Published in
TDS Archive

A new computational fabric for Graph Neural Networks

Are graphs the right computational fabric for GNNs? A recent line of papers challenges this assumption.

Jun 10, 2022

A new computational fabric for Graph Neural Networks

Jun 10, 2022

Published in
TDS Archive

Neural Sheaf Diffusion for deep learning on graphs

Cellular sheaf theory, a branch of algebraic topology, provides new insights into how Graph Neural Networks work and how to design new…

May 16, 2022

Neural Sheaf Diffusion for deep learning on graphs

May 16, 2022

Published in
TDS Archive

Announcing the Learning on Graphs Conference

Graph Machine Learning has become large enough of a field to deserve its own standalone event: the Learning on Graphs Conference (LoG).

Apr 15, 2022

Announcing the Learning on Graphs Conference

Apr 15, 2022

Published in
TDS Archive

Graph Neural Networks beyond Weisfeiler-Lehman and vanilla Message Passing

Physics-inspired continuous learning models on graphs allow to overcome the limitations of traditional GNNs

Mar 3, 2022

Graph Neural Networks beyond Weisfeiler-Lehman and vanilla Message Passing

Mar 3, 2022

Published in
TDS Archive

Learning on graphs with missing features

Feature Propagation is a simple and surprisingly efficient solution for learning on graphs with missing node features

Feb 3, 2022

Learning on graphs with missing features

Feb 3, 2022

Published in
TDS Archive

Predictions and hopes for Geometric & Graph ML in 2022

Leading researchers in Geometric & Graph ML summarise the progress in 2021 and make predictions for 2022

Jan 24, 2022

Predictions and hopes for Geometric & Graph ML in 2022

Jan 24, 2022

Published in
TDS Archive

Using subgraphs for more expressive GNNs

GNNs have limited expressive power due to their equivalence to WL test. Recent works show how to improve expressivity by using subgraphs.

Dec 20, 2021

Using subgraphs for more expressive GNNs

Dec 20, 2021

Michael Bronstein

Michael Bronstein

11.7K Followers

DeepMind Professor of AI @Oxford. Serial startupper. ML for graphs, biochemistry, drug design, and animal communication.

Following

Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams