Stories by Michael Bronstein on Medium

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

Michael Bronstein — Mon, 18 Mar 2024 09:15:32 GMT

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

Michael Bronstein — Wed, 06 Dec 2023 01:35:56 GMT

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

Michael Bronstein — Thu, 19 Oct 2023 14:46:22 GMT

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

Michael Bronstein — Mon, 19 Jun 2023 14:07:06 GMT

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

Michael Bronstein — Thu, 08 Jun 2023 13:59:04 GMT

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

Michael Bronstein — Sun, 30 Apr 2023 12:03:49 GMT

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

Michael Bronstein — Thu, 20 Apr 2023 10:27:15 GMT

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

Michael Bronstein — Fri, 14 Oct 2022 14:29:52 GMT

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

Michael Bronstein — Mon, 25 Jul 2022 07:16:44 GMT

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

Michael Bronstein — Mon, 18 Jul 2022 08:21:20 GMT

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