Actor–critic networks with analogue memristors mimicking reward-based learning
Kevin Portner, Till Zellweger, Flavio Martinelli, Laura Bégon-Lours, Valeria Bragaglia, Christoph Weilenmann, Daniel Jubin, Donato Francesco Falcone, Felix Hermann, Oscar Hrynkevych, Tommaso Stecconi, Antonio La Porta, Ute Drechsler, Antonis Olziersky, Bert Jan Offrein, Wulfram Gerstner, Mathieu Lui · 2025
Read Actor–critic networks with analogue memristors mimicking reward-based learning by Kevin Portner, Till Zellweger, Flavio Martinelli, Laura Bégon-Lours, Valeria Bragaglia, Christoph Weilenmann, Daniel Jubin, Donato Francesco Falcone, Felix Hermann, Oscar Hrynkevych, Tommaso Stecconi, Antonio La Porta, Ute Drechsler, Antonis Olziersky, Bert Jan Offrein, Wulfram Gerstner, Mathieu Lui (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-12719229.txt.
Description
Portner and a large team build a learning network from analogue memory devices in twenty twenty-five. The devices, called memristors, hold and change values like the links between brain cells. Most past designs used them for just one part of a learning task. Here they store the link weights, work out updates, and pick the next move.
The network learns by reward, much like animals learn from treats and failures. It steered through two maze tasks, one on a grid and one water maze. All training took place inside the memory itself, with no data shuttled in and out. The team calls it a first step toward full in memory learning machines.
Catalog record
| CREATOR | Kevin Portner, Till Zellweger, Flavio Martinelli, Laura Bégon-Lours, Valeria Bragaglia, Christoph Weilenmann, Daniel Jubin, Donato Francesco Falcone, Felix Hermann, Oscar Hrynkevych, Tommaso Stecconi, Antonio La Porta, Ute Drechsler, Antonis Olziersky, Bert Jan Offrein, Wulfram Gerstner, Mathieu Lui |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 76 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |