Hide Comments
Comments
You must login before you can post a comment.
Wednesday 30 November 17:00 - 18:00
Huxley Building
180 Queen's Gate
Clore lecture theatre
London
SW7 2AZ
Registration
Imperial College ICARL seminars - Alhussein Fawzi, DeepMind
Science & Technology
AlphaTensor: Discovering faster matrix multiplication algorithms with reinforcement learning
We are delighted to be hosting Alhussein Fawzi from DeepMind leading author of the recent breakthrough of AlphaTensor, an AI system for discovering novel, efficient, and exact algorithms for matrix multiplication.
The talk will be hybrid and you will be able to listen and ask questions both in person or in remote.
Abstract
Improving the efficiency of algorithms for fundamental computations can have a widespread impact, as it can affect the overall speed of a large amount of computations. Matrix multiplication is one such primitive task, occurring in many systems—from neural networks to scientific computing routines. The automatic discovery of algorithms using machine learning offers the prospect of reaching beyond human intuition and outperforming the current best human-designed algorithms. However, automating the algorithm discovery procedure is intricate, as the space of possible algorithms is enormous. Here we report a deep reinforcement learning approach based on AlphaZero1 for discovering efficient and provably correct algorithms for the multiplication of arbitrary matrices. Our agent, AlphaTensor, is trained to play a single-player game where the objective is finding tensor decompositions within a finite factor space. AlphaTensor discovered algorithms that outperform the state-of-the-art complexity for many matrix sizes. Particularly relevant is the case of 4 × 4 matrices in a finite field, where AlphaTensor’s algorithm improves on Strassen’s two-level algorithm for the first time, to our knowledge, since its discovery 50 years ago2. We further showcase the flexibility of AlphaTensor through different use-cases: algorithms with state-of-the-art complexity for structured matrix multiplication and improved practical efficiency by optimizing matrix multiplication for runtime on specific hardware. Our results highlight AlphaTensor’s ability to accelerate the process of algorithmic discovery on a range of problems, and to optimize for different criteria.
About the speaker
Alhussein is a Staff Research Scientist at DeepMind. He obtained his PhD from the signal processing laboratory in EPFL in 2016, and his MSc in Electrical Engineering from EPFL in 2012.
Sponsor
This event is sponsored by DeepMind
