We applied AlphaEvolve to a fundamental problem in computer science: discovering algorithms for matrix multiplication. It managed to identify multiple new algorithms. This significantly advances our previous model AlphaTensor, which AlphaEvolve outperforms using its better and

In @NatMachIntell, we show how our AI system AlphaTensor-Quantum can make quantum computing more efficient. 🖥️⚡. By optimizing quantum circuits, it’s helping run calculations faster to save resources and accelerate discoveries. ↓

From its name, I thought AlphaEvolve would be a fancy self-play mcts thing, like AlphaTensor (left). Turns out it's much more like a small #dspy program (right)

So it's strictly better than AlphaTensor on the domain that AlphaTensor was RL'd on, without using RL.

forget alphatensor, how many kernels can be ultra sped up by magic numbers?

Congrats to @GoogleDeepMind for the breakthrough achieved with AlphaEvolve. Earlier AlphaTensor lowered the tensor rank of matrix multiplication of 4x4 matrices over Z2 from 49=7^2 (Strassen) to 47. AlphaEvolve does the job for standard arithmetic: 49 -> 48

AlphaTensor. #動画で学ぶデータサイエンス.

Brillaint 🔥. @GoogleDeepMind launched AlphaEvolve, a Gemini-powered coding agent that discovers and evolves algorithms, outperforming its predecessor AlphaTensor. It achieved 23% matrix kernel speedup, 32.5% GPU kernel boost, and 0.7% compute recovery in data centers, driving

Quantum circuit optimization with AlphaTensor. Fault-tolerant quantum computers aim to run powerful algorithms by using quantum bits (qubits), but they have to deal with hardware errors. Scientists handle these errors using special error-correction techniques, which make some

Discovering novel algorithms with AlphaTensor. @DeepMind. #AI .Cc @DeepLearn007 @ahier @psb_dc @jblefevre60 @pierrepinna @Ym78200

DeepMind AlphaTensor: The delicate balance between human and artificial intelligence. @bendee983 via @bdtechtalks. #AI #MachineLearning . Cc @jblefevre60 @SpirosMargaris @HaroldSinnott @Ym78200 @Nicochan33 @PawlowskiMario

🔥🔥 ¡NUEVO VÍDEO! 🔥🔥. IA + CIENCIA = ¡Explosión de Inteligencia!. La verdadera revolución de la IA se vivirá cuando empecemos a ver cómo se usa para acelerar los descubrimientos científicos. Y DeepMind lo sabe. ¿Conocéis a AlphaTensor? 😏. 👉 LINK:

Alhussein Fawzi, a researcher at DeepMind, helped lead the team that developed AlphaTensor, a neural network that has discovered thousands of speedy algorithms for multiplying matrices.

i used to find it appalling but realized. where else are we gonna get those resources in that quantity without decimating earth. certainly we can mitigate that by figuring how to build faster computers and better algos but ai does that too. alphatensor

If you though AlphaTensor was a big deal (if you heard me in spaces — it’s a massive deal for multiplying the impact of existing dedicated ASICS for models). Check this out:. Deepmind sped up sorting algorithm implementations. has anyone done this since

RL optimizes quantum computing!. Quantum circuits often use a special type of gate, the “T gate" for universal quantum computation. A big goal is to use as few T gates as possible. @GoogleDeepMind's AlphaTensor-Quantum optimizes this by applying deep reinforcement learning (RL).

We are looking for a London-based researcher. During the last couple of years our team did AlphaTensor and FunSearch and our goal hasn't changed: using AI to discover novel algorithms. Join us!

It seems like AlphaTensor started off (2022) this trend of finding efficient matrix multiplication methods and from that point on it's like people are searching for better matrix multiplication algo, it's like you're mining to find the gem that makes your mining tool even sharper.

🧠 AI Outpacing Humans: Proof in the Code.Here’s where recent studies and breakthroughs show AI already eclipsing human capability:1. Mathematical Problem Solving. DeepMind’s AlphaTensor and cap set solution cracked a 132-year-old open math problem — a feat no human had