We introduce 🌸✨ AlphaEvolve ✨🌸, an evolutionary coding agent using LLMs coupled with automatic evaluators, to tackle open scientific problems 🧑‍🔬 and optimize critical pieces of compute infra ⚙️.

RT @jeanfrancois287: 📢New Paper on Reward Modelling📢. Ever wondered how to choose the best comparisons when building a preference dataset f….

RT @chrysalis_ai: Take a look at our latest work. One of my proudest works so far :).

RT @stepjamUK: 🚨Important update from our Robot Learning Lab in London. Following recent news, we’re moving on after a wonderful 2 years….T….

Joint work with @hyunjik11, @MatthiasSBauer, @lucastheis, @schwarzjn_.

For technical details, please refer to the paper and code. 📜: 🧑‍💻: ⚙️: We hope this is a step towards making neural codecs a practical reality ✨.

We introduce C3, which significantly improves COOL-CHIC compression performance, approaching the SOTA neural codec (MLIC+) while requiring 200x fewer FLOPs to decode. We also extend C3 to videos 🎥

COOL-CHIC (Ladune et al., 2023), learns a decoder per image, as well as a latent grid and an entropy model *per image*. This dramatically improves compression at a very low decoding cost.

COIN (Dupont et al, 2021) learns a small decoder *per image*, leading to low decoding cost. However, compression performance is weak.

Traditional neural compression models are based on autoencoders trained on datasets of natural images or videos. While these achieve good compression, the decoder is often large as it needs to generalize to arbitrary images, leading to expensive decoding.

We build neural codecs from a *single* image or video, achieving compression performance close to SOTA models trained on large datasets, while requiring ~100x fewer FLOPs for decoding ⚡ #CVPR2024.

Work done with the amazing @ber24, @BarekatainAmin, @SashaVNovikov, @matejbalog, @Mpawankumar123, @franciscuto, @JSEllenberg, @PengmingWang, Omar Fawzi, @pushmeet, @AlhusseinFawzi!.

We present #FunSearch in @Nature today - a system combining LLMs with evolutionary search to generate new discoveries in math and computer science! 👩‍🔬🔬✨.

RT @jinxu06: We construct neural processes by iteratively transforming a simple stochastic process into an expressive one, similar to flow/….

RT @itsbautistam: Introducing Manifold Diffusion Fields (MDF), our new work on learning generative models over fields defined on curved geo….

RT @jinxu06: We introduce a new class of stochastic process models, which are constructed by stacking sequences of neural parameterised Mar….

RT @hyunjik11: Drop by our #ICLR2023 workshop tmrw (Thurs) on "Neural Fields Across Fields: Methods and Applications of INRs"!.Schedule: ht….

RT @schwarzjn_: Very happy to announce that our latest paper on Neural data compression with INRs, Meta Learning & Sparse Subnetwork select….

RT @bobby_he: Can deep transformers be trained without skip connections nor normalisation layers?. Our ICLR 2023 paper shows you how, using….

RT @hyunjik11: Previously we had introduced *functa*, a framework for representing data as neural functions (aka neural fields, INRs) and d….