Antoine Bordes

AI is reshaping the world. Defence is protecting it. Women are underrepresented in both. We intend to change that. Women in AI launches on 23rd April in Paris, a series for practitioners, researchers, and leaders. Then Munich and London. Speakers from @GoogleDeepMind, @AIatMeta, @MistralAI, and Helsing. Register below. eventbrite.com/e/women-in-ai-…

La France, bien positionnée pour devenir un acteur-clé dans l'#IA de défense ? "On a tout ce qu'il faut : les talents et les géants industriels avec qui collaborer", assure @AntoineBordes vice-président de #Helsing dont l'ambition est d'être "l'OpenAI de la défense" @franceinter

"En Ukraine, depuis deux ans, on a déjà vécu plusieurs révolutions technologiques sur la manière de faire la guerre", constate @AntoineBordes le vice-président de #Helsing, société spécialisée dans l'#IntelligenceArtificielle de défense #Narretepasleco @franceinter

This is huge: Llama-v2 is open source, with a license that authorizes commercial use! This is going to change the landscape of the LLM market. Llama-v2 is available on Microsoft Azure and will be available on AWS, Hugging Face and other providers Pretrained and fine-tuned models are available with 7B, 13B and 70B parameters. Llama-2 website: ai.meta.com/llama/ Llama-2 paper: ai.meta.com/research/publi… A number of personalities from industry and academia have endorsed our open source approach: about.fb.com/news/2023/07/l…

@bremen79 @david_picard @gautamcgoel @roydanroy @AdenIshaq @OmarRivasplata @ylecun Brings back memories! And indeed LaSVM was a modified SMO

@david_picard @gautamcgoel @roydanroy @AdenIshaq @OmarRivasplata @ylecun @AntoineBordes Also LaSVM was not SGD but essentially SMO in a sequential way, as they also clearly say in the paper

I recall an old paper (by Leon Bottou?) with a figure showing SGD following the gradient of the population whereas GD chasing the gradient of the sample and using this to argue for SGD.

#VivaTech 🚀 "Can democracy survive AI ?", modéré par @LouettePierre, PDG du Groupe @LesEchos-@le_Parisien, avec @Aurelie_JEAN, Fondatrice et PDG In Silico Veritas et @AntoineBordes, Vice Président AI @HelsingAI depuis la Stage 1 de @VivaTech.

In 2021, we created a research demo that brought amateur drawings to life through animation — today, we're open-sourcing the code + releasing a first-of-its-kind dataset of nearly 180K annotated amateur drawings to help researchers keep innovating in this space. More details ⬇️

LLMs can also learn and predict protein structures! And fast! Great job @alexrives and team.

« Helsing a l'ambition d'être un accélérateur pour l'IA de défense » @AntoineBordes @Armees_Gouv @DGA @thalesgroup @AtosFR @eavidian @Airbus @AJDPresse @AJPAE

Big news - FAIR's @AntoineBordes is joining Helsing as VP Artificial Intelligence. bloomberg.com/news/articles/…

Crazy times for AI! And it is true for defence too. Time for a new chapter for me: bloomberg.com/news/articles/…

Do LLMs work like the human brain? Not clear as shown in this piece by @c_caucheteux @JeanRemiKing & @agramfort . Check it out

This release is great step towards understanding and reproducibility of LLMs!

Today we release LLaMA, 4 foundation models ranging from 7B to 65B parameters. LLaMA-13B outperforms OPT and GPT-3 175B on most benchmarks. LLaMA-65B is competitive with Chinchilla 70B and PaLM 540B. The weights for all models are open and available at research.facebook.com/publications/l… 1/n

🎉 New paper 🎉 Introducing the Toolformer, a language model that teaches itself to use various tools in a self-supervised way. This significantly improves zero-shot performance and enables it to outperform much larger models. 🧰 🔗 Link: arxiv.org/abs/2302.04761

cc @apeyrache

Awesome new work from FAIR at @MetaAI on blind AI agents learning to build internal spatial representations ("mental maps") of their environment while learning to navigate!

The pace of progress in #GenerativeAI is accelerating day after day! Now you can generate 4D assets from text !!

Introducing Make-A-Video3D! Generating 3D dynamic (mini) scenes from input text. That is, text --> 4D! Needs no 4D data (i.e., no dynamic 3D data), no static 3D data, no paired text-video data. Paper: arxiv.org/abs/2301.11280 Website: make-a-video3d.github.io