Happy to share that our work has been accepted to ECCV 2022! "STEEX: Steering Counterfactual Explanations with Semantics"🚙🔎 w/ P. Jacob, H. Ben-Younes (Hedi Ben younes), Mickael Chen, P. Pérez, and M. Cord (Matthieu Cord). paper: arxiv.org/abs/2111.09094 code: github.com/valeoai/steex

One of my all-time favorite illusions: The spinning dancer If you look at the dancer on the left and the one in the middle, the one in the middle spins clockwise. If you look at the dancer on the *right* and the one in the middle, the one in the middle spins counterclockwise.

Our paper "LaRa: Latents and Rays for Multi-Camera Bird's-Eye-View Semantic Segmentation" made it to #CoRL2022 🥳 w/ F. Bartoccioni, A. Bursuc (Andrei Bursuc), P. Pérez, Matthieu Cord (Matthieu Cord) & K. Alahari (Karteek Alahari) arxiv: arxiv.org/abs/2206.13294 code: github.com/valeoai/LaRa

Congratulations to Hugo Touvron who brightly defended his PhD thesis today! 👏👏👏 Thank you for the very interesting presentation of your work! Good luck for the future!

I’d like to address the serious matter of some newcomers to AI experiencing imposter syndrome, where someone wonders if they’re a fraud or really belong in the AI community. Lets build a community that encourages and welcomes everyone. deeplearning.ai/the-batch/issu…

Cookbook for Vision Transformers deeplearning.ai/the-batch/issu… Good insight about our DeiT III in the Andrew Ng’s news letter I will present it with Hugo at ECCV 2022!

DiffEdit: Diffusion-based semantic image editing with mask guidance abs: arxiv.org/abs/2210.11427

First #ECCV2022 evening with Matthieu Cord in a nice restaurant in town. It's been a long time since the last time and we had a lot of things to talk about. There are few things as enjoyable as meeting your PhD supervisor at a conference long after you defended. 10/10, would recommend.

Our researchers are attending NeurIPS Conference #NeurIPS2023 Do not hesitate to reach them and discuss with them. Review once again our papers in the following thread ⬇️⬇️⬇️

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

Christian Wolf (🦋🦋🦋) Kosta Derpanis Same. My biggest fear is not the Terminator apocalypse, but more what could happen if a large number of people ask an LLM "Sum up the candidates ideas and tell me which one is the most aligned with me" during the next elections. Not voting, because yes/no is too restrictive.

PhD defense of Corentin Dancette Corentin Dancette at Sorbonne University, Paris, on Shortcut learning in VQA. Committee: Matthieu Cord nicolasthome Damien Teney Marcus Rohrbach Elisa Fromont and yours, truly.

Congratulations to Corentin! Bravo!

We're happy to announce the #ICCV2023 BRAVO Workshop: roBust and Reliability of Autonomous Vehicles in the Open-world. We welcome full papers and extended abstracts on  robustness, generalization, transparency, and verification of computer vision models valeoai.github.io/bravo/

Félicitations à Hugo Touvron qui a obtenu le prix de thèse Signal, Image et Vision 2023 décerné par le GdR ISIS ! "Architectures and Training for visual understanding" et supervisée par Matthieu Cord et Hervé Jegou from @meta tel.archives-ouvertes.fr/tel-03945782v1 Sorbonne Université ISIR - Robotique

The team is active at #ICCV2023, come and say Hi to the Chordettes! They are talking about LLMs, multimodality, foundation models, diffusion models and more! Check their papers: 👇

UnIVAL: Unified Model for Image, Video, Audio and Language Tasks. 👉A unified foundation model that can process images, videos and audios-language tasks, trained with academic compute budget, on public datasets. Project: unival-model.github.io Workshop: Monday in MMFM

Thanks all for coming to Paris #ICCV2025 and making it such an amazing event. As warned, we brought the whole valeo.ai family, interns and seniors alike, and it was a blast! We are so happy to share good papers, discussions and laughs. What an amazing community #ICCV2023

At #ICCV2023, presenting our work ep-ALM with Mustafa Shukor and Matthieu Cord. Check out the website at mshukor.github.io/eP-ALM.github.…