Interesting new paper for image classification that introduces a new type of convolution (OctaveConv): achieves higher accuracy with fewer FLOPs on ImageNet. But no public code yet: implementations needed! paperswithcode.com/paper/drop-an-…

TextVQA challenge deadline extended to May 27th!

TextVQA Challenge 2020 is now live! Deadline: May 15th More Info: textvqa.org/challenge We have M4C (TextVQA SoTA) as a baseline available in Pythia now. github.com/facebookresear… Stay safe!

We're hosting the ActivityNet-Entities Object Localization (Grounding) Challenge, a part of the annual ActivityNet Challenge at CVPR'20. See details here: lnkd.in/gQkTiWz The hidden test set and eval server will be available in early April. Stay tuned for more details!

The "new" modality of OCR has lot's of interesting modeling opportunities for vision & language research.

Interested in referring expression localization? Video captioning? Grounding? Submit to our ActivityNet-Entities Challenge ahead of May 29! It’s co-hosted at @cvprconf and you can learn more at: github.com/facebookresear… #CVPR2020

Excited to announce our CVPR 2021 paper: KRISP: Integrating Implicit and Symbolic Knowledge for Open-Domain Knowledge-based VQA Collab @facebookai w/ Marcus Rohrbach, Xinlei Chen, Abhinav Gupta and Devi Parikh openaccess.thecvf.com/content/CVPR20… Poster happening now at #CVPR2021!

Starting in 5 minutes! Access via ICCV2021 Or live streaming on YouTube: youtube.com/watch?v=GzIphB…

IEEE Spectrum editors apparently fell for Russian propaganda! Please sign the open letter to IEEE Spectrum: docs.google.com/forms/d/e/1FAI…

Wann versteht auch endlich unser #Bundeskanzler und unsere #Bundesregierung das Putin nur mit Waffen gestoppt werden kann? Beginnt unsere Landesverteidigung nicht im Donbas? Für was brauchen wir Panzer etc. in Deutschland wenn die Front wanders verläuft... n-tv.de/23265258

This work was done during an internship a @MetaAI , in collaboration with S. Whitehead, R. Maheshwary, Rama Vedantam, S. Scherer, Xinlei Chen, Matthieu Cord, Marcus Rohrbach

Excited to start an Alexander von Humboldt Professorship at TU Darmstadt, Germany. Looking for PhDs and PostDocs (English-speaking lab). If you are interested in Reliable Multimodal AI, find me at #CVPR2023 or rohrbach.vision/apply/

Excited to be at the Hessian #AIcon in two weeks! Reach out if you want to meet!

I am excited to open several Ph.D. and PostDoc positions in the broad field of Multimodal Reliable AI, see rohrbach.vision/apply/

Applications to the prestigious Ellis PhD Program are open! If you are interested in the broad space of Multimodal / Reliable / Trustworthy / Modular AI, consider selecting me as a potential PhD advisor (we do have several open positions). Extensive options for collaboration!

🥳We passed the first step in the German excellence competition with our ideas to make AI more reasonable Kristian Kersting Marcus Rohrbach Carsten Binnig @ConstantinRoth1 Jan Peters Georgia Chalvatzaki Martin Mundt and others hessian.AI and Computer Science, TU Darmstadt

📣We have an opening for a professorship in AI programming hessian.AI and Computer Science, TU Darmstadt Assistant TT or Full professor. Looking forward to see outstanding researchers at the crossroads of PL and AI people applying! tu-darmstadt.de/universitaet/k…

We have an opening for a PostDoc on Efficient Multimodal Models. Details see here: linkedin.com/jobs/view/4027… Apply here: informatik.tu-darmstadt.de/mai/multimodal…