Natalia Neverova

@NataliaNeverova

Research Scientist (FAIR): computer vision, deep learning, AI

Vrijeme pridruživanja: kolovoz 2010.

Tweetovi

Blokirali ste korisnika/cu @NataliaNeverova

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @NataliaNeverova

  1. Prikvačeni tweet
    2. velj 2018.

    Meet DensePose! We introduce a new computer vision task of dense human pose estimation in the wild, which aims on mapping a RGB image of a person to 3D surface of the human body. First results are in , data and code are coming soon.

    Poništi
  2. 24. sij

    Facebook’s AI Residency program is now accepting applications in Menlo Park, NYC, Seattle and London (!). Apply by January 31st.

    Poništi
  3. 20. pro 2019.

    Today, we are excited to present CoPhy: a new benchmark and framework for Counterfactual Learning of Physics. Paper: Project page: with , J.Mille, ,

    Poništi
  4. 2. stu 2019.
    Prikaži ovu nit
    Poništi
  5. 27. lis 2019.
    Poništi
  6. proslijedio/la je Tweet
    23. lis 2019.
    Poništi
  7. 10. lis 2019.

    Detectron2 in PyTorch is officially released today: including Mask R-CNN, , Panoptic Segmentation, and more. Congratulations to the entire team!

    Poništi
  8. proslijedio/la je Tweet
    23. ruj 2019.

    1/5 3 days to the workshop deadline . We have an amazing lineup of speakers, panellists, moderators and advisors in 3 sessions: Theory, Vision and Further applications, and a closing panel on *The Role of Communication at Large*.

    Science meets Engineering of Deep Learning workshop at NeurIPS
    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    3. ruj 2019.
    Poništi
  10. proslijedio/la je Tweet
    29. kol 2019.

    1st Vision session at kicking off with Manohar Paluri, . 🦁🌍

    Poništi
  11. 6. kol 2019.

    We are happy to announce opening of the track of COCO + Mapillary Joint Recognition Challenge ! Submission deadline October 4, 2019. Note changes in evaluation and rules regarding submissions and awards.

    Poništi
  12. proslijedio/la je Tweet
    2. kol 2019.

    Laura Sevilla and I are coordinating the computer vision parallel track on Thursday at the . Laura (CV in videos; University of Edinburgh), (3D CV in the Wild; FAIR), and Manohar Paluri (CV at Scale; FAIR) are our 3 wonderful speakers 🤓🤩

    Poništi
  13. proslijedio/la je Tweet
    29. srp 2019.

    A recent work on image classification with my fellow colleagues Matthijs and Andrea, and our intern the smart Hugo Touvron. Simple but extremely effective on all the architectures and classification tasks we have tested.

    Poništi
  14. 14. lip 2019.

    A joint work with , Rıza Alp Güler, Iasonas Kokkinos and Andrea Vedaldi, will be presented on Thursday at

    Poništi
  15. proslijedio/la je Tweet
    8. tra 2019.

    The first version of Habitat (v0.1.0) was developed entirely in stealth mode. Moving forward, we are switching from the "Cathedral" model to the "Bazaar" model of development. All issues, WIPs, PRs in the open. Come join us:

    Poništi
  16. proslijedio/la je Tweet

    We're celebrating 5 years of Facebook AI Research and reflecting on the hard lessons learned along the way. Here's a post from our CTO Mike Schroepfer on the challenges we faced in building an lab with an ambitious goal of understanding intelligence:

    Poništi
  17. proslijedio/la je Tweet
    24. lis 2018.

    Today we are releasing Mask R-CNN Benchmark: a fast and modular implementation for Faster R-CNN and Mask R-CNN written entirely in 1.0. It brings up to 30% speedup compared to mmdetection during training. Check out the webcam demo!

    Poništi
  18. proslijedio/la je Tweet
    19. ruj 2018.
    Poništi
  19. 13. ruj 2018.

    Mathilde Caron and Matthijs Douze showing that visual representations on ImageNet can be learned unsipervised using simple clustering, while being only a couple points behind supervised baselines! Joint work with Piotr Bojanowski and Armand Joulin.

    Poništi
  20. proslijedio/la je Tweet
    12. ruj 2018.

    Hervé Jégou , Matthijs Douze and Cordelia Schmid win the ECCV test of time award for "Hamming embedding and weak geometric consistency for large scale image search" - well deserved!

    Poništi
  21. proslijedio/la je Tweet
    11. ruj 2018.

    I'm thrilled to finally release "Visual Rhythm and Beat," our SIGGRAPH 2018 work on manipulating dance in video.

    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·