Mahmoud (Mido) Assran
@mido_assran
Followers
378
Following
442
Media
0
Statuses
46
Research Scientist @AIatMeta. PhD at @mcgillu, @Mila_Quebec. Working on self-supervised representation learning.
Montréal, Québec
Joined June 2020
#ICCV2021.Hello world! Super excited to announce that our recent paper on semi-supervised visual representation learning has been accepted as an oral presentation at ICCV!.
6
37
167
RT @AIatMeta: Our vision is for AI that uses world models to adapt in new and dynamic environments and efficiently learn new skills. We’re….
0
353
0
How much can a model learn about the world just by passively watching video? We’re releasing the results of some of this research today, along with the code and models!.
Today we’re releasing V-JEPA, a method for teaching machines to understand and model the physical world by watching videos. This work is another important step towards @ylecun’s outlined vision of AI models that use a learned understanding of the world to plan, reason and
1
0
14
RT @AIatMeta: Today we’re releasing V-JEPA, a method for teaching machines to understand and model the physical world by watching videos. T….
0
547
0
RT @arimorcos: Excited and honored that our paper on beating power law scaling via data pruning earned an outstanding paper award at #NeurI….
0
9
0
RT @randall_balestr: Self-Supervised Learning methods have strong a priori on the type of data distribution you train on. With @mido_assran….
0
6
0
RT @MetaAI: Tired of manually annotating vast quantities of data to achieve good performance? Our Masked Siamese Networks (MSN) is a self-s….
0
145
0
RT @julienmairal: Congratulations to Dr. Mathilde Caron @mcaron31, who successfully defended her PhD **in person** after a brilliant presen….
0
4
0
Interested in semi-supervised learning of visual representations? Check out our recent work!.
Newest semi/self-supervised learning work from FAIR.
0
0
0
Thank you to all the amazing collaborators involved in this project! @mcaron31, @armandjoulin, @p_bojanowski, @imisra_, Michael Rabbat, and Nicolas Ballas.
0
0
4
The code for our new method, PAWS, is on GitHub, and I'm incredibly excited about all the ways people have already been using PAWS in their own research.
1
0
3
We show how one can leverage a small quantity of labeled data during training in a self-supervised objective to obtain new state-of-the-art semi-supervised learning performance with significantly less computational effort than previous approaches.
1
0
8
RT @MLStreetTalk: We dive deep into self-supervised learning with Dr. Ishan Misra @imisra_ from FAIR @facebookai and cover their recent cl….
0
58
0
RT @an_open_mind: More progress on self-supervision for computer vision - better segmentation and more efficiency h….
0
4
0
RT @facebookai: Working with @Inria researchers, we’ve developed DINO, a method to train Vision Transformers (ViT) with no supervision. Thi….
0
186
0
RT @facebookai: In addition to sharing DINO, a self-supervised model that can discover and segment objects in an image or video with no sup….
0
84
0
RT @alexvoica: Object segmentation is considered one of the hardest challenges in computer vision because it requires that AI truly underst….
0
6
0
RT @onucharlesc: We have internship positions available at @ubenwa_ai for people interested in using machine learning to tackle infant heal….
0
37
0