Date / Heure
Date(s) - 14/06/2019
12h30 - 14h30
Algolia 55 Rue d'Amsterdam · Paris
Zero, one or few shot learning is the new thing. The marriage between ML and big data is considering a divorce. Have we reached peak big data? We’ll discuss the ever growing list of domains where models use a few shots or less.
We meet every Friday for a peer-to-peer discussion of a pre-selected machine learning research paper. The paper and discussion is in English. Tuesdays we meet to discuss applied topics.
Following will be frequently updated.
1. Can AI approach human ability to generalize from few examples?
2. Can AI models surpass human ability to learn from a few shot learning?
3. How important will transfer learning become to few shot modeling?
1. Few-Shot Adversarial Learning of Realistic Neural Talking Head Models: https://arxiv.org/abs/1905.08233
2. Fun with Snapchat’s Gender Swapping Filter: https://blog.evjang.com/2019/05/fun-with-snapchats-gender-swapping.html
3. NVIDIA AI Enables Low-Data Face Swap for Pets: https://medium.com/syncedreview/nvidia-ai-enables-low-data-face-swap-for-pets-56ba3be78ec0
1. Few-Shot Adversarial Learning of Realistic Neural Talking Head Models: https://www.youtube.com/watch?v=p1b5aiTrGzY&feature=youtu.be