VOS: Learning What You Don't Know By Virtual Outlier Synthesis | Paper Explained
❤️ Become The AI Epiphany Patreon ❤️
/ theaiepiphany
👨👩👧👦 Join our Discord community 👨👩👧👦
/ discord
In this paper I cover the "VOS: Learning What You Don't Know By Virtual Outlier Synthesis" paper - where they introduce a clever way of sampling OOD (out-of-distribution) data in the feature space in order to produce a more robust ID (in-distribution) image classification/object detection that's OOD aware.
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
✅ Paper: https://arxiv.org/abs/2202.01197
✅ Code: https://github.com/deeplearning-wisc/vos
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
⌚️ Timetable:
00:00 Intro to the OOD problem
04:40 High-level VOS explanation
09:38 Alternative synthesis approach (GANs)
11:40 Diving deeper into the method
17:40 Uncertainty loss component
22:45 Inference-time OOD detection
24:05 Method step-by-step overview
26:35 Results
27:45 Computational cost
29:00 Ablations, visualization
30:23 Outro
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💰 BECOME A PATREON OF THE AI EPIPHANY ❤️
If these videos, GitHub projects, and blogs help you,
consider helping me out by supporting me on Patreon!
The AI Epiphany - / theaiepiphany
One-time donation - https://www.paypal.com/paypalme/theai...
Huge thank you to these AI Epiphany patreons:
Eli Mahler
Petar Veličković
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💼 LinkedIn - / aleksagordic
🐦 Twitter - / gordic_aleksa
👨👩👧👦 Discord - / discord
📺 YouTube - / theaiepiphany
📚 Medium - / gordicaleksa
💻 GitHub - https://github.com/gordicaleksa
📢 AI Newsletter - https://aiepiphany.substack.com/
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
#generalization #OOD #virtualoutliersynthesis