Unit 11 - Zero-shot Computer Vision

[mmhamdy]

This is the outline of Unit 11 which will discuss zero-shot learning in computer vision. Each section represents a chapter in this unit (similar to other HF courses).

The issue for the draft outline has some more details, check it out here #43

• Introduction (assigned to @mmhamdy)
  • On Generalization
  • Zero-shot Learning (ZSL), History and Definitions
  • Comparison With Other Techniques
  • Relationship with Transfer Learning
• Zero-shot Learning (assigned to @albertkao227)
  • ZSL vs. Generalized ZSL
  • Semantic Information
  • Embedding Spaces
  • ZSL Algorithms
  • Benchmarks
• Zero-shot Image Recognition (assigned to @mmhamdy)
  • Methods (candidates are: TransZero, ViT-ZSL)
• Zero-shot Object Detection (assigned to @maddy011 )
  • Methods (candidates are: OWL-ViT)
• Zero-shot Image Segmentation
  • Methods (candidates are: CLIPSeg)
• Challenges and Frontiers
  • Advantages of Zero-shot Learning
  • Applications of Zero-shot Learning
  • Challenges and Limitations of Zero-shot Learning
  • Frontiers
• Summary

[maddy011]

Heyy @mmhamdy

Can I work on Zero-shot Object Detection and Segmentation if it's not taken already?
If yes, then let me know how you're planning to structure the Zero-shot Image Recognition part so that all three parts can be uniform.

Thank you :)

[mmhamdy]

Cool! I just assigned you to it.

[seshupavan]

Hello @mmhamdy, the commits for unit-11 have been quite for a while compared to the syllabus. May i know the reason? is it because of a inactive team?

[mmhamdy]

Hi @seshu-pavan. The only active members in this group are me and @albertkao227. The 2nd section is almost done, and the 3rd and last 2 are WIP. other than that, are unassigned which I'm planning to work on once I finish WIP.

Thanks for asking by the way.

[seshupavan]

There is nothing you can do if others are inactive but thanks for holding this far. Can you just copy this response and post there so everything stays align: #185

[mmhamdy]

Sure!