Towards Realistic Predictors - EN

Later version of the "Towards Realistic Predictors" paper review, in English.

Transcript

1. Towards Realistic Predictors
   Pei Wang and Nuno Vasconcelos
   Statistical and Visual Computing Lab, UC San Diego
   mlKitchen X
   2018.11.29
   @_lunardog_
   http://openaccess.thecvf.com/content_ECCV_2018/papers/Pei_Wang_Towards_Realistic_Predictors_ECCV_2018_paper.pdf
   

   View Slide

2. Self-Introduction
   ● Call me Leszek
   ● Born in Poland
   ● Living in Japan since 2010
   ● Cookpad R&D since 2016
   ● I consume too much science fiction
   ● I’m bad at selfies
   

   View Slide

3. View Slide

4. View Slide

5. View Slide

6. View Slide

7. Towards Realistic Predictors
   Pei Wang and Nuno Vasconcelos
   Statistical and Visual Computing Lab, UC San Diego
   http://openaccess.thecvf.com/content_ECCV_2018/papers/Pei_Wang_Towards_Realistic_Predictors_ECCV_2018_paper.pdf
   

   View Slide

8. Define “Realistic”
   optimistic
   pessimistic realistic
   

   View Slide

9. https://snappygoat.com/
   

   View Slide

10. https://snappygoat.com/
    

    View Slide

11. https://www.abc15.com
    

    View Slide

12. A more benign example...
    

    View Slide

13. Classifier
    Food / non-food classifier
    food
    plant
    person
    pet
    …
    other
    

    View Slide

14. View Slide

15. View Slide

16. View Slide

17. View Slide

18. View Slide

19. View Slide

20. Let’s start with what we can easily classify
    

    View Slide

21. View Slide

22. View Slide

23. View Slide

24. What is “hard” anyway?
    

    View Slide

25. HP-Net = Hardness Predictor
    HP-Net hardness
    

    View Slide

26. Classifier
    HP-Net
    Adversarial training with a hardness predictor
    

    View Slide

27. Hardness Predictor Loss
    bi y s e t y s u w r e s ma c
    mi zi t K l a k-Le b di g e b en t
    di r i n d a m i m n = 1 − p
    c
    

    View Slide

28. Classifier Loss weighted by hardness
    we t ro -en p
    ma h er p e (la r ) mo po n , w i as e m s
    (lo s) ar en s or c
    

    View Slide

29. Classifier
    HP-Net
    Training
    1. train classifier F and
    HP-Net S jointly on training
    set D
    2. run S on D and eliminate
    hard examples, to create
    realistic training set D′
    3. learn realistic classifier F′
    on D′, with S fixed
    4. output pair S, F′
    5. GOTO 1
    D
    F
    S
    

    View Slide

30. Can’t we just use confidence scores?
    

    View Slide

31. View Slide

32. Hardness
    progression
    during training
    

    View Slide

33. Do we need two separate models?
    

    View Slide

34. Classifier
    +
    HP-Net
    +
    

    View Slide

35. View Slide

36. Do we need to fine tune?
    

    View Slide

37. C - normal classifier
    F - realistic predictor without fine-tuning (just rejection)
    F’ - realistic predictor, fine-tuned on samples accepted by HP-Net
    

    View Slide

38. View Slide

39. Conclusions
    ● There are times when it’s OK to skip hard samples
    ● ...and times when it’s BEST to reject hard samples
    ● The paper introduces a GAN-like architecture to train
    any classifier with its own hardness predictor
    ● Training with a hardness predictor improves accuracy
    ● HP-Net should be trained jointly with the classifier, but in an alternating order
    ● HP-Net solves a different problem from the Classifier, should be a separate model
    ● ….but best results are when the architectures are the same
    

    View Slide

40. https://ja.wikipedia.org/wiki/2001年宇宙の旅
    　　　うちゅうのたび
    2001年宇宙の旅
    『2001年宇宙の旅』（にせんいちねんう
    ちゅうのたび、原題：2001: A Space
    Odyssey）は、アーサー・C・クラークとスタン
    リー・キューブリックのアイデアをまとめた
    ストーリーに基いて製作された、SF映画お
    よびSF小説である。
    2001: A Space Odyssey is a 1968 epic science
    fiction film produced and directed by Stanley
    Kubrick. The screenplay was written by Kubrick
    and Arthur C. Clarke, and was inspired by
    Clarke's short story "The Sentinel". A novel also
    called 2001: A Space Odyssey, written
    concurrently with the screenplay, was
    published soon after the film was released.
    2001: A Space Odyssey
    https://en.wikipedia.org/wiki/2001:_A_Space_Odyssey_(film)
    

    View Slide

41. It’s the movie with the mysterious black block, and classical music in space.
    

    View Slide

42. Spaceships don’t make a “Whoosh!” sound, there’s classical music instead.
    

    View Slide

43. Fra Dav
    HA 9000 is r pe
    in f e n a om
    d i n p e h
    Frank and Dave don’t trust HAL.
    HAL is a realistic predictor and doesn’t always follow orders.
    

    View Slide

44. Open the pod bay doors, HAL!
    

    View Slide

45. I’m sorry, Dave.
    I’m afraid I can’t do that.
    Fra
    HAL kills Frank.
    

    View Slide

46. Dave has no choice but to deactivate HAL.
    Daisy… Daisy…
    

    View Slide

47. Dedicated to Douglas Rain
    (March 13, 1928 – November 11, 2018)
    known as the Voice of HAL
    

    View Slide