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
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
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
8. Define “Realistic” optimistic pessimistic realistic
9. https://snappygoat.com/
10. https://snappygoat.com/
11. https://www.abc15.com
12. A more benign example...
13. Classifier Food / non-food classifier food plant person pet … other
20. Let’s start with what we can easily classify
24. What is “hard” anyway?
25. HP-Net = Hardness Predictor HP-Net hardness
26. Classifier HP-Net Adversarial training with a hardness predictor
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
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
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
30. Can’t we just use confidence scores?
32. Hardness progression during training
33. Do we need two separate models?
34. Classifier + HP-Net +
36. Do we need to fine tune?
37. C - normal classifier F - realistic predictor without fine-tuning (just rejection) F’ - realistic predictor, fine-tuned on samples accepted by HP-Net
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
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)
41. It’s the movie with the mysterious black block, and classical music in space.
42. Spaceships don’t make a “Whoosh!” sound, there’s classical music instead.
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.
44. Open the pod bay doors, HAL!
45. I’m sorry, Dave. I’m afraid I can’t do that. Fra HAL kills Frank.
46. Dave has no choice but to deactivate HAL. Daisy… Daisy…
47. Dedicated to Douglas Rain (March 13, 1928 – November 11, 2018) known as the Voice of HAL