poster - å ¯æcat.phys.s.u-tokyo.ac.jp/~zliu/slides/neurips2019_poster...Title Microsoft...

1
Deep Gamblers: Learning to Abstain with Portfolio Theory Liu Ziyin Liu (UTokyo), Zhikang T. Wang(UTokyo), Paul Liang(CMU), Ruslan salakhutdinov (CMU), Louis-Philippe Morency (CMU), Masahito Ueda (UTokyo), Classification and the Inadequacy of loss Want to find: = arg max Pr(|) In practice, minimize ( loss): min − log (|) The proposed method: the gambler’s loss max log = max ୀଵ log( + ) SOTA Performance… Surprising Benefit: -Training with gambler’s loss reduces overfit -Improved performance when noisy label is present The Learned Representation is Better Separable: Toy Example: Image Rotation.. Toy Example: Identifying Disconfident Images.. Intuition: Prediction as Horse Race Horse Race with Reservation horses Betting strategy: ୀଵ Chance of winning: Payoff if we bet on the winning horse: Return after winning: = + Objective: maximize doubling rate: max = max log = max ୀଵ log( + ) Classification Problem = Betting problem with Reservation with = 1, =0 Classification Problem Betting problem with Reservation

Transcript of poster - å ¯æcat.phys.s.u-tokyo.ac.jp/~zliu/slides/neurips2019_poster...Title Microsoft...

Page 1: poster - å ¯æcat.phys.s.u-tokyo.ac.jp/~zliu/slides/neurips2019_poster...Title Microsoft PowerPoint - poster - å ¯æ ¬.pptx Author a4045 Created Date 10/28/2019 10:57:19 AM

Deep Gamblers: Learning to Abstain with Portfolio TheoryLiu Ziyin Liu (UTokyo), Zhikang T. Wang(UTokyo), Paul Liang(CMU), Ruslan salakhutdinov (CMU), Louis-Philippe Morency (CMU), Masahito Ueda (UTokyo),

Classification and the Inadequacy of 𝑛𝑙𝑙 lossWant to find: 𝜃 = arg max Pr(𝑌|𝜃)

In practice, minimize 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒 𝑙𝑜𝑔 𝑙𝑜𝑠𝑠(𝑛𝑙𝑙 loss): min − log 𝑝(𝑌|𝜃)

The proposed method: the gambler’s loss

max 𝐸 log 𝑆 = max 𝑝 log(𝑜 𝑏 + 𝑏 )

SOTA Performance…

Surprising Benefit:-Training with gambler’s loss reduces overfit-Improved performance when noisy label is present

The Learned Representation is Better Separable:

Toy Example: Image Rotation..

Toy Example: Identifying Disconfident Images..

Intuition: Prediction as Horse RaceHorse Race with Reservation

𝑚 horses Betting strategy: ∑ 𝑏 → ∑ 𝑏 Chance of winning: 𝑝Payoff if we bet on the winning horse: 𝑜Return after winning: 𝑆 = 𝑜 𝑏 → 𝑜 𝑏 + 𝑏

Objective: maximize doubling rate:

max 𝑊 = max 𝐸 log 𝑆 = max 𝑝 log(𝑜 𝑏 + 𝑏 )

Classification Problem = Betting problem with Reservation with 𝑜 = 1, 𝑏 = 0

Classification Problem ≤ Betting problem with Reservation