Timezone: »
Open category detection is the problem of detecting "alien" test instances that belong to categories or classes that were not present in the training data. In many applications, reliably detecting such aliens is central to ensuring the safety and accuracy of test set predictions. Unfortunately, there are no algorithms that provide theoretical guarantees on their ability to detect aliens under general assumptions. Further, while there are algorithms for open category detection, there are few empirical results that directly report alien detection rates. Thus, there are significant theoretical and empirical gaps in our understanding of open category detection. In this paper, we take a step toward addressing this gap by studying a simple, but practically-relevant variant of open category detection. In our setting, we are provided with a "clean" training set that contains only the target categories of interest and an unlabeled "contaminated'' training set that contains a fraction alpha of alien examples. Under the assumption that we know an upper bound on alpha we develop an algorithm with PAC-style guarantees on the alien detection rate, while aiming to minimize false alarms. Empirical results on synthetic and standard benchmark datasets demonstrate the regimes in which the algorithm can be effective and provide a baseline for further advancements.
Author Information
Si Liu (Oregon State University)
Risheek Garrepalli (Oregon State University)
Thomas Dietterich ((organization))
Alan Fern (Oregon State University)
Dan Hendrycks (UC Berkeley)
Related Events (a corresponding poster, oral, or spotlight)
-
2018 Oral: Open Category Detection with PAC Guarantees »
Thu. Jul 12th 03:20 -- 03:30 PM Room A6
More from the Same Authors
-
2023 Social: ML Safety Social »
Dan Hendrycks -
2021 : Live Panel Discussion »
Thomas Dietterich · Chelsea Finn · Kamalika Chaudhuri · Yarin Gal · Uri Shalit -
2021 : RL Explainability & Interpretability Panel »
Ofra Amir · Finale Doshi-Velez · Alan Fern · Zachary Lipton · Omer Gottesman · Niranjani Prasad -
2021 : RL Foundation Panel »
Matthew Botvinick · Thomas Dietterich · Leslie Kaelbling · John Langford · Warrren B Powell · Csaba Szepesvari · Lihong Li · Yuxi Li -
2021 Poster: Re-understanding Finite-State Representations of Recurrent Policy Networks »
Mohamad H Danesh · Anurag Koul · Alan Fern · Saeed Khorram -
2021 Spotlight: Re-understanding Finite-State Representations of Recurrent Policy Networks »
Mohamad H Danesh · Anurag Koul · Alan Fern · Saeed Khorram -
2019 : Panel Discussion (moderator: Tom Dietterich) »
Max Welling · Kilian Weinberger · Terrance Boult · Dawn Song · Thomas Dietterich -
2018 Poster: Visualizing and Understanding Atari Agents »
Samuel Greydanus · Anurag Koul · Jonathan Dodge · Alan Fern -
2018 Poster: Discovering and Removing Exogenous State Variables and Rewards for Reinforcement Learning »
Thomas Dietterich · George Trimponias · Zhitang Chen -
2018 Oral: Discovering and Removing Exogenous State Variables and Rewards for Reinforcement Learning »
Thomas Dietterich · George Trimponias · Zhitang Chen -
2018 Oral: Visualizing and Understanding Atari Agents »
Samuel Greydanus · Anurag Koul · Jonathan Dodge · Alan Fern