MetaREVEAL: RL-based Meta-learning from ...
Type de document :
Communication dans un congrès avec actes
URL permanente :
Titre :
MetaREVEAL: RL-based Meta-learning from Learning Curves
Auteur(s) :
Nguyen, Manh Hung [Auteur]
Grinsztajn, Nathan [Auteur]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Scool [Scool]
Guyon, Isabelle [Auteur]
Laboratoire Interdisciplinaire des Sciences du Numérique [LISN]
Sun-Hosoya, Lisheng [Auteur]
TAckling the Underspecified [TAU]
Grinsztajn, Nathan [Auteur]
Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189 [CRIStAL]
Scool [Scool]
Guyon, Isabelle [Auteur]
Laboratoire Interdisciplinaire des Sciences du Numérique [LISN]
Sun-Hosoya, Lisheng [Auteur]
TAckling the Underspecified [TAU]
Titre de la manifestation scientifique :
Workshop on Interactive Adaptive Learning co-located with European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2021)
Ville :
Bilbao/Virtual
Pays :
Espagne
Date de début de la manifestation scientifique :
2021-09-13
Mot(s)-clé(s) en anglais :
Meta-Learning
Learning Curves
Reinforcement Learning
Learning Curves
Reinforcement Learning
Discipline(s) HAL :
Informatique [cs]
Résumé en anglais : [en]
This paper addresses a cornerstone of Automated Machine Learning: the problem of rapidly uncovering which machine learning algorithm performs best on a new dataset. Our approach leverages performances of such algorithms ...
Lire la suite >This paper addresses a cornerstone of Automated Machine Learning: the problem of rapidly uncovering which machine learning algorithm performs best on a new dataset. Our approach leverages performances of such algorithms on datasets to which they have been previously exposed, i.e., implementing a form of meta-learning. More specifically, the problem is cast as a REVEAL Reinforcement Learning (RL) game: the meta-learning problem is wrapped into a RL environment in which an agent can start, pause, or resume training various machine learning algorithms to progressively "reveal" their learning curves. The learned policy is then applied to quickly uncover the best algorithm on a new dataset. While other similar approaches, such as Freeze-Thaw, were proposed in the past, using Bayesian optimization, our methodology is, to the best of our knowledge, the first that trains a RL agent to do this task on previous datasets. Using real and artificial data, we show that our new RL-based meta-learning paradigm outperforms Free-Thaw and other baseline methods, with respect to the Area under the Learning curve metric, a form of evaluation of Anytime learning (i.e., the capability of interrupting the algorithm at any time while obtaining good performance).Lire moins >
Lire la suite >This paper addresses a cornerstone of Automated Machine Learning: the problem of rapidly uncovering which machine learning algorithm performs best on a new dataset. Our approach leverages performances of such algorithms on datasets to which they have been previously exposed, i.e., implementing a form of meta-learning. More specifically, the problem is cast as a REVEAL Reinforcement Learning (RL) game: the meta-learning problem is wrapped into a RL environment in which an agent can start, pause, or resume training various machine learning algorithms to progressively "reveal" their learning curves. The learned policy is then applied to quickly uncover the best algorithm on a new dataset. While other similar approaches, such as Freeze-Thaw, were proposed in the past, using Bayesian optimization, our methodology is, to the best of our knowledge, the first that trains a RL agent to do this task on previous datasets. Using real and artificial data, we show that our new RL-based meta-learning paradigm outperforms Free-Thaw and other baseline methods, with respect to the Area under the Learning curve metric, a form of evaluation of Anytime learning (i.e., the capability of interrupting the algorithm at any time while obtaining good performance).Lire moins >
Langue :
Anglais
Comité de lecture :
Oui
Audience :
Internationale
Vulgarisation :
Non
Collections :
Source :
Date de dépôt :
2021-12-29T02:00:46Z
Fichiers
- https://hal.inria.fr/hal-03502358/document
- Accès libre
- Accéder au document