Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hongler, Clément, Gabriel, Franck, Hartmann, Valentin, Renard, Arthur, Emil, Andrew
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2603.22479
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908956633333760
author Hongler, Clément
Gabriel, Franck
Hartmann, Valentin
Renard, Arthur
Emil, Andrew
author_facet Hongler, Clément
Gabriel, Franck
Hartmann, Valentin
Renard, Arthur
Emil, Andrew
contents Defining a constructive process to build general capabilities for language models in an automatic manner is considered an open problem in artificial intelligence. Towards this, we consider the problem of building a curriculum of tasks that grows a model via relevant skill discovery. We provide a concrete framework for this task, using a family of tasks called Cross-Entropy Games, which we postulate is universal in a suitable sense. We show that if it is possible to grow the curriculum for relevant skill discovery by iterating a greedy optimization algorithm, then, under natural assumptions, there is essentially only one meta-objective possible (up to a few hyper-parameters). We call the resulting process cognitive training. We postulate that, given sufficiently capable language models as players and meta-samplers, cognitive training provides a principled way to relevant skill discovery; and hence to the extent general capabilities are achievable via greedy curriculum learning, cognitive training would be a solution.
format Preprint
id arxiv_https___arxiv_org_abs_2603_22479
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Cognitive Training for Language Models: Towards General Capabilities via Cross-Entropy Games
Hongler, Clément
Gabriel, Franck
Hartmann, Valentin
Renard, Arthur
Emil, Andrew
Optimization and Control
Artificial Intelligence
Defining a constructive process to build general capabilities for language models in an automatic manner is considered an open problem in artificial intelligence. Towards this, we consider the problem of building a curriculum of tasks that grows a model via relevant skill discovery. We provide a concrete framework for this task, using a family of tasks called Cross-Entropy Games, which we postulate is universal in a suitable sense. We show that if it is possible to grow the curriculum for relevant skill discovery by iterating a greedy optimization algorithm, then, under natural assumptions, there is essentially only one meta-objective possible (up to a few hyper-parameters). We call the resulting process cognitive training. We postulate that, given sufficiently capable language models as players and meta-samplers, cognitive training provides a principled way to relevant skill discovery; and hence to the extent general capabilities are achievable via greedy curriculum learning, cognitive training would be a solution.
title Cognitive Training for Language Models: Towards General Capabilities via Cross-Entropy Games
topic Optimization and Control
Artificial Intelligence
url https://arxiv.org/abs/2603.22479