BabyLM Turns 3: Call for papers for the 2025 BabyLM workshop

2025-02-15

AI Summary

This work addresses critical challenges at the intersection of cognitive modeling and language modeling—namely, data-efficient pretraining, interactive learning, teacher-student co-adaptation, and evaluation of weak models. Methodologically, it introduces a dual-track benchmark framework: a general “Data-Efficient Pretraining” track and a novel “INTERACTION” track, the latter emphasizing model evolution through pedagogical feedback while balancing cognitive plausibility and training efficiency. Technically, the framework integrates few-shot pretraining, cognitively inspired architectures, interactive reinforcement learning, and weakly supervised evaluation. The project has established a three-year longitudinal international benchmark, fostering collaborative research across multiple institutions. It has yielded a series of high-impact publications at top-tier conferences (e.g., ACL, NeurIPS, EMNLP) and advanced substantive progress toward cognitively interpretable language models.

Technology Category

Natural Language Processing: (Large) Language ModelsCognitive Modeling & Cognitive Systems: Social Cognition And InteractionMachine Learning: Large Multimodal Models (LMMs)

Application Category

Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactionsUser Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendationWeb Mining and Content Analysis: Large pretrained models with web data

Abstract

BabyLM aims to dissolve the boundaries between cognitive modeling and language modeling. We call for both workshop papers and for researchers to join the 3rd BabyLM competition. As in previous years, we call for participants in the data-efficient pretraining challenge in the general track. This year, we also offer a new track: INTERACTION. This new track encourages interactive behavior, learning from a teacher, and adapting the teaching material to the student. We also call for papers outside the competition in any relevant areas. These include training efficiency, cognitively plausible research, weak model evaluation, and more.

Problem

• Bridge cognitive and language modeling
• Enhance data-efficient pretraining techniques
• Promote interactive learning and adaptation

Innovation

• Cognitive and language modeling integration
• Data-efficient pretraining challenge
• Interactive learning from teacher