Image, Word and Thought: A More Challenging Language Task for the Iterated Learning Model

2026-01-06

arXiv.org

AI Summary

This study investigates how linguistic structures can simultaneously exhibit expressivity, compositionality, and stability when transmitted across generations in the context of complex semantics, such as seven-segment digit images. To this end, the authors propose a semi-supervised iterated learning model that integrates supervised and unsupervised learning within an autoencoder architecture to simulate intergenerational language evolution. This work represents the first successful application of iterated learning to high-complexity semantic communication tasks. Experimental results demonstrate that agents develop languages assigning unique codes to all 128 glyphs (strong expressivity), with signals systematically aligned to semantic constituents (robust compositionality), while maintaining structural fidelity over multiple transmission generations. These findings confirm that the three core linguistic properties can co-emerge through cultural transmission under appropriate learning constraints.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Evolutionary Learning
• Search and Optimization: Learning to Search

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Web Mining and Content Analysis: Models for Web evolution

Abstract

The iterated learning model simulates the transmission of language from generation to generation in order to explore how the constraints imposed by language transmission facilitate the emergence of language structure. Despite each modelled language learner starting from a blank slate, the presence of a bottleneck limiting the number of utterances to which the learner is exposed can lead to the emergence of language that lacks ambiguity, is governed by grammatical rules, and is consistent over successive generations, that is, one that is expressive, compositional and stable. The recent introduction of a more computationally tractable and ecologically valid semi supervised iterated learning model, combining supervised and unsupervised learning within an autoencoder architecture, has enabled exploration of language transmission dynamics for much larger meaning-signal spaces. Here, for the first time, the model has been successfully applied to a language learning task involving the communication of much more complex meanings: seven-segment display images. Agents in this model are able to learn and transmit a language that is expressive: distinct codes are employed for all 128 glyphs; compositional: signal components consistently map to meaning components, and stable: the language does not change from generation to generation.

Problem

• iterated learning
• language emergence
• compositional language
• language transmission
• complex meaning

Innovation

• iterated learning model
• semi-supervised learning
• autoencoder
• compositionality
• language emergence