Sprache verbessert das Lernen in künstlichen Netzwerken: Wenn ein künstliches Netzwerk mit einem zweiten durch Nachrichten kommuniziert kann das zweite „Schüler-Netzwerk“ schneller lernen.

Language helps artificial networks to learn


Bonn researchers get to the bottom of the social aspect of communication for mental activity – Across all species, important skills for survival, such as hunting prey, are passed on from parents to offspring through communication. Researchers at the University Hospital Bonn (UKB) and the University of Bonn were able to show that effective language-like communication is a two-way process between sender and receiver by using artificial networks as agents that took on the role of teachers and trainees. The results of their study have been published in the scientific journal „Nature Communications.“

Bonn, September, 05th, 2024 Communication, in the form of sounds, smells or movements, is important for the survival of humans in several respects. The social aspect is particularly important for cognition. Our task descriptions in the brain are not only shaped by sensory experiences, but also by the information communicated to us.

‘We know from our everyday lives that social communication improves our learning abilities in the real world, which is summarised by the saying ‘teaching is learning for the second time’,’ says Prof. Tatjana Tchumatchenko, from the Institute of Experimental Epileptology and Cognition Research at the UKB.


Brain creates abstractions for our real world

The results showed that both roles can develop a language that enables the learner to learn from the teacher. Interestingly, this language was influenced by both the task to be solved and the performance of the learner. ‘What we found is consistent with what is known about language formation in animals,’ says Carlos Wert-Carvajal, co-corresponding author and PhD student at the University of Bonn in Prof Tchumatchenko’s research group at the UKB. He emphasises that the way our brain encodes our world is not only determined by our own experiences, but also creates abstractions that are understandable to others: ‘For example, we don’t say ‘sweet, crunchy, round red or green fruit’, but use the single word ‘apple’. Such a word exists because our language has evolved to represent a shared experience that provides a pleasant reward.’ In other words, every language must describe the world as efficiently as possible.

This efficiency meant not only a concise message, but also one that contained as much information as possible. Good language had to combine both the teacher’s and trainee’s internal descriptions of the task and the actual characteristics of the real world. ‘When we gave feedback on how well the trainee performed the task, the teacher changed their language to convey more useful information,’ explains first author Tobias Wieczorek. This process shows that effective communication is a two-way process. ‘Both the sender and the receiver need to work together to ensure that the information exchanged is clear, precise and truly useful,’ says Prof Tchumatchenko, who led the study.

Language closes the loop in communication as a shared experience

Remarkably, by ‘closing the loop’ – that is, by returning the language of the learner to itself – the Bonn researchers were able to enable learners to teach each other. Despite lacking explicit teaching skills, the agents effectively communicated essential information and demonstrated the robustness of the language they had developed. ‘Although they did not know how to ‘teach’, they were still able to use their language to convey important information,’ says co-corresponding author Dr Maximilian Eggl, who until recently was a postdoc at the University of Bonn in Prof Tchumatchenko’s research group at UKB.

This research highlights the fundamental role of language-like communication as a shared cognitive experience and demonstrates its critical importance for learning and generalisation. The results provide valuable insights into the design of biological and artificial communication systems that optimise learning and task performance in different environments.

Original publication:

Tobias J. Wieczorek, Tatjana Tchumatchenko, Carlos Wert-Carvajal and Maximilian F. Eggl: A framework for the emergence and analysis of language in social learning agents; Nature Communications15, 7590 (2024); DOI: 10.1038/s41467-024-51887-5

Further information:

https://doi.org/10.1038/s41467-024-51887-5 Publikation


Beitrag veröffentlicht

in

von

Kommentare

Schreibe einen Kommentar

Deine E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert