Interactive natural language acquisition in a multi-modal recurrent neural architecture
For the complex human brain that enables us to communicate in natural language, we gathered good understandings of principles underlying language acquisition and processing, knowledge about sociocultural conditions, and insights into activity patterns in the brain. However, we were not yet able to u...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
Taylor & Francis Group
2018-01-01
|
Series: | Connection Science |
Subjects: | |
Online Access: | http://dx.doi.org/10.1080/09540091.2017.1318357 |
_version_ | 1797684096329056256 |
---|---|
author | Stefan Heinrich Stefan Wermter |
author_facet | Stefan Heinrich Stefan Wermter |
author_sort | Stefan Heinrich |
collection | DOAJ |
description | For the complex human brain that enables us to communicate in natural language, we gathered good understandings of principles underlying language acquisition and processing, knowledge about sociocultural conditions, and insights into activity patterns in the brain. However, we were not yet able to understand the behavioural and mechanistic characteristics for natural language and how mechanisms in the brain allow to acquire and process language. In bridging the insights from behavioural psychology and neuroscience, the goal of this paper is to contribute a computational understanding of appropriate characteristics that favour language acquisition. Accordingly, we provide concepts and refinements in cognitive modelling regarding principles and mechanisms in the brain and propose a neurocognitively plausible model for embodied language acquisition from real-world interaction of a humanoid robot with its environment. In particular, the architecture consists of a continuous time recurrent neural network, where parts have different leakage characteristics and thus operate on multiple timescales for every modality and the association of the higher level nodes of all modalities into cell assemblies. The model is capable of learning language production grounded in both, temporal dynamic somatosensation and vision, and features hierarchical concept abstraction, concept decomposition, multi-modal integration, and self-organisation of latent representations. |
first_indexed | 2024-03-12T00:24:39Z |
format | Article |
id | doaj.art-ce566375e7884de1add7545cb9b9ac46 |
institution | Directory Open Access Journal |
issn | 0954-0091 1360-0494 |
language | English |
last_indexed | 2024-03-12T00:24:39Z |
publishDate | 2018-01-01 |
publisher | Taylor & Francis Group |
record_format | Article |
series | Connection Science |
spelling | doaj.art-ce566375e7884de1add7545cb9b9ac462023-09-15T10:47:58ZengTaylor & Francis GroupConnection Science0954-00911360-04942018-01-013019913310.1080/09540091.2017.13183571318357Interactive natural language acquisition in a multi-modal recurrent neural architectureStefan Heinrich0Stefan Wermter1Universität HamburgUniversität HamburgFor the complex human brain that enables us to communicate in natural language, we gathered good understandings of principles underlying language acquisition and processing, knowledge about sociocultural conditions, and insights into activity patterns in the brain. However, we were not yet able to understand the behavioural and mechanistic characteristics for natural language and how mechanisms in the brain allow to acquire and process language. In bridging the insights from behavioural psychology and neuroscience, the goal of this paper is to contribute a computational understanding of appropriate characteristics that favour language acquisition. Accordingly, we provide concepts and refinements in cognitive modelling regarding principles and mechanisms in the brain and propose a neurocognitively plausible model for embodied language acquisition from real-world interaction of a humanoid robot with its environment. In particular, the architecture consists of a continuous time recurrent neural network, where parts have different leakage characteristics and thus operate on multiple timescales for every modality and the association of the higher level nodes of all modalities into cell assemblies. The model is capable of learning language production grounded in both, temporal dynamic somatosensation and vision, and features hierarchical concept abstraction, concept decomposition, multi-modal integration, and self-organisation of latent representations.http://dx.doi.org/10.1080/09540091.2017.1318357language acquisitionrecurrent neural networksembodied cognitionmulti-modal integrationdevelopmental robotics |
spellingShingle | Stefan Heinrich Stefan Wermter Interactive natural language acquisition in a multi-modal recurrent neural architecture Connection Science language acquisition recurrent neural networks embodied cognition multi-modal integration developmental robotics |
title | Interactive natural language acquisition in a multi-modal recurrent neural architecture |
title_full | Interactive natural language acquisition in a multi-modal recurrent neural architecture |
title_fullStr | Interactive natural language acquisition in a multi-modal recurrent neural architecture |
title_full_unstemmed | Interactive natural language acquisition in a multi-modal recurrent neural architecture |
title_short | Interactive natural language acquisition in a multi-modal recurrent neural architecture |
title_sort | interactive natural language acquisition in a multi modal recurrent neural architecture |
topic | language acquisition recurrent neural networks embodied cognition multi-modal integration developmental robotics |
url | http://dx.doi.org/10.1080/09540091.2017.1318357 |
work_keys_str_mv | AT stefanheinrich interactivenaturallanguageacquisitioninamultimodalrecurrentneuralarchitecture AT stefanwermter interactivenaturallanguageacquisitioninamultimodalrecurrentneuralarchitecture |