Cognitive Science in the era of Artificial Intelligence: A roadmap for reverse-engineering the infant language-learner
By Emmanuel Dupoux. 2016
188 citations recorded by the source corpus.
Spectacular progress in the information processing sciences (machine learning, wearable sensors) promises to revolutionize the study of cognitive development. Here, we analyse the conditions under which 'reverse engineering' language development, i.e., building an effective system that mimics infant's achievements, can contribute to our scientific understanding of early language development. We argue that, on the computational side, it is important to move from toy problems to the full complexity of the learning situation, and take as input as faithful reconstructions of the sensory signals available to infants as possible. On the data side, accessible but privacy-preserving repositories of home data have to be setup. On the psycholinguistic side, specific tests have to be constructed to benchmark humans and machines at different linguistic levels. We discuss the feasibility of this approach and present an overview of current results.
Subfields: Artificial Intelligence, Developmental and Educational Psychology, Natural Language Processing, Linguistics
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- James R. Glass
- Wei-Ning Hsu
- David F. Harwath
- Karen Livescu
- Herman Kamper
- Emmanuel Dupoux
- Gabriel Synnaeve
- Michael Auli
In Developmental and Educational Psychology
- James R. Glass
- Herman Kamper
- Tal Linzen
- Laurent Besacier
- William N. Havard
- Jean-Pierre Chevrot
- Aren Jansen
- Sharon Goldwater
In Natural Language Processing
- Wei-Ning Hsu
- Yonatan Belinkov
- James R. Glass
- Awni Y. Hannun
- Tal Linzen
- Herman Kamper
- Gabriel Synnaeve
- Ann Lee
In Linguistics
- Wei-Ning Hsu
- James R. Glass
- David F. Harwath
- Laurent Besacier
- William N. Havard
- Jean-Pierre Chevrot
- Tal Linzen
- Sharon Goldwater
Citations in the corpus
- How Can We Accelerate Progress Towards Human-like Linguistic Generalization? (2020, background): “Systems that generalize like humans would be useful not only for NLP, but also for the scientific study of human language acquisition and processing (Keller, 2010; Dupoux, 2018).”
- Learning Hierarchical Discrete Linguistic Units from Visually-Grounded Speech (2019, background): “By 8 months of age, human infants learn to recognize not only the names of their caregivers and common objects, but also the contrast between the different vowels and consonants which comprise these words (Dupoux, 2018).”
- On Generative Spoken Language Modeling from Raw Audio (2021, neutral): “An open question for AI research is creating systems that learns from natural interactions as infants learn their first language(s): from raw uncu-rated data, and without access to text or expert labels (Dupoux, 2018).”
- Semi-Supervised Speech Recognition via Local Prior Matching (2020, background): “Instead, humans possess an inherent ability to learn from vast quantities of unlabeled speech (Chomsky, 1986; Kuhl, 2004; Glass, 2012; Dupoux, 2018).”
- Jointly Discovering Visual Objects and Spoken Words from Raw Sensory Input (2018, background): “Babies learn to understand speech and recognize objects in an extremely weakly supervised fashion, aided not by ground-truth annotations, but by observation, repetition, multimodal context, and environmental interaction (Dupoux 2018; Spelke 1990).”
- An embedded segmental K-means model for unsupervised segmentation and clustering of speech (2017, background): “In addition, such methods may shed light on how human infants acquire language [2, 3].”
- Models of Visually Grounded Speech Signal Pay Attention to Nouns: A Bilingual Experiment on English and Japanese (2019, methodology): “This work was supported by grants from NeuroCoG IDEX UGA as part of of the ”Investissements d’avenir” program (ANR-15-IDEX-02) cesses at work in humans as suggested by [15].”
- Coincidence, Categorization, and Consolidation: Learning to Recognize Sounds with Minimal Supervision (2019, background): “It is only after this pre-verbal exposure that our abilities of object tracking, color discrimination, object recognition, word and phoneme recognition, and environmental sound recognition emerge [1, 2, 3, 4, 5, 6].”
- Early phonetic learning without phonetic categories: Insights from large-scale simulations on realistic input (2021, background): “The learning algorithms considered have been argued to be particularly relevant for modeling how infants learn in general, and learn language in particular (70).”