From Child Language to AI: Large-Scale Multimodal Data for Cognitive Research and Application

Description

In an era of rapid developments in generative AI and digital technology, many fields are facing significant challenges. Despite the tremendous power of AI models, it has become increasingly clear that high-quality linguistic and non-linguistic (multimodal) data are crucial, not only for human learning but also machine learning. Many current AI models suffer from biases, imprecision, and hallucination because they are trained on random or non-embodied text data, and the models’ success also rests on the amount of multimodal large-scale data for training or pre-training. In this talk, I provide examples to illustrate how cognitive scientists should leverage high-quality multimodal data in the study of language acquisition, language representation, text comprehension, and the neurocognition of language.

Period	1 Jun 2025
Event title	International Workshop on Cross-linguistic Databases and Norms: IWCDN Workshop
Event type	Workshop
Location	ChinaShow on map
Degree of Recognition	International