Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study contrasts the vocabulary acquisition processes of children and language models, revealing that children exhibit accelerating returns in word learning, where each new linguistic experience yields greater learning than the last. In contrast, language models demonstrate constant proportional returns, failing to replicate the accelerating learning observed in children despite being trained on child-directed speech. The findings suggest that children's efficient use of linguistic input, requiring significantly less training data, may explain their superior vocabulary growth compared to language models.
Children learn words exponentially faster than language models, highlighting a critical gap in language acquisition efficiency.
Children learn hundreds of words over the first years of their lives, in a process that begins slowly but quickly picks up speed. Prior models describe vocabulary growth as evidence accumulation over time. Here we show that the process is best characterized as accelerating accumulation: children learn more from each additional unit of linguistic experience than they did from the one before. In contrast to children, language models -- even those trained on child-directed speech -- do not accelerate. Instead, they show constant proportional returns on new data, consistent with scaling laws. Children learn using many orders of magnitude less training data than language models; their increasingly efficient use of their learning input is a candidate explanation.