|
What can language models teach us about structure in human language?| title | What can language models teach us about structure in human language? |
|---|
| start_date | 2025/01/15 |
|---|
| schedule | 16h30-17h30 |
|---|
| online | yes |
|---|
| location_info | on zoom |
|---|
| summary | This talk will examine how linguistic theory can benefit from the surprising successes of language models in modeling human language production. Language models provide linguists with an unprecedented empirical tool to expand and test our theoretical hypotheses about language. I will go over two main methodologies for taking advantage of language models as an empirical tool. Firstly, using language model training as an empirical testbed, examining what kinds of environments make statistical language learning possible or hard. Secondly, examining language model internals as functional theories for how language might be represented to lead to capabilities, focusing on how grammatical role is represented and used by language models. Both methodologies showcase the importance of developing empirical paradigms that narrow the gap between computational methods and linguistic concerns in order to make language models into valid tools for theoretical linguistics. |
|---|
| responsibles | Bernard |
|---|
Workflow history| from state (1) | to state | comment | date |
|---|
| submitted | published | | 2025/01/13 13:25 UTC |
| |
|