|
Language technology for everyone| old_uid | 15667 |
|---|
| title | Language technology for everyone |
|---|
| start_date | 2015/05/22 |
|---|
| schedule | 15h |
|---|
| online | no |
|---|
| location_info | salle 118 |
|---|
| summary | High-quality NLP tools - from tokenization to semantic parsers - exist
for 10-15 of the world’s seven thousand languages, of which we have
digital texts for at least a quarter. Even for the major languages,
such as English, our tools only fair reasonably well on standard
language, and not on informal language or dialect. We even see gender
and age biases affect our tools’ performance. In addition our tools
often over-fit arbitrary annotation choices, arguably making them even
less robust to lingustic diversity. This talk surveys recent efforts
in the COASTAL group to bridge these gaps. |
|---|
| responsibles | Candito |
|---|
| |
|