Matthew E. Peters
E800453
Matthew E. Peters is a computer scientist and researcher known for co-developing the ELMo deep contextualized word representation model in natural language processing.
All labels observed (1)
| Label | Occurrences |
|---|---|
| Matthew E. Peters canonical | 1 |
Statements (27)
| Predicate | Object |
|---|---|
| instanceOf |
computer scientist
ⓘ
researcher ⓘ |
| approach | bidirectional language modeling for word representations ⓘ |
| citationsForWork | Deep contextualized word representations is highly cited in NLP research NERFINISHED ⓘ |
| coAuthorOf | Deep contextualized word representations NERFINISHED ⓘ |
| coDeveloperOf | ELMo NERFINISHED ⓘ |
| contributedTo | advances in word representation models ⓘ |
| field |
artificial intelligence
ⓘ
machine learning ⓘ natural language processing ⓘ |
| impact |
ELMo became a standard baseline for contextual word embeddings
NERFINISHED
ⓘ
ELMo improved performance on multiple NLP benchmarks NERFINISHED ⓘ |
| influenced |
contextual embedding methods in NLP
ⓘ
pretrained language model research ⓘ |
| knownFor |
ELMo
NERFINISHED
ⓘ
contextual word embeddings ⓘ deep contextualized word representations ⓘ |
| publicationTitle | Deep contextualized word representations ⓘ |
| publishedIn | NAACL 2018 NERFINISHED ⓘ |
| researchArea |
neural networks for language
ⓘ
sequence modeling ⓘ text representation ⓘ |
| usedMethods |
language model pretraining
ⓘ
recurrent neural networks ⓘ |
| workFocus |
neural language models
ⓘ
representation learning for NLP ⓘ transfer learning in NLP ⓘ |
Referenced by (1)
Full triples — surface form annotated when it differs from this entity's canonical label.