Ilya Sutskever
Ilya Sutskever
collaborator
line weight = collaboration weight
Works in this corpus
| Title | Position | Year | Cited |
|---|---|---|---|
| Dropout: a simple way to prevent neural networks from overfitting | middle | 2014 | 34,236 |
| Distributed Representations of Words and Phrases and their Compositionality | middle | 2013 | 18,054 |
| Mastering the game of Go with deep neural networks and tree search | middle | 2016 | 16,014 |
| Sequence to Sequence Learning with Neural Networks | first | 2014 | 13,351 |
| Improving neural networks by preventing co-adaptation of feature detectors | middle | 2012 | 6,653 |
| Intriguing properties of neural networks | middle | 2013 | 5,739 |
| On the importance of initialization and momentum in deep learning | first | 2013 | 3,523 |
| Sequence to Sequence Learning with Neural Networks | first | 2014 | 3,514 |
| Language Models are Few-Shot Learners | middle | 2020 | 3,020 |
| Recurrent Neural Network Regularization | middle | 2014 | 2,289 |
| Exploiting Similarities among Languages for Machine Translation | last | 2013 | 1,442 |
| An Empirical Exploration of Recurrent Network Architectures | last | 2015 | 1,406 |
Is this one person?
This record is probably one person, but at least one signal is weak. Check the evidence below before relying on the totals.
- neutralNo ORCID on this record, so its identity rests on inference alone.
- supports2 distinct name form(s) across this row's works, counting a spelled-out given name and its initial as one form.
- supportsAt most 2 distinct institution(s) inside any five-year window, which is a normal career.
- supports92% of this row's works sit in its single largest field.
- neutralConfidence is judged on the 12 work(s) this corpus holds, not on the author's whole output. A single-work row carries little evidence either way.