@misc{oord2018representation,
  abstract = {While supervised learning has enabled great progress in many applications,
unsupervised learning has not seen such widespread adoption, and remains an
important and challenging endeavor for artificial intelligence. In this work,
we propose a universal unsupervised learning approach to extract useful
representations from high-dimensional data, which we call Contrastive
Predictive Coding. The key insight of our model is to learn such
representations by predicting the future in latent space by using powerful
autoregressive models. We use a probabilistic contrastive loss which induces
the latent space to capture information that is maximally useful to predict
future samples. It also makes the model tractable by using negative sampling.
While most prior work has focused on evaluating representations for a
particular modality, we demonstrate that our approach is able to learn useful
representations achieving strong performance on four distinct domains: speech,
images, text and reinforcement learning in 3D environments.},
  added-at = {2022-07-11T20:02:57.000+0200},
  author = {Oord, Aaron van den and Li, Yazhe and Vinyals, Oriol},
  biburl = {https://www.bibsonomy.org/bibtex/2dfd444d61e25894a2d2351126f7863f4/simonh},
  description = {[1807.03748] Representation Learning with Contrastive Predictive Coding},
  interhash = {a60649d0354c4bd776a4614048625268},
  intrahash = {dfd444d61e25894a2d2351126f7863f4},
  keywords = {},
  note = {cite arxiv:1807.03748},
  timestamp = {2022-07-12T10:08:51.000+0200},
  title = {Representation Learning with Contrastive Predictive Coding},
  url = {http://arxiv.org/abs/1807.03748},
  year = 2018
}

{"_id":"YutAJRk5r4cE7wCfc","bibbaseid":"oord-li-vinyals-representationlearningwithcontrastivepredictivecoding-2018","authorIDs":[],"author_short":["Oord, A. v. d.","Li, Y.","Vinyals, O."],"bibdata":{"bibtype":"misc","type":"misc","abstract":"While supervised learning has enabled great progress in many applications, unsupervised learning has not seen such widespread adoption, and remains an important and challenging endeavor for artificial intelligence. In this work, we propose a universal unsupervised learning approach to extract useful representations from high-dimensional data, which we call Contrastive Predictive Coding. The key insight of our model is to learn such representations by predicting the future in latent space by using powerful autoregressive models. We use a probabilistic contrastive loss which induces the latent space to capture information that is maximally useful to predict future samples. It also makes the model tractable by using negative sampling. While most prior work has focused on evaluating representations for a particular modality, we demonstrate that our approach is able to learn useful representations achieving strong performance on four distinct domains: speech, images, text and reinforcement learning in 3D environments.","added-at":"2022-07-11T20:02:57.000+0200","author":[{"propositions":[],"lastnames":["Oord"],"firstnames":["Aaron","van","den"],"suffixes":[]},{"propositions":[],"lastnames":["Li"],"firstnames":["Yazhe"],"suffixes":[]},{"propositions":[],"lastnames":["Vinyals"],"firstnames":["Oriol"],"suffixes":[]}],"biburl":"https://www.bibsonomy.org/bibtex/2dfd444d61e25894a2d2351126f7863f4/simonh","description":"[1807.03748] Representation Learning with Contrastive Predictive Coding","interhash":"a60649d0354c4bd776a4614048625268","intrahash":"dfd444d61e25894a2d2351126f7863f4","keywords":"","note":"cite arxiv:1807.03748","timestamp":"2022-07-12T10:08:51.000+0200","title":"Representation Learning with Contrastive Predictive Coding","url":"http://arxiv.org/abs/1807.03748","year":"2018","bibtex":"@misc{oord2018representation,\n abstract = {While supervised learning has enabled great progress in many applications,\r\nunsupervised learning has not seen such widespread adoption, and remains an\r\nimportant and challenging endeavor for artificial intelligence. In this work,\r\nwe propose a universal unsupervised learning approach to extract useful\r\nrepresentations from high-dimensional data, which we call Contrastive\r\nPredictive Coding. The key insight of our model is to learn such\r\nrepresentations by predicting the future in latent space by using powerful\r\nautoregressive models. We use a probabilistic contrastive loss which induces\r\nthe latent space to capture information that is maximally useful to predict\r\nfuture samples. It also makes the model tractable by using negative sampling.\r\nWhile most prior work has focused on evaluating representations for a\r\nparticular modality, we demonstrate that our approach is able to learn useful\r\nrepresentations achieving strong performance on four distinct domains: speech,\r\nimages, text and reinforcement learning in 3D environments.},\n added-at = {2022-07-11T20:02:57.000+0200},\n author = {Oord, Aaron van den and Li, Yazhe and Vinyals, Oriol},\n biburl = {https://www.bibsonomy.org/bibtex/2dfd444d61e25894a2d2351126f7863f4/simonh},\n description = {[1807.03748] Representation Learning with Contrastive Predictive Coding},\n interhash = {a60649d0354c4bd776a4614048625268},\n intrahash = {dfd444d61e25894a2d2351126f7863f4},\n keywords = {},\n note = {cite arxiv:1807.03748},\n timestamp = {2022-07-12T10:08:51.000+0200},\n title = {Representation Learning with Contrastive Predictive Coding},\n url = {http://arxiv.org/abs/1807.03748},\n year = 2018\n}\n\n","author_short":["Oord, A. v. d.","Li, Y.","Vinyals, O."],"key":"oord2018representation","id":"oord2018representation","bibbaseid":"oord-li-vinyals-representationlearningwithcontrastivepredictivecoding-2018","role":"author","urls":{"Paper":"http://arxiv.org/abs/1807.03748"},"metadata":{"authorlinks":{}},"downloads":0,"html":""},"bibtype":"misc","biburl":"http://www.bibsonomy.org/bib/author/van den oord?items=1000","creationDate":"2019-10-25T03:26:15.067Z","downloads":0,"keywords":[],"search_terms":["representation","learning","contrastive","predictive","coding","oord","li","vinyals"],"title":"Representation Learning with Contrastive Predictive Coding","year":2018,"dataSources":["fjacg9txEnNSDwee6","k67YLoDHfisS7tu6Q"]}