More Annotations

Favourite Annotations

Text

2021 CONFERENCE

navigation.

NEURIPS 2020 : PAPERS This is the public, feature-limited version of the conference webpage. After Registration and login please visit the full version. CALL FOR DATASETS BENCHMARKS NeurIPS 2021 Datasets and Benchmarks Track. The Datasets and Benchmarks track serves as a novel venue for high-quality publications, talks, and posters on highly valuable machine learning datasets and benchmarks, as well as a forum for discussions on how to improve dataset development. Datasets and benchmarks are crucial for the development of machine learning methods, but also require their PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

PAPERINFORMATION / STYLEFILES You must use the NeurIPS 2021 LaTeX style file. The maximum file size for submissions is 50MB. Submissions that violate the NeurIPS style (e.g., by decreasing margins or font sizes) or page limits may be rejected without further review. Please make sure that your paper

prints well.

SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAINING Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain CONSTRAINED EPISODIC REINFORCEMENT LEARNING IN CONCAVE Constrained Markov decision process. We work with MDPs that have resource consumption in addition to rewards. Formally, a constrained MDP (CMDP) is a triple M=

2021 CONFERENCE

navigation.

for

prints well.

2019 CONFERENCE

The Expo is a one-day industry day with talks, panels, demos and workshops from our sponsors. It takes place on Sunday Dec 8th. If you plan to ONLY attend the expo, then please visit Expo Only Registration � otherwise you may register for the expo as you register for the

main conference.

PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

CALL FOR DATASETS BENCHMARKS NeurIPS 2021 Datasets and Benchmarks Track. The Datasets and Benchmarks track serves as a novel venue for high-quality publications, talks, and posters on highly valuable machine learning datasets and benchmarks, as well as a forum for discussions on how to improve dataset development. Datasets and benchmarks are crucial for the development of machine learning methods, but also require their PAPERINFORMATION / STYLEFILES You must use the NeurIPS 2021 LaTeX style file. The maximum file size for submissions is 50MB. Submissions that violate the NeurIPS style (e.g., by decreasing margins or font sizes) or page limits may be rejected without further review. Please make sure that your paper

prints well.

REVIEW FOR NEURIPS PAPER: RECONSTRUCTING PERCEPTIVE IMAGES Review 2. Summary and Contributions: This paper used a new method for separately decoding shape and semantic information about images from the corresponding fMRI responses, and then combining the shape and semantic information together in order to reconstruct the image.This method seems highly effective – much more so than previous “all-in-one” methods. REVIEW FOR NEURIPS PAPER: TRAINING GENERATIVE ADVERSARIAL Review 1. Summary and Contributions: This paper proposes a new effective method for training GANs from small size of data by incorporating non-leaking data augmentation.For achieving this, the authors design an adaptive discriminator augmentation (ADA). They extensively analyze why data augmentation can harm the GAN performance with diverse experiments. REVIEW FOR NEURIPS PAPER: UNFOLDING THE ALTERNATING Review 1. Summary and Contributions: This paper tackles Blind-SR by unfolding the two steps approach of IKC (Gu'19) into one trainable network that requires no test-time optimization.. Strengths: 1.Simple and elegant way to idealize the two steps. 2. Impressive results, large meaningful margin. Weaknesses: All weaknesses are related to experiments, analysis and understanding. GPIPE: EFFICIENT TRAINING OF GIANT NEURAL NETWORKS USING Figure 2: (a) An example neural network with sequential layers is partitioned across four accelerators. F k is the composite forward computation function of the k-th cell. B k is the back-propagation function, which depends on both B k+1 from the upper layer and F k. SEE, HEAR, EXPLORE: CURIOSITY VIA AUDIO-VISUAL ASSOCIATION Audio and visual information are closely linked, and since we commonly have access to both in the form of video, this is a rich area for

self-supervision.

PAPERS.NEURIPS.CC

papers.neurips.cc

2021 CONFERENCE

Conference Sessions, Tutorials, Workshops and Expo. Mon Dec 6th through Tue the 14th. Abstract Submission Deadline. May 21 '21 08:00 PM UTC *. 00 weeks 00 days 00:00:00. Applications for Workshops Open. May 28 '21 04:00 PM UTC *. 00 weeks 04 days 12:07:12. Paper submission and co-author registration deadline. 2021 DATES AND DEADLINES Jul 04 '21 (Anywhere on Earth) Datasets and Benchmarks Submission deadline (2nd round) Aug 27 '21 (Anywhere on Earth) Datasets Reviews released and start of discussions. Sep 24 '21 (Anywhere on Earth) Datasets Author notification for the second round. NEURIPS 2021 COMPETITION TRACK NeurIPS 2021 Competition Track. Below you will find a brief summary of accepted competitions NeurIPS 2021. Regular competitions take place before the NeurIPS, whereas live competitions will have their final phase during the competition session @NeurIPS2021. Competitions are listed in alphabetical order, all prizes are tentative and depend PAPERINFORMATION / ETHICSREVIEW PaperInformation / EthicsReview. NeurIPS | 2021. Thirty-fifth Conference on Neural Information Processing Systems. Toggle

navigation.

CALL FOR DATASETS BENCHMARKS NeurIPS 2021 Datasets and Benchmarks Track. The Datasets and Benchmarks track serves as a novel venue for high-quality publications, talks, and posters on highly valuable machine learning datasets and benchmarks, as well as a forum for discussions on how to improve dataset development. Datasets and benchmarks are crucial for the development of machine learning methods, but also require their PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

A CAUSAL VIEW ON ROBUSTNESS OF NEURAL NETWORKS2 A Causal View on Robustness of Neural Networks Cheng Zhang Microsoft Research Cheng.Zhang@microsoft.com Kun Zhang Carnegie Mellon University kunz1@cmu.edu SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR8 Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAINING12 Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain GRAPH TRANSFORMER NETWORKS signiﬁcantly affected by the choice of meta-paths. Unlike these approaches, our Graph Transformer Networks can operate on a heterogeneous graph and transform the graph for tasks while learning

2021 CONFERENCE

navigation.

CALL FOR DATASETS BENCHMARKS NeurIPS 2021 Datasets and Benchmarks Track. The Datasets and Benchmarks track serves as a novel venue for high-quality publications, talks, and posters on highly valuable machine learning datasets and benchmarks, as well as a forum for discussions on how to improve dataset development. Datasets and benchmarks are crucial for the development of machine learning methods, but also require their PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

2019 CONFERENCE

main conference.

LOGIN - NEURIPS.CC

Do not remove: This comment is monitored to verify that the site is

working properly

prints well.

REVIEW FOR NEURIPS PAPER: TRAINING GENERATIVE ADVERSARIAL Review 1. Summary and Contributions: This paper proposes a new effective method for training GANs from small size of data by incorporating non-leaking data augmentation.For achieving this, the authors design an adaptive discriminator augmentation (ADA). They extensively analyze why data augmentation can harm the GAN performance with diverse experiments. REVIEW FOR NEURIPS PAPER: UNFOLDING THE ALTERNATING Review 1. Summary and Contributions: This paper tackles Blind-SR by unfolding the two steps approach of IKC (Gu'19) into one trainable network that requires no test-time optimization.. Strengths: 1.Simple and elegant way to idealize the two steps. 2. Impressive results, large meaningful margin. Weaknesses: All weaknesses are related to experiments, analysis and understanding. LEARNING WITH OPTIMIZED RANDOM FEATURES: EXPONENTIAL Authors. Hayata Yamasaki, Sathyawageeswar Subramanian, Sho Sonoda, Masato Koashi. Abstract. Kernel methods augmented with random features give scalable algorithms for learning from big data. GPIPE: EFFICIENT TRAINING OF GIANT NEURAL NETWORKS USING Figure 2: (a) An example neural network with sequential layers is partitioned across four accelerators. F k is the composite forward computation function of the k-th cell. B k is the back-propagation function, which depends on both B k+1 from the upper layer and F k.

PAPERS.NEURIPS.CC

papers.neurips.cc

2021 CONFERENCE

navigation.

CALL FOR DATASETS BENCHMARKS NeurIPS 2021 Datasets and Benchmarks Track. The Datasets and Benchmarks track serves as a novel venue for high-quality publications, talks, and posters on highly valuable machine learning datasets and benchmarks, as well as a forum for discussions on how to improve dataset development. Datasets and benchmarks are crucial for the development of machine learning methods, but also require their PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

A CAUSAL VIEW ON ROBUSTNESS OF NEURAL NETWORKS2 A Causal View on Robustness of Neural Networks Cheng Zhang Microsoft Research Cheng.Zhang@microsoft.com Kun Zhang Carnegie Mellon University kunz1@cmu.edu SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR6 Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAINING11 Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain GRAPH TRANSFORMER NETWORKS signiﬁcantly affected by the choice of meta-paths. Unlike these approaches, our Graph Transformer Networks can operate on a heterogeneous graph and transform the graph for tasks while learning

2021 CONFERENCE

navigation.

CALL FOR DATASETS BENCHMARKS NeurIPS 2021 Datasets and Benchmarks Track. The Datasets and Benchmarks track serves as a novel venue for high-quality publications, talks, and posters on highly valuable machine learning datasets and benchmarks, as well as a forum for discussions on how to improve dataset development. Datasets and benchmarks are crucial for the development of machine learning methods, but also require their PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

A CAUSAL VIEW ON ROBUSTNESS OF NEURAL NETWORKS2 A Causal View on Robustness of Neural Networks Cheng Zhang Microsoft Research Cheng.Zhang@microsoft.com Kun Zhang Carnegie Mellon University kunz1@cmu.edu SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR6 Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAINING11 Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain GRAPH TRANSFORMER NETWORKS signiﬁcantly affected by the choice of meta-paths. Unlike these approaches, our Graph Transformer Networks can operate on a heterogeneous graph and transform the graph for tasks while learning PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

prints well.

2019 CONFERENCE

main conference.

LOGIN - NEURIPS.CC

Do not remove: This comment is monitored to verify that the site is

working properly

REVIEW FOR NEURIPS PAPER: TRAINING GENERATIVE ADVERSARIAL Review 1. Summary and Contributions: This paper proposes a new effective method for training GANs from small size of data by incorporating non-leaking data augmentation.For achieving this, the authors design an adaptive discriminator augmentation (ADA). They extensively analyze why data augmentation can harm the GAN performance with diverse experiments. REVIEW FOR NEURIPS PAPER: UNFOLDING THE ALTERNATING Review 1. Summary and Contributions: This paper tackles Blind-SR by unfolding the two steps approach of IKC (Gu'19) into one trainable network that requires no test-time optimization.. Strengths: 1.Simple and elegant way to idealize the two steps. 2. Impressive results, large meaningful margin. Weaknesses: All weaknesses are related to experiments, analysis and understanding. REVIEW FOR NEURIPS PAPER: UNCERTAINTY-AWARE LEARNING FOR Review 1. Summary and Contributions: The paper proposes to include in zero-shots semantic segmentation uncertainty estimation elements to model labeling and prediction uncertainties.The experimental results show that including these Bayesian uncertainty estimation elements provide noticeable gains in the obtained zero-shot segmentations. REVIEW FOR NEURIPS PAPER: TASK-ROBUST MODEL-AGNOSTIC META Review 1. Summary and Contributions: The submission proposes a modification of the multi-task meta-learning objective from the average of the per-task losses to the maximum of those losses.The argument is that this will force the learner to learn all tasks to a comparable amount, even the worst-case ones, so no task can be

ignored.

2021 CONFERENCE

navigation.

CALL FOR DATASETS BENCHMARKS NeurIPS 2021 Datasets and Benchmarks Track. The Datasets and Benchmarks track serves as a novel venue for high-quality publications, talks, and posters on highly valuable machine learning datasets and benchmarks, as well as a forum for discussions on how to improve dataset development. Datasets and benchmarks are crucial for the development of machine learning methods, but also require their PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

A CAUSAL VIEW ON ROBUSTNESS OF NEURAL NETWORKS2 A Causal View on Robustness of Neural Networks Cheng Zhang Microsoft Research Cheng.Zhang@microsoft.com Kun Zhang Carnegie Mellon University kunz1@cmu.edu SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR6 Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAINING11 Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain GRAPH TRANSFORMER NETWORKS signiﬁcantly affected by the choice of meta-paths. Unlike these approaches, our Graph Transformer Networks can operate on a heterogeneous graph and transform the graph for tasks while learning

2021 CONFERENCE

navigation.

CALL FOR DATASETS BENCHMARKS NeurIPS 2021 Datasets and Benchmarks Track. The Datasets and Benchmarks track serves as a novel venue for high-quality publications, talks, and posters on highly valuable machine learning datasets and benchmarks, as well as a forum for discussions on how to improve dataset development. Datasets and benchmarks are crucial for the development of machine learning methods, but also require their PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

A CAUSAL VIEW ON ROBUSTNESS OF NEURAL NETWORKS2 A Causal View on Robustness of Neural Networks Cheng Zhang Microsoft Research Cheng.Zhang@microsoft.com Kun Zhang Carnegie Mellon University kunz1@cmu.edu SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR6 Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAINING11 Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain GRAPH TRANSFORMER NETWORKS signiﬁcantly affected by the choice of meta-paths. Unlike these approaches, our Graph Transformer Networks can operate on a heterogeneous graph and transform the graph for tasks while learning PAPERINFORMATION / PAPERCHECKLIST The NeurIPS Paper Checklist is designed to encourage best practices for responsible machine learning research, addressing issues of reproducibility, transparency, research ethics, and societal impact. For each question in the checklist: You should answer yes, no, or n/a. You should reference the section (s) of the paper that provide support

for

prints well.

2019 CONFERENCE

main conference.

LOGIN - NEURIPS.CC

Do not remove: This comment is monitored to verify that the site is

working properly

REVIEW FOR NEURIPS PAPER: TRAINING GENERATIVE ADVERSARIAL Review 1. Summary and Contributions: This paper proposes a new effective method for training GANs from small size of data by incorporating non-leaking data augmentation.For achieving this, the authors design an adaptive discriminator augmentation (ADA). They extensively analyze why data augmentation can harm the GAN performance with diverse experiments. REVIEW FOR NEURIPS PAPER: UNFOLDING THE ALTERNATING Review 1. Summary and Contributions: This paper tackles Blind-SR by unfolding the two steps approach of IKC (Gu'19) into one trainable network that requires no test-time optimization.. Strengths: 1.Simple and elegant way to idealize the two steps. 2. Impressive results, large meaningful margin. Weaknesses: All weaknesses are related to experiments, analysis and understanding. REVIEW FOR NEURIPS PAPER: UNCERTAINTY-AWARE LEARNING FOR Review 1. Summary and Contributions: The paper proposes to include in zero-shots semantic segmentation uncertainty estimation elements to model labeling and prediction uncertainties.The experimental results show that including these Bayesian uncertainty estimation elements provide noticeable gains in the obtained zero-shot segmentations. REVIEW FOR NEURIPS PAPER: TASK-ROBUST MODEL-AGNOSTIC META Review 1. Summary and Contributions: The submission proposes a modification of the multi-task meta-learning objective from the average of the per-task losses to the maximum of those losses.The argument is that this will force the learner to learn all tasks to a comparable amount, even the worst-case ones, so no task can be

ignored.

REVIEW FOR NEURIPS PAPER: TASK-ROBUST MODEL-AGNOSTIC META Meta Review. This paper stirred a lot of discussion between reviewers. The reviewers appreciated the new experiments on MiniImagenet. The primary outstanding reviewer concerns were: (a) limited motivation for using a max loss (particularly since, if tasks are of varying difficulty, then the max loss will focus solely on the hardest task rather than taking all tasks into account) (b) the NEURIPS - 2020 CONFERENCEFUTURE MEETINGSMY REGISTRATIONSCREATE PROFILEPAPERSRESET PASSWORDVIDEOLAKE Conference Sessions, Tutorials, Workshops and Expo. Mon Dec 6th through Tue the 14th. Abstract Submission Deadline. May 21 '21 08:00 PM UTC *. 00 weeks 00 days 00:00:00. Applications for Workshops Open. May 28 '21 04:00 PM UTC *. 00 weeks 04 days 12:07:12. Paper submission and co-author registration deadline. 2021 DATES AND DEADLINES Oct 06 '21 (Anywhere on Earth) PaperSubmission. Abstract Submission Deadline. May 21 '21 08:00 PM UTC. Paper submission and co-author registration deadline. May 28 '21 08:00 PM UTC.

2020 CONFERENCE

Neil D. Lawrence, Cambridge University. Daniel D. Lee, University of Pennsylvania. Marc'Aurelio Ranzato, Facebook (term beginning in 2021) Masashi Sugiyama, RIKEN & The University of Tokyo. Ulrike von Luxburg, University of Tübingen (term ending in 2020) Hanna Wallach, Microsoft

Research.

PAPERINFORMATION / ETHICSREVIEW PaperInformation / EthicsReview. NeurIPS | 2021. Thirty-fifth Conference on Neural Information Processing Systems. Toggle

navigation.

for

prints well.

SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAINING Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain TRANSFER LEARNING VIA $\ELL_1$ REGULARIZATION Critical issues lie in how to effectively adapt models under an ever-changing environment. We propose a method for transferring knowledge from a source domain to a target domain via ℓ1 ℓ 1 regularization in high dimension. We incorporate ℓ1 ℓ 1 regularization of differences between source and target parameters in addition to an ordinary

CONCRETE DROPOUT

Concrete Dropout Yarin Gal yarin.gal@eng.cam.ac.uk University of Cambridge and Alan Turing Institute, London Jiri Hron jh2084@cam.ac.uk University of Cambridge NEURIPS - 2020 CONFERENCEFUTURE MEETINGSMY REGISTRATIONSCREATE PROFILEPAPERSRESET PASSWORDVIDEOLAKE Conference Sessions, Tutorials, Workshops and Expo. Mon Dec 6th through Tue the 14th. Abstract Submission Deadline. May 21 '21 08:00 PM UTC *. 00 weeks 00 days 00:00:00. Applications for Workshops Open. May 28 '21 04:00 PM UTC *. 00 weeks 04 days 12:07:12. Paper submission and co-author registration deadline. 2021 DATES AND DEADLINES Oct 06 '21 (Anywhere on Earth) PaperSubmission. Abstract Submission Deadline. May 21 '21 08:00 PM UTC. Paper submission and co-author registration deadline. May 28 '21 08:00 PM UTC.

2020 CONFERENCE

Research.

PAPERINFORMATION / ETHICSREVIEW PaperInformation / EthicsReview. NeurIPS | 2021. Thirty-fifth Conference on Neural Information Processing Systems. Toggle

navigation.

for

prints well.

SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAINING Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain TRANSFER LEARNING VIA $\ELL_1$ REGULARIZATION Critical issues lie in how to effectively adapt models under an ever-changing environment. We propose a method for transferring knowledge from a source domain to a target domain via ℓ1 ℓ 1 regularization in high dimension. We incorporate ℓ1 ℓ 1 regularization of differences between source and target parameters in addition to an ordinary

CONCRETE DROPOUT

Concrete Dropout Yarin Gal yarin.gal@eng.cam.ac.uk University of Cambridge and Alan Turing Institute, London Jiri Hron jh2084@cam.ac.uk University of Cambridge

LOGIN - NEURIPS

Submit. Calls 2021. Call for Papers. Paper FAQ. Become or recommend a revewer. Call for Tutorials. Call for Competitions. Call for Demonstrations. Call for Workshops. NEURIPS 2021 CALL FOR PAPERS Call For Papers Abstract submission deadline: Wednesday, May 19, 2021 01:00 PM PDT Friday, May 21, 2021 01:00 PM PDT Full paper submission and co-author registration deadline: Wednesday, May 26, 2021 01:00 PM PDT Friday, May 28, 2021 01:00 PM PDT Supplementary material submission deadline: Wednesday, June 2, 2021 01:00 PM PDT Friday, June 4, 2021 01:00 PM PDT

LIST OF PROCEEDINGS

Advances in Neural Information Processing Systems 7 (NIPS 1994) Advances in Neural Information Processing Systems 6 (NIPS 1993) Advances in Neural Information Processing Systems 5 (NIPS 1992) Advances in Neural Information Processing Systems 4 (NIPS 1991) Advances in Neural Information Processing Systems 3 (NIPS 1990)

Advances in Neural

REGISTRATION

Registration Instructions. Update your profile. We now allow you to choose non-binary genders, pronouns, and dietary restrictions in your profile. Password must be 9 characters if creating a profile. Email yourself a receipt in section 3. Registration Cancellation Policy. Registrations are not Transferable. It is not necessary to register

for

2019 CONFERENCE

main conference.

PRICING - NEURIPS.CC Pricing. NeurIPS | 2019. Thirty-third Conference on Neural Information Processing Systems. Toggle navigation. Dates. Schedule. Full Schedule (mobile friendly) Multitrack Schedule. Conference Program PDF. LEARNING WITH OPTIMIZED RANDOM FEATURES: EXPONENTIAL Authors. Hayata Yamasaki, Sathyawageeswar Subramanian, Sho Sonoda, Masato Koashi. Abstract. Kernel methods augmented with random features give scalable algorithms for learning from big data. COMMUNITY DETECTION USING FAST LOW-CARDINALITY SEMIDEFINITE Algorithm 1 The Leiden-Locale method 1: procedure LEIDEN-LOCALE(Graph G, Partition P) 2: do 3: E LocaleEmbeddings (G;P).Replace the LocalMove in Leiden 4: PLocaleRounding(G ;E) 5: G ;P done LeidenRefineAggregate(G;P) . 6: while not done 7: return P 8: end procedure Because we still use the reﬁnement step from the Leiden algorithm, we have the following guarantee. SANITY-CHECKING PRUNING METHODS: RANDOM TICKETS CAN WIN Sanity-Checking Pruning Methods: Random Tickets can Win the Jackpot Jingtong Su1 ;Yihang Chen2 Tianle Cai3 4 Tianhao Wu2 Ruiqi Gao3 ;4 Liwei Wang5 6 y Jason D. Lee3;y 1Yuanpei College, Peking University 2School of Mathematical Sciences, Peking University 3Department of Electrical Engineering, Princeton University 4Zhongguancun Haihua Institute for Frontier Information Technology NEURIPS - 2020 CONFERENCEFUTURE MEETINGSMY REGISTRATIONSCREATE PROFILEPAPERSRESET PASSWORDVIDEOLAKE General Chair. Marc'Aurelio Ranzato, Facebook AI Research. Program Chair. Alina Beygelzimer, Yahoo Research. Program Co-chairs. Percy Liang, Stanford University 2021 DATES AND DEADLINES Datasets and Benchmarks Datasets and Benchmarks Submission deadline (1st round) Jun 07 '21(Anywhere on Earth) Dataset and Benchmarks Reviews released and start of discussions NEURIPS 2020 : PAPERS This is the public, feature-limited version of the conference webpage. After Registration and login please visit the full version.

2020 CONFERENCE

submission!

PAPERINFORMATION / ETHICSREVIEW Do not remove: This comment is monitored to verify that the site is

working properly

SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAININGCONFIDENCE REGULARIZED SELF TRAININGSELF TRAINING MEANINGSELF TRAINING MODELTRID TRAINING SELF CERTIFICATION ANSWERSSELF TRAINING A SERVICE DOG Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain TRANSFER LEARNING VIA $\ELL_1$ REGULARIZATION Do not remove: This comment is monitored to verify that the site is

working properly

CONCRETE DROPOUT

Concrete Dropout Yarin Gal yarin.gal@eng.cam.ac.uk University of Cambridge and Alan Turing Institute, London Jiri Hron jh2084@cam.ac.uk University of Cambridge NEURIPS - 2020 CONFERENCEFUTURE MEETINGSMY REGISTRATIONSCREATE PROFILEPAPERSRESET PASSWORDVIDEOLAKE General Chair. Marc'Aurelio Ranzato, Facebook AI Research. Program Chair. Alina Beygelzimer, Yahoo Research. Program Co-chairs. Percy Liang, Stanford University 2021 DATES AND DEADLINES Datasets and Benchmarks Datasets and Benchmarks Submission deadline (1st round) Jun 07 '21(Anywhere on Earth) Dataset and Benchmarks Reviews released and start of discussions NEURIPS 2020 : PAPERS This is the public, feature-limited version of the conference webpage. After Registration and login please visit the full version.

2020 CONFERENCE

submission!

PAPERINFORMATION / ETHICSREVIEW Do not remove: This comment is monitored to verify that the site is

working properly

SELF-SUPERVISED GRAPH TRANSFORMER ON LARGE-SCALE MOLECULAR Self-Supervised Graph Transformer on Large-Scale Molecular Data Yu Rong 1, Yatao Bian , Tingyang Xu , Weiyang Xie , Ying Wei1, Wenbing Huang2y,Junzhou Huang1 1Tencent AI Lab 2 Beijing National Research Center for Information Science and Technology(BNRist), Department of Computer Science and Technology, Tsinghua University RETHINKING PRE-TRAINING AND SELF-TRAININGCONFIDENCE REGULARIZED SELF TRAININGSELF TRAINING MEANINGSELF TRAINING MODELTRID TRAINING SELF CERTIFICATION ANSWERSSELF TRAINING A SERVICE DOG Rethinking Pre-training and Self-training Barret Zoph⇤, Golnaz Ghiasi ⇤, Tsung-Yi Lin ⇤, Yin Cui, Hanxiao Liu, Ekin D. Cubuk, Quoc V. Le Google Research, Brain TRANSFER LEARNING VIA $\ELL_1$ REGULARIZATION Do not remove: This comment is monitored to verify that the site is

working properly

CONCRETE DROPOUT

Concrete Dropout Yarin Gal yarin.gal@eng.cam.ac.uk University of Cambridge and Alan Turing Institute, London Jiri Hron jh2084@cam.ac.uk University of Cambridge

2020 CONFERENCE

Post Conference Announcements. If you need a certificate of attendance, they're available from the Registration page in Section 3.Click the button "Email Certificate of Attendance." All conference content, except discussions held in RocketChat will become free to the public in late January. NEURIPS 2021 CALL FOR PAPERS Call For Papers Abstract submission deadline: Wednesday, May 19, 2021 01:00 PM PDT Friday, May 21, 2021 01:00 PM PDT Full paper submission and co-author registration deadline: Wednesday, May 26, 2021 01:00 PM PDT Friday, May 28, 2021 01:00 PM PDT Supplementary material submission deadline: Wednesday, June 2, 2021 01:00 PM PDT Friday, June 4, 2021 01:00 PM PDT

LIST OF PROCEEDINGS

Advances in Neural Information Processing Systems 33 (NeurIPS 2020) Advances in Neural Information Processing Systems 32 (NeurIPS 2019) Advances in Neural Information Processing Systems 31

REGISTRATION

If you are open to being recruited by a sponsor, see section 6. Student ID required to get student discount. See more. Early and late pricing cost the same.; Agree to the code of conduct, and privacy policy. Please watch our video titled, "Where Everyone Has a Voice" NEURIPS 2020 CALL FOR PAPERS Do not remove: This comment is monitored to verify that the site is

working properly

2019 CONFERENCE

Registration. To register using access to reserved tickets, you must be logged in. - General Admission Lottery. If you are not an author, workshop organizer, tutorial speaker or invited speaker, then you should join the registraiton lottery.Read a press release about the

lottery here

LEARNING WITH OPTIMIZED RANDOM FEATURES: EXPONENTIAL Authors. Hayata Yamasaki, Sathyawageeswar Subramanian, Sho Sonoda, Masato Koashi. Abstract. Kernel methods augmented with random features give scalable algorithms for learning from big data. COMMUNITY DETECTION USING FAST LOW-CARDINALITY SEMIDEFINITE Algorithm 1 The Leiden-Locale method 1: procedure LEIDEN-LOCALE(Graph G, Partition P) 2: do 3: E LocaleEmbeddings (G;P).Replace the LocalMove in Leiden 4: PLocaleRounding(G ;E) 5: G ;P done LeidenRefineAggregate(G;P) . 6: while not done 7: return P 8: end procedure Because we still use the reﬁnement step from the Leiden algorithm, we have the following guarantee. SANITY-CHECKING PRUNING METHODS: RANDOM TICKETS CAN WIN Sanity-Checking Pruning Methods: Random Tickets can Win the Jackpot Jingtong Su1 ;Yihang Chen2 Tianle Cai3 4 Tianhao Wu2 Ruiqi Gao3 ;4 Liwei Wang5 6 y Jason D. Lee3;y 1Yuanpei College, Peking University 2School of Mathematical Sciences, Peking University 3Department of Electrical Engineering, Princeton University 4Zhongguancun Haihua Institute for Frontier Information Technology

Toggle navigation

Toggle navigation Login

__

NEURIPS | 2020

Thirty-fourth Conference on Neural Information Processing Systems

Year (2020)

* 2020

*

* 2019

*

* 2018

*

* 2017

*

* 2016

*

* 2015

*

* 2014

*

* 2013

*

* 2012

*

* 2011

*

* 2010

*

* 2009

*

* 2008

*

* 2007

*

* 2006

*

* Earlier Conferences

Help

* FAQ

*

* Contact NeurIPS

*

* NeurIPS Foundation

*

* Edit Profile

*

* Create an Account

*

* Reset Password

*

* __ Merge Profiles

*

* Privacy Policy

* My Registrations

Profile

* Edit Profile

*

* Change Password

*

* __ Merge Profiles

*

* __ Create New Profile

*

* __ Reset Password

*

* Log In

*

* Log Out

Contact NeurIPS Sponsor Info Publications Future Meetings

Video Archives

Diversity & Inclusion

New in ML Code of

Conduct About Us NeurIPS Blog

Press

Board 2020

Toggle navigation

* Schedule

* Dates

* Submit

* CALLS 2020

* Call for Papers

* Paper FAQ * Recommend or become a reviewer * Call for Tutorials * Call for Competitions * Call for Demonstrations * Call for Workshops * Workshop FAQ

* Call for Meetups

* Call for Socials

* Call for Expo Non-Profit Proposals

*

* AUTHOR RESOURCES * NeurIPS 2020 CMT Site * AC and SAC Guidelines * Reviewer Guidelines

* Subject Areas

* Style Files

* Funding Disclosure * Code Submission Policy * Conflicts of Interest

* Organizers

* NeurIPS Board

*

* Organizing Committee

*

* Program Committee

*

* NeurIPS Foundation NeurIPS Thirty-fourth Annual Conference on Neural Information

Processing Systems

NeurIPS 2020 is a Virtual-only Conference Sun Dec 6th through Sat the 12th (Sunday is an industry expo)

SCHEDULE

Virtual Conference Schedule » More information about the schedule and virtual conference is available in this blog post

.

The schedule is beta and may not be in it's final form.

REGISTRATION

Sun

EXPO Day

Mon

Tutorials

Tue-Thr

Main Conference

Fri-Sat

Workshops

Pricing » Registration 2020 » Registration will provide access to the workshops, posters, demonstrations, socials, expo, competitions, and the option to be included in recruiting materials provided to our sponsors. The online tutorials, keynotes and orals will be made available free of charge. NeurIPS is proud to provide registration fee assistance to individuals for whom these fees would cause a financial burden. Students with accepted papers are particularly encouraged to apply. Applications close on 11/27/2020. Financial Assistance Application »

The Foundation »

ANNOUNCEMENTS

* July 27, 2020 -- Check out our blog post

for

this year's list of invited speakers! * June 12, 2020 -- NeurIPS 2020 will be held entirely online. See

our blog post

for more information. * June 2, 2020 -- Important notice to all authors: the paper submission deadline has been extended by 48 hours. The new deadline is Friday June 5, 2020 at 1pm PDT. More information can be found at

this page .

* May 20, 2020 -- To all authors: in the event that there are technical problems with CMT during paper submission, please check here and on Twitter (@NeurIPSConf

) for updates.

* Apr 17, 2020 -- Notice to all authors: the submission deadline has been extended by three weeks. Check the updated Call for Papers and the NeurIPS-FAQ . More information can be found in our blog , in the post titled, Extension to Submission Deadline

.

* See the Submit menu above for Call for Papers, Call for Tutorials, and Call for Workshops, and the Call For Meetups and Socials.

IMPORTANT DATES

Expo (Industry) Day

Sun Dec. 6th

Conference Sessions, Tutorials, Workshops and Expo Mon Dec 7th through Sat the 12th Abstract Submission Deadline May 27 01:00 PM PDT * 00 weeks 00 days 00:00:00 Paper Submission Deadline Jun 05 01:00 PM PDT * 00 weeks 00 days 00:00:00 Tutorial Application Deadline Jun 17 04:59 PM PDT * 00 weeks 00 days 00:00:00 Review Period Begins Jul 06 12:00 PM PDT * 00 weeks 00 days 00:00:00

Review Period Ends

Jul 24 01:00 PM PDT * 00 weeks 00 days 00:00:00 Author Response Begins Aug 07 01:00 PM PDT * 00 weeks 00 days 00:00:00 Author Response Ends Aug 13 01:00 PM PDT * 00 weeks 00 days 00:00:00 Demonstration Proposal Deadline Aug 28 (Anywhere on Earth) 00 weeks 00 days 00:00:00 Volunteer and Financial Assistant Applications Open Sep 01 10:40 AM PDT * 00 weeks 00 days 00:00:00

Registration Opens

Sep 15 01:00 AM PDT * 00 weeks 00 days 00:00:00

Author Notification

Sep 25 08:00 PM PDT * DemonstrationNotifications Oct 06 (Anywhere on Earth) 00 weeks 00 days 00:00:00 Volunteer and Travel Application Deadline Oct 08 11:59 PM PDT * 00 weeks 00 days 00:00:00 Camera Ready Paper Deadline Oct 22 01:00 PM PDT * 00 weeks 00 days 00:00:00 Meetup Submission Deadline Oct 25 (Anywhere on Earth) 00 weeks 00 days 00:00:00 Socials Submission Deadline Oct 29 (Anywhere on Earth) 00 weeks 00 days 00:00:00 Socials Notification Nov 09 12:00 AM PST * Financial Assistance Application Closes Nov 27 10:00 AM PST * 00 weeks 00 days 00:00:00 SlidesLive stops accepting recordings Dec 01 04:00 AM PST * 00 weeks 00 days 00:00:00

RegistrationClose

Dec 07 06:00 PM PST * 00 weeks 06 days 10:55:59

All dates »

* Dates above are in pacific time

ABOUT NEURIPS

The purpose of the Neural Information Processing Systems annual meeting is to foster the exchange of research on neural information processing systems in their biological, technological, mathematical, and theoretical aspects. The core focus is peer-reviewed novel research which is presented and discussed in the general session, along with invited talks by leaders in their field. On Sunday is an Expo, where our top industry sponsors give talks, panels, demos, and workshops on topics that are of academic interest. On Monday are tutorials, which cover a broad background on current lines of inquiry, affinity group meetings, and the opening talk & reception. The general sessions are held Tuesday - Thursday, and include talks, posters, and demonstrations. Friday - Saturday are the workshops, which are smaller meetings focused on current topics, and provide an informal, cutting edge venue for discussion.

__

#neurips2020

Successful Page Load Do not remove: This comment is monitored to verify that the site is

working properly

Details

Image Url

HTML Url

Moderation By

More Annotations

David Lawrence

2020-02-28 15:12:56

David Lawrence

2020-02-28 15:13:20

David Lawrence

2020-02-28 15:13:21

David Lawrence

2020-02-28 15:13:46

David Lawrence

2020-02-28 15:13:51

David Lawrence

2020-02-28 15:14:15

David Lawrence

2020-02-28 15:14:23

David Lawrence

2020-02-28 15:14:40

David Lawrence

2020-02-28 15:14:40

David Lawrence

2020-02-28 15:14:57

David Lawrence

2020-02-28 15:15:03

David Lawrence

2020-02-28 15:15:12

Favourite Annotations

David Lawrence

2021-06-06 14:29:11

David Lawrence

2021-06-06 14:29:12

David Lawrence

2021-06-06 14:29:12

David Lawrence

2021-06-06 14:29:13

David Lawrence

2021-06-06 14:29:13

David Lawrence

2021-06-06 14:29:14

David Lawrence

2021-06-06 14:29:14

David Lawrence

2021-06-06 14:29:15

David Lawrence

2021-06-06 14:29:16

David Lawrence

2021-06-06 14:29:16

David Lawrence

2021-06-06 14:29:16

David Lawrence

2021-06-06 14:29:18

Text

2021 CONFERENCE

navigation.

for

prints well.

2021 CONFERENCE

navigation.

for

prints well.

2019 CONFERENCE

main conference.

for

prints well.

self-supervision.

PAPERS.NEURIPS.CC

papers.neurips.cc

2021 CONFERENCE

navigation.

for

2021 CONFERENCE

navigation.

for

for

2019 CONFERENCE

main conference.

LOGIN - NEURIPS.CC

working properly

prints well.

PAPERS.NEURIPS.CC

papers.neurips.cc