publications | Fei Sha

2023

arxiv

SEEDS: Emulation of Weather Forecast Ensembles with Diffusion Models

Lizao Li, Rob Carver, Ignacio Lopez-Gomez, and 2 more authors

2023

@misc{li23seeds,
  title = {{SEEDS}: Emulation of Weather Forecast Ensembles with Diffusion Models},
  author = {Li, Lizao and Carver, Rob and Lopez-Gomez, Ignacio and Sha, Fei and Anderson, John},
  year = {2023},
  eprint = {2306.14066},
  archiveprefix = {arXiv},
}

ICCV

Encyclopedic VQA: Visual questions about detailed properties of fine-grained categories

Thomas Mensink, Jasper Uijlings, Lluis Castrejon, and 6 more authors

In Proc. of ICCV, 2023

Bib

@inproceedings{mensink23vqa,
  title = {{Encyclopedic VQA}: Visual questions about detailed properties of fine-grained categories},
  booktitle = {Proc. of ICCV},
  author = {Mensink, Thomas and Uijlings, Jasper and Castrejon, Lluis and Goel, Arushi and Cadar, Felipe and Zhou, Howard and Sha, Fei and Araujo, Andr\'{e} and Ferrari, Vittorio},
  year = {2023},
  eprint = {2306.09224},
  archiveprefix = {arXiv}
}

arxiv

Neural Ideal Large Eddy Simulation: Modeling Turbulence with Neural Stochastic Differential Equations

Anudhyan Boral, Zhong Yi Wan, Leonardo Zepeda-Núñez, and 5 more authors

2023

Bib

@misc{boral23les,
  title = {Neural Ideal Large Eddy Simulation: Modeling Turbulence with Neural Stochastic Differential Equations},
  author = {Boral, Anudhyan and Wan, Zhong Yi and Zepeda-N\'{u}\~{n}ez, Leonardo and Lottes, James and Wang, Qing and Chen, Yi-fan and Anderson, John Roberts and Sha, Fei},
  year = {2023},
  eprint = {2306.01174},
  archiveprefix = {arXiv}
}

arxiv

Debias Coarsely, Sample Conditionally: Statistical Downscaling through Optimal Transport and Probabilistic Diffusion Models

Zhong Yi Wan, Ricardo Baptista, Yi-fan Chen, and 4 more authors

2023

Bib

@misc{wan23debias,
  title = {Debias Coarsely, Sample Conditionally: Statistical Downscaling through Optimal Transport and Probabilistic Diffusion Models},
  author = {Wan, Zhong Yi and Baptista, Ricardo and Chen, Yi-fan and Anderson, John and Boral, Anudhyan and Sha, Fei and Zepeda-N\'{u}\~{n}ez, Leonardo},
  year = {2023},
  eprint = {2305.15618},
  archiveprefix = {arXiv}
}

ICML

User-defined Event Sampling and Uncertainty Quantification in Diffusion Models for Physical Dynamical Systems

Marc Finzi, Anudhyan Boral, Andrew Gordon Wilson, and 2 more authors

In ICML, 2023

Bib

@inproceedings{finzi23user,
  title = {User-defined Event Sampling and Uncertainty Quantification in Diffusion Models for Physical Dynamical Systems},
  author = {Finzi, Marc and Boral, Anudhyan and Wilson, Andrew Gordon and Sha, Fei and Zepeda-N\'{u}\~{n}ez, Leonardo},
  booktitle = {ICML},
  year = {2023},
}

ICML

Pre-computed memory or on-the-fly encoding? A hybrid approach to retrieval augmentation makes the most of your compute

Michiel Jong, Yury Zemlyanskiy, Nicholas FitzGerald, and 4 more authors

In ICML, 2023

Bib

@inproceedings{dejong23lumen,
  title = {Pre-computed memory or on-the-fly encoding? A hybrid approach to retrieval augmentation makes the most of your compute},
  author = {de Jong, Michiel and Zemlyanskiy, Yury and Fitz{G}erald, Nicholas and Ainslie, Joshua and Sanghai, Sumit and Sha, Fei and Cohen, William},
  booktitle = {ICML},
  year = {2023},
}

ACL

FiDO: Fusion-in-Decoder optimized for stronger performance and faster inference

Michiel Jong, Yury Zemlyanskiy, Joshua Ainslie, and 4 more authors

In Findings of ACL, 2023

Bib

@inproceedings{dejong23fido,
  author = {de Jong, Michiel and Zemlyanskiy, Yury and Ainslie, Joshua and FitzGerald, Nicholas and Sanghai, Sumit and Sha, Fei and Cohen, William},
  booktitle = {Findings of ACL},
  year = {2023},
}

ICLR

Evolve Smoothly, Fit Consistently: Learning Smooth Latent Dynamics For Advection-Dominated Systems

Zhong Yi Wan, Leonardo Zepeda-Núñez, Anudhyan Boral, and 1 more author

In ICLR, 2023

Bib

@inproceedings{wan23evolve,
  author = {Wan, Zhong Yi and Zepeda-N\'{u}\~{n}ez, Leonardo and Boral, Anudhyan and Sha, Fei},
  title = {Evolve Smoothly, Fit Consistently: Learning Smooth Latent Dynamics For Advection-Dominated Systems},
  booktitle = {ICLR},
  year = {2023},
}

2022

PAMI

Drinking from a firehose: Continual learning with web-scale natural language

Hexiang Hu, Ozan Sener, Fei Sha, and 1 more author

IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022

Bib

@article{hu22firehose,
  title = {Drinking from a firehose: Continual learning with web-scale natural language},
  author = {Hu, Hexiang and Sener, Ozan and Sha, Fei and Koltun, Vladlen},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2022},
  volume = {45},
  no = {5},
}

EMNLP

Evaluating the Impact of Model Scale for Compositional Generalization in Semantic Parsing

Linlu Qiu, Peter Shaw, Panupong Pasupat, and 5 more authors

In EMNLP, 2022

Bib

@inproceedings{liu22evaluating,
  author = {Qiu, Linlu and Shaw, Peter and Pasupat, Panupong and Shi, Tianze and Herzig, Jonathan and Pitler, Emily and Sha, Fei and Toutanova, Kristina},
  title = {Evaluating the Impact of Model Scale for Compositional Generalization in Semantic Parsing},
  booktitle = {EMNLP},
  year = {2022},
}

COLING

Generate-and-Retrieve: use your predictions to improve retrieval for semantic parsing

Yury Zemlyanskiy, Michiel Jong, Joshua Ainslie, and 5 more authors

In COLING, 2022

Bib

@inproceedings{zemlyansky22generate,
  author = {Zemlyanskiy, Yury and de Jong, Michiel and Ainslie, Joshua and Pasupat, Panupong and Shaw, Peter and Qiu, Linlu and Sanghai, Sumit and Sha, Fei},
  title = {Generate-and-Retrieve: use your predictions to improve retrieval for semantic parsing},
  booktitle = {COLING},
  year = {2022},
}

NeurIPS

ALMA: Hierarchical Learning for Composite Multi-Agent Tasks

Shariq Iqbal, Robby Costales, and Fei Sha

In NeurIPS, 2022

Bib

@inproceedings{iqbal22alma,
  author = {Iqbal, Shariq and Costales, Robby and Sha, Fei},
  title = {{ALMA}: Hierarchical Learning for Composite Multi-Agent Tasks
  },
  booktitle = {NeurIPS},
  year = {2022},
}

ICLR

Mention Memory: incorporating textual knowledge into Transformers through entity mention attention

Michiel Jong, Yury Zemlyanskiy, Nicholas FitzGerald, and 2 more authors

In ICLR, 2022

Bib

@inproceedings{dejong22mention,
  title = {Mention Memory: incorporating textual knowledge into Transformers through entity mention attention},
  author = {de Jong, Michiel and Zemlyanskiy, Yury and FitzGerald, Nicholas and Sha, Fei and Cohen, William},
  booktitle = {ICLR},
  year = {2022},
}

ICLR

Possibility Before Utility: Learning And Using Hierarchical Affordances

Robby Costales, Shariq Iqbal, and Fei Sha

2022

Bib

@inproceedinngs{costales22affordance,
  title = {Possibility Before Utility: Learning And Using Hierarchical Affordances},
  author = {Costales, Robby and Iqbal, Shariq and Sha, Fei},
  booktitle = {ICLR},
  year = {2022},
}

AISTAS

Policy Learning and Evaluation with Randomized Quasi-Monte Carlo

Sebastien M. R. Arnold, Pierre L’Ecuyer, Liyu Chen, and 2 more authors

In AISTATS, 2022

Bib

@inproceedings{arnold22policy,
  title = {Policy Learning and Evaluation with Randomized Quasi-Monte Carlo},
  author = {Arnold, Sebastien M. R. and L'Ecuyer, Pierre and Chen, Liyu and Chen, Yi-fan and Sha, Fei},
  booktitle = {AISTATS},
  year = {2022},
}

2021

EMNLP

Systematic Generalization on gSCAN: what is nearly solved and what is next?

Linlu Qiu, Hexiang Hu, Bowen Zhang, and 2 more authors

In Proc. of EMNLP, 2021

Bib

@inproceedings{qiu21gscan,
  author = {Qiu, Linlu and Hu, Hexiang and Zhang, Bowen and Shaw, Pete and Sha, Fei},
  title = {Systematic Generalization on {gSCAN}: what is nearly solved and what is next?},
  booktitle = {Proc. of EMNLP},
  year = {2021},
}

EMNLP

Visually Grounded Concept Composition

Bowen Zhang, Hexiang Hu, Linlu Qiu, and 2 more authors

In Findings of EMNLP, 2021

Bib

@inproceedings{zhang21crg,
  author = {Zhang, Bowen and Hu, Hexiang and Qiu, Linlu and and Pete Shaw and Sha, Fei},
  title = {Visually Grounded Concept Composition},
  booktitle = {Findings of EMNLP},
  year = {2021},
}

ICML

Randomized Entity-wise Factorization for Multi-Agent Reinforcement Learning

Shariq Iqbal, Christian A. Schroeder Witt, Bei Peng, and 3 more authors

In Proc. of Intl. Conf. on Machine Learning (ICML), 2021

Bib

@inproceedings{iqbal2020aiqmix,
  title = {Randomized Entity-wise Factorization for Multi-Agent Reinforcement Learning},
  author = {Iqbal, Shariq and de Witt, Christian A. Schroeder and Peng, Bei and B\"{o}hmer, Wendelin and Whiteson, Shimon and Sha, Fei},
  year = {2021},
  booktitle = {Proc. of Intl. Conf. on Machine Learning (ICML)},
}

AISTATS

When MAML Can Adapt Fast and How to Assist When It Cannot

Sebastien M. R. Arnold, Shariq Iqbal, and Fei Sha

In AISTATS, 2021

Bib

@inproceedings{arnold2021maml,
  title = {When MAML Can Adapt Fast and How to Assist When It Cannot},
  author = {Arnold, Sebastien M. R. and Iqbal, Shariq and Sha, Fei},
  booktitle = {AISTATS},
  year = {2021},
}

ICDAR

CoMSum and SIBERT: A Dataset and Neural Model for Query-Based Multi-Document Summarization

Sayali Kulkarn, Sheide Chammas, Wan Zhu, and 2 more authors

In Proceedings of Intl. Conf. on Document Analysis and Recognition, 2021

Bib

@inproceedings{kulkarni21comsum,
  author = {Kulkarn, Sayali and Chammas, Sheide and Zhu, Wan and Sha, Fei and Ie, Eugene},
  title = { {CoMSum and SIBERT}: A Dataset and Neural Model for Query-Based Multi-Document Summarization},
  booktitle = {Proceedings of Intl. Conf. on Document Analysis and Recognition},
  year = {2021},
}

EACL

DOCENT: Learning Self-Supervised Entity Representations from Large Document Collections

Yury Zemlyanskiy, Sudeep Gandhe, Ruining He, and 5 more authors

In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, Apr 2021

Bib

@inproceedings{zemlyanskiy21docent,
  title = {{DOCENT}: Learning Self-Supervised Entity Representations from Large Document Collections},
  author = {Zemlyanskiy, Yury and Gandhe, Sudeep and He, Ruining and Kanagal, Bhargav and Ravula, Anirudh and Gottweis, Juraj and Sha, Fei and Eckstein, Ilya},
  booktitle = {Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume},
  month = apr,
  year = {2021},
  address = {Online},
  publisher = {Association for Computational Linguistics},
  pages = {2540--2549},
}

NAACL

ReadTwice: Reading Very Large Documents with Memories

Yury Zemlyanskiy, Joshua Ainslie, Michiel Jong, and 3 more authors

In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Jun 2021

Bib

@inproceedings{zemlyanskiy21readtwice,
  title = {{R}ead{T}wice: Reading Very Large Documents with Memories},
  author = {Zemlyanskiy, Yury and Ainslie, Joshua and de Jong, Michiel and Pham, Philip and Eckstein, Ilya and Sha, Fei},
  booktitle = {Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
  month = jun,
  year = {2021},
  address = {Online},
  pages = {5189--5195},
}

2020

Sci. R
Deep learned tissue "fingerprints" classify breast cancers by ER/PR/Her2 status from H&E images

Rishi R Rawat, Itzel Ortega, Preeyam Roy, and 4 more authors

Scientific reports, Jun 2020

Abs Bib

Because histologic types are subjective and difficult to reproduce between pathologists, tissue morphology often takes a back seat to molecular testing for the selection of breast cancer treatments. This work explores whether a deep-learning algorithm can learn objective histologic H&E features that predict the clinical subtypes of breast cancer, as assessed by immunostaining for estrogen, progesterone, and Her2 receptors (ER/PR/Her2). Translating deep learning to this and related problems in histopathology presents a challenge due to the lack of large, well-annotated data sets, which are typically required for the algorithms to learn statistically significant discriminatory patterns. To overcome this limitation, we introduce the concept of “tissue fingerprints,” which leverages large, unannotated datasets in a label-free manner to learn H&E features that can distinguish one patient from another. The hypothesis is that training the algorithm to learn the morphological differences between patients will implicitly teach it about the biologic variation between them. Following this training internship, we used the features the network learned, which we call “fingerprints,” to predict ER, PR, and Her2 status in two datasets. Despite the discovery dataset being relatively small by the standards of the machine learning community (n = 939), fingerprints enabled the determination of ER, PR, and Her2 status from whole slide H&E images with 0.89 AUC (ER), 0.81 AUC (PR), and 0.79 AUC (Her2) on a large, independent test set (n = 2531). Tissue fingerprints are concise but meaningful histopathologic image representations that capture biological information and may enable machine learning algorithms that go beyond the traditional ER/PR/Her2 clinical groupings by directly predicting theragnosis.
@article{rawat20fingerprints, title = {{Deep learned tissue "fingerprints" classify breast cancers by ER/PR/Her2 status from H\&E images}}, author = {Rawat, Rishi R and Ortega, Itzel and Roy, Preeyam and Sha, Fei and Shibata, Darryl and Ruderman, Daniel and Agus, David B}, affiliation = {Lawrence J. Ellison Institute for Transformative Medicine, University of Southern California, 12414 Exposition Blvd, Los Angeles, CA, 90064, USA. DASH Center at USC, 1002 Childs Way, MCB 114, Los Angeles, CA, 90089-0005, USA. Department of Pathology, University of Southern California Health Sciences Campus, NOR 1441 Eastlake Ave, Los Angeles, 90033, USA. Lawrence J. Ellison Institute for Transformative Medicine, University of Southern California, 12414 Exposition Blvd, Los Angeles, CA, 90064, USA. ruderman@usc.edu.}, journal = {Scientific reports}, volume = {10}, number = {1}, pages = {7275}, year = {2020}, language = {en}, issn = {2045-2322}, pmid = {32350370}, }

IJCV

Classifier and Exemplar Synthesis for Zero-Shot Learning

Soravit Changpinyo, Wei-Lun Chao, Boqing Gong, and 1 more author

Int. J. Comput. Vis., Jun 2020

Bib

@article{changpinyo20classifier,
  author = {Changpinyo, Soravit and Chao, Wei{-}Lun and Gong, Boqing and Sha, Fei},
  title = {Classifier and Exemplar Synthesis for Zero-Shot Learning},
  journal = {Int. J. Comput. Vis.},
  volume = {128},
  number = {1},
  pages = {166--201},
  year = {2020},
}

EMNLP

Learning to Represent Image and Text with Denotation Graph

Bowen Zhang, Hexiang Hu, Vihan Jain, and 2 more authors

In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), Nov 2020

Bib

@inproceedings{zhang20vdg,
  title = {Learning to Represent Image and Text with Denotation Graph},
  author = {Zhang, Bowen and Hu, Hexiang and Jain, Vihan and Ie, Eugene and Sha, Fei},
  booktitle = {Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)},
  month = nov,
  year = {2020},
  address = {Online},
  publisher = {Association for Computational Linguistics},
  pages = {823--839},
}

ACL

BabyWalk: Going Farther in Vision-and-Language Navigation by Taking Baby Steps

Wang Zhu, Hexiang Hu, Jiacheng Chen, and 4 more authors

In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, ACL 2020, Online, July 5-10, 2020, Nov 2020

Bib

@inproceedings{zhu20babywalk,
  author = {Zhu, Wang and Hu, Hexiang and Chen, Jiacheng and Deng, Zhiwei and Jain, Vihan and Ie, Eugene and Sha, Fei},
  editor = {Jurafsky, Dan and Chai, Joyce and Schluter, Natalie and Tetreault, Joel R.},
  title = {BabyWalk: Going Farther in Vision-and-Language Navigation by Taking
                 Baby Steps},
  booktitle = {Proceedings of the 58th Annual Meeting of the Association for Computational
                 Linguistics, {ACL} 2020, Online, July 5-10, 2020},
  pages = {2539--2556},
  publisher = {Association for Computational Linguistics},
  year = {2020},
}

CVPR

Few-Shot Learning via Embedding Adaptation With Set-to-Set Functions

Han-Jia Ye, Hexiang Hu, De-Chuan Zhang, and 1 more author

In The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun 2020

Bib

@inproceedings{ye20fsl,
  author = {Ye, Han-Jia and Hu, Hexiang and Zhang, De-Chuan and Sha, Fei},
  title = {Few-Shot Learning via Embedding Adaptation With Set-to-Set Functions},
  booktitle = {The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  month = jun,
  year = {2020},
}

AISTATS

Amortized Inference of Variational Bounds for Learning Noisy-OR

Yiming Yan, Melissa Ailem, and Fei Sha

In Proc. of AISTATS, Jun 2020

Bib

@inproceedings{yan20noisy,
  title = {Amortized Inference of Variational Bounds for Learning Noisy-OR},
  author = {Yan, Yiming and Ailem, Melissa and Sha, Fei},
  booktitle = {Proc. of AISTATS},
  year = {2020},
}

2019

ICML

Actor-Attention-Critic for Multi-Agent Reinforcement Learning

Shariq Iqbal, and Fei Sha

In Proceedings of the 36th International Conference on Machine Learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA, Jun 2019

Bib

@inproceedings{iqbal19acc,
  author = {Iqbal, Shariq and Sha, Fei},
  editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan},
  title = {Actor-Attention-Critic for Multi-Agent Reinforcement Learning},
  booktitle = {Proceedings of the 36th International Conference on Machine Learning,
                 {ICML} 2019, 9-15 June 2019, Long Beach, California, {USA}},
  series = {Proceedings of Machine Learning Research},
  volume = {97},
  pages = {2961--2970},
  publisher = {{PMLR}},
  year = {2019},
}

2016

CVPR

Synthesized classifiers for zero-shot learning

Soravit Changpinyo, Weilun Chao, Boqing Gong, and 1 more author

In Proc. of CVPR, Jun 2016

Bib

@inproceedings{changpinyo16zero,
  author = {Changpinyo, Soravit and Chao, Weilun and Gong, Boqing and Sha, Fei},
  title = {Synthesized classifiers for zero-shot learning},
  booktitle = {Proc. of CVPR},
  year = {2016},
}

2013

NeurIPS

Similarity Component Analysis

Soravit Changpinyo, Kuan Liu, and Fei Sha

In Proc. of Annual Conference on Neural Information Processing Systems (NIPS), Jun 2013

Bib

@inproceedings{chanpingyo13sca,
  author = {Changpinyo, Soravit and Liu, Kuan and Sha, Fei},
  title = {Similarity Component Analysis},
  booktitle = {Proc. of Annual Conference on Neural Information Processing Systems (NIPS)},
  year = {2013},
}

2012

CVPR

Geodesic Flow Kernel for Unsupervised Domain Adaptation

Boqing Gong, Yuan Shi, Fei Sha, and 1 more author

In Proceedings of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), Jun 2012

Bib

@inproceedings{gong12gfk,
  author = {Gong, Boqing and Shi, Yuan and Sha, Fei and Grauman, Kristen},
  title = {Geodesic Flow Kernel for Unsupervised Domain Adaptation},
  booktitle = {Proceedings of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR)},
  address = {Providence, Rhode Island},
  year = {2012}
}

2010

NeurIPS

Unsupervised Kernel Dimension Reduction

Meihong Wang, Fei Sha, and Michael I. Jordan

In Proceedings of Neural Information Processing (NIPS), Jun 2010

Bib

@inproceedings{wang10ukdr,
  author = {Wang, Meihong and Sha, Fei and Jordan, Michael I.},
  title = {Unsupervised Kernel Dimension Reduction},
  booktitle = {Proceedings of Neural Information Processing (NIPS)},
  year = {2010},
}

2007

NeurIPS

Large margin hidden Markov models for automatic speech recognition

Fei Sha, and Lawrence K. Saul

In Advances in Neural Information Processing Systems 19, Jun 2007

Bib

@inproceedings{sha07hmm,
  author = {Sha, Fei and Saul, Lawrence K.},
  title = {Large margin hidden {M}arkov models for automatic speech recognition},
  booktitle = {Advances in Neural Information Processing Systems 19},
  year = {2007},
  editor = {Sch\"{o}lkopf, B. and Platt, J.C. and Hofmann, T.},
  pages = {1249--1256},
  address = {Cambridge, MA},
  publisher = {MIT Press},
  owner = {feisha},
}

2003

NAACL-HLT

Shallow Parsing with Conditional Random Fields

Fei Sha, and Fernando Pereira

In Proceedings of Human Language Technology-NAACL 2003, Jun 2003

Bib

@inproceedings{sha03shallow,
  author = {Sha, Fei and Pereira, Fernando},
  title = {Shallow Parsing with Conditional Random Fields},
  booktitle = {Proceedings of Human Language Technology-NAACL 2003},
  year = {2003},
  pages = {213--220},
  address = {Edmonton, Canada}
}