Task-agnostic Pretraining

Task-Agnostic Pretraining is a self-supervised learning framework that enables robotic systems to acquire generalizable representations from unlabeled data without reliance on specific downstream task annotations, thereby distinguishing itself from task-specific pretraining methods by promoting broad adaptability across diverse robotic tasks.

Publications

5.8K

Citations

361

Authors

Institutions

Publications per year

2017–2026

Parents

Self-supervised Robot Learning

Publications

MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers

Wenhui Wang, Furu Wei, Li Dong et al. · arXiv (Cornell University) · 2020 · 638 citations · Full text

Computing, Computer Science, Computational Linguistics +12

SUPERB: Speech Processing Universal PERformance Benchmark

Shu-Wen Yang, Po-Han Chi, Yung-Sung Chuang et al. · 2021 · 637 citations

Computing, Computer Science, Computer Vision +11

RT-1: Robotics Transformer for Real-World Control at Scale

Anthony Brohan, Noah Brown, Justice Carbajal et al. · 2023 · 520 citations · Full text

Computing, Computer Science, Robotics Engineering +14

Big Self-Supervised Models are Strong Semi-Supervised Learners

Ting Chen, Simon Kornblith, Kevin Swersky et al. · arXiv (Cornell University) · 2020 · 476 citations · Full text

Computing, Computer Science, Computer Vision +16

RGB-D object recognition and pose estimation based on pre-trained convolutional neural network features

Max Schwarz, Hannes Schulz, Sven Behnke · 2015 · 321 citations

Computing, Computer Science, Robotics Engineering +16

1–5 of 61

Authors

361

FW
Furu Wei
Microsoft Research Asia (China), China
5 publications · 919 citations
5
h-index
HB
Hangbo Bao
Microsoft Research (United Kingdom), United Kingdom
4 publications · 900 citations
4
h-index
LD
Li Dong
Microsoft (United States), United States
4 publications · 900 citations
4
h-index
SW
Shinji Watanabe
Johns Hopkins University, United States
2 publications · 688 citations
2
h-index
KR
Kanishka Rao
Google (United States), United States
2 publications · 559 citations
2
h-index

1–5 of 361

Institutions

1–8 of 88

Venues

arXiv (Cornell University)
Repository · 21 publications · 2K citations
15
h-index
Proceedings of the AAAI Conference on Artificial Intelligence
Conference · 3 publications · 142 citations
3
h-index
Neural Information Processing Systems
Conference · 3 publications · 80 citations
3
h-index
Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
Conference · 2 publications · 109 citations
2
h-index
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
Conference · 1 publications · 105 citations
1
h-index

1–5