Results 21 to 30 of about 2,585,016 (323)
Knowledge Distillation Meets Self-Supervision [PDF]
European Conference on Computer Vision, 2020 Knowledge distillation, which involves extracting the "dark knowledge" from a teacher network to guide the learning of a student network, has emerged as an important technique for model compression and transfer learning.
Guodong Xu +3 more
semanticscholar +1 more source
Differentiable Volumetric Rendering: Learning Implicit 3D Representations Without 3D Supervision [PDF]
Computer Vision and Pattern Recognition, 2019 Learning-based 3D reconstruction methods have shown impressive results. However, most methods require 3D supervision which is often hard to obtain for real-world datasets. Recently, several works have proposed differentiable rendering techniques to train
Michael Niemeyer +3 more
semanticscholar +1 more source
Libri-Light: A Benchmark for ASR with Limited or No Supervision [PDF]
IEEE International Conference on Acoustics, Speech, and Signal Processing, 2019 We introduce a new collection of spoken English audio suitable for training speech recognition systems under limited or no supervision. It is derived from open-source audio books from the LibriVox project.
Jacob Kahn +14 more
semanticscholar +1 more source
Bayesian Loss for Crowd Count Estimation With Point Supervision [PDF]
IEEE International Conference on Computer Vision, 2019 In crowd counting datasets, each person is annotated by a point, which is usually the center of the head. And the task is to estimate the total count in a crowd scene.
Zhiheng Ma +3 more
semanticscholar +1 more source
Boosting Few-Shot Visual Learning With Self-Supervision [PDF]
IEEE International Conference on Computer Vision, 2019 Few-shot learning and self-supervised learning address different facets of the same problem: how to train a model with little or no labeled data. Few-shot learning aims for optimization methods and models that can learn efficiently to recognize patterns ...
Spyros Gidaris +4 more
semanticscholar +1 more source
Learning by Analogy: Reliable Supervision From Transformations for Unsupervised Optical Flow Estimation [PDF]
Computer Vision and Pattern Recognition, 2020 Unsupervised learning of optical flow, which leverages the supervision from view synthesis, has emerged as a promising alternative to supervised methods. However, the objective of unsupervised learning is likely to be unreliable in challenging scenes. In
L. Liu +9 more
semanticscholar +1 more source
Snorkel: Rapid Training Data Creation with Weak Supervision [PDF]
Proceedings of the VLDB Endowment, 2017 Labeling training data is increasingly the largest bottleneck in deploying machine learning systems. We present Snorkel, a first-of-its-kind system that enables users to train state-of- the-art models without hand labeling any training data.
Alexander J. Ratner +5 more
semanticscholar +1 more source
Distant supervision for relation extraction without labeled data
Annual Meeting of the Association for Computational Linguistics, 2009 Modern models of relation extraction for tasks like ACE are based on supervised learning of relations from small hand-labeled corpora. We investigate an alternative paradigm that does not require labeled corpora, avoiding the domain dependence of ACE ...
Mike D. Mintz +3 more
semanticscholar +1 more source
Ordinal Depth Supervision for 3D Human Pose Estimation [PDF]
2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018 Our ability to train end-to-end systems for 3D human pose estimation from single images is currently constrained by the limited availability of 3D annotations for natural images. Most datasets are captured using Motion Capture (MoCap) systems in a studio
G. Pavlakos +2 more
semanticscholar +1 more source
Bank Regulation and Supervision: What Works Best?
, 2001 The authors draw on their new database on bank regulation and supervision in 107 countries to assess different governmental approaches to bank regulation and supervision and evaluate the efficacy of different regulatory and supervisory policies.
James R. Barth, G. Caprio, Ross Levine
semanticscholar +1 more source