Multiclass transfer learning from unconstrained priors
By Jie Luo, Tatiana Tommasi, Barbara Caputo. 2011
112 citations recorded by the source corpus.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Yi Yang
- Mingsheng Long
- Jianmin Wang
- Tao A. Xiang
- Gabriella Csurka
- Jiashi Feng
- Kate Saenko
- Barbara Caputo
In Computer Vision and Pattern Recognition
- Dacheng Tao
- Yi Yang
- Luc Van Gool
- Trevor Darrell
- Xiaogang Wang
- Jiashi Feng
- Tao A. Xiang
- Ming-Hsuan Yang
Citations in the corpus
- From N to N+1: Multiclass Transfer Incremental Learning (2013, background): “Success in this setting is defined as optimizing the accuracy of the dog detector, with a minimal number of annotated training samples (Figure 1, left).”
- ImageCLEF 2014: Overview and Analysis of the Results (2014, methodology): “The approach considers source classifiers as experts, and it combines their confidence output with a high-level cue integration scheme, as opposed to a mid-level one as proposed in [10].”
- Harnessing Lab Knowledge for Real-World Action Recognition (2014, background): “– Multiple Kernel Transfer Learning (MKTL) (Luo et al. 2011): a recent multi-class transfer learning algorithm built within a multiple kernel learning framework.”
- Transfer Feature Learning with Joint Distribution Adaptation (2013, background): “Transfer learning has proven to be promising in image classification [24, 12] and tagging [19, 25], object recognition [14, 2, 7, 10], and feature learning [13, 11, 17].”
- Bayesian Joint Modelling for Object Localisation in Weakly Labelled Images (2015, background): “More recently, transductive SVM [37], Multiple Kernel Learning (MKL) [38], and instance constraints [39] have been exploited.”
- Transfer Joint Matching for Unsupervised Domain Adaptation (2014, methodology): “Specifically, TJM aims to reduce the domain difference by jointly matching the features and reweighting the instances across domains in a principled dimensionality reduction procedure, and construct new feature representation that is invariant to both the distribution difference and the irrelevant…”