Domain Adaptation on the Statistical Manifold

By Mahsa Baktash, Mehrtash Harandi, Brian C. Lovell, Mathieu Salzmann. 2014

140 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

In Computer Vision and Pattern Recognition

Citations in the corpus

Domain Generalization for Object Recognition with Multi-task Autoencoders (2015, background): “Many domain adaptation algorithms have been proposed for object recognition (see, e.g ., [2, 31]).”
Domain Adaptation for Visual Applications: A Comprehensive Survey (2017, methodology): “They go a step further in [11] and based on the fact that probability distributions lie on a Riemannian manifold, propose the Statistically Invariant Embedding 7 (SIE) that uses the Hellinger distance on this manifold to compare kernel density estimates between of the source and target data.”
Information Bottleneck Domain Adaptation with Privileged Information for Visual Recognition (2016, unknown): “More recently, [2] applied manifold learning to achieve the above goal by minimizing the Hellinger distance between cross-domain data distributions.”
Visual Domain Adaptation with Manifold Embedded Distribution Alignment (2018, methodology): “l structures [5, 18] that can avoid distortion in the original space. And Gcan facilitate classi•er learning by treating the original d-dimensional subspace (i.e. feature vector) as its basic element [4]. Additionally, feature transformation and distribution alignment o›en have e†cient numerical for”
A Review of Domain Adaptation without Target Labels (2019, methodology): “A further step can be taken by assuming that there exists a manifold of transformations between the source and target domain [94], [124], [125].”
Scatter Component Analysis: A Unified Framework for Domain Adaptation and Domain Generalization (2015, methodology): “One of the methods proposed in [51] follows a similar intuition by using Hellinger distance as an alternative to MMD.”
An introduction to domain adaptation and transfer learning (2018, methodology): “For example, it might consist of a set of camera optics parameters, where each setting would measure the data in one vector space basis [86,7].”
Semi-supervised Domain Adaptation with Subspace Learning for visual recognition (2015, background): “Similar in spirit, Baktashmotlagh et al. [1] made use of the Riemannian metric on the statistical manifold as a measure of distance between the source and target distributions for domain adaptation.”
Transfer Learning via Learning to Transfer (2018, unknown): “• Common latent space based transfer learning algorithms: TCA (Pan et al., 2011), ITL (Shi & Sha, 2012), CMF (Long et al., 2014), LSDT (Zhang et al., 2016), STL (Raina et al., 2007), DIP (Baktashmotlagh et al., 2013) and SIE (Baktashmotlagh et al., 2014).”
A Unified Framework for Metric Transfer Learning (2017, methodology): “Unlike the other baselines, SISS attempts to minimize the distance between the re-weighted source distribution and the target distribution, JDA tries to learn a new feature representation to jointly match both the marginal distributions and conditional distributions.”