Unsupervised Adaptation Across Domain Shifts by Generating Intermediate Data Representations

By Raghuraman Gopalan, Ruonan Li, Ramalingam Chellappa. 2014

213 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 Adaptation for Visual Applications: A Comprehensive Survey (2017, methodology): “To better take advantage of the presence of multiple source domains, extensions to methods previously designed for a single source versus target case were proposed in [107, 27, 116, 265].”
An introduction to domain adaptation and transfer learning (2018, background): “This assumption is useful when selecting a set of transformations as a geodesic path along the manifold [82,81,84,6,36].”
Domain adaptive sparse representation-based classification (2015, methodology): “Rather than working with the information conveyed by the source and target domains alone, [4] proposed an incremental learning technique based on gradually following the geodesic path between the source and target domains.”
Deep domain adaptation for describing people based on fine-grained clothing attributes (2015, methodology): “Many methods have been proposed for domain adaptation in visual recognition [16, 18, 15].”
Source Data-Absent Unsupervised Domain Adaptation Through Hypothesis Transfer and Labeling Transfer (2020, background): “ous visual applications like object recognition [2], [39], [42] and semantic segmentation [3], [8], [9], [43].”
Visual Domain Adaptation: A survey of recent advances (2015, methodology): “The manifold-based approach [34] outperforms the existing unsupervised domain adaptation methods in two of the three source-target combinations.”
Deep Domain Adaptation in Action Space (2018, methodology): “In the subspace based approach [11, 16], a linear transformation is learned to align the source and target distribution.”
A Review of Single-Source Deep Unsupervised Visual Domain Adaptation (2020, methodology): “Single-source UDA methods can be divided into two categories [26, 27]: sample re-weighting [28, 29] and intermediate subspace transformation [30–34].”
Structure-Preserved Multi-source Domain Adaptation (2016, background): “the adaptation process to shift the discriminative function from the source domain and applies it for the target domain [35]; Low-rank Transfer Subspace Learning (LTSL) first learns a subspace with the low-rank constraints, then applies PCA or LDA for the adaptation [29]; SGF trains the discriminati”
Analysing Domain Shift Factors between Videos and Images for Object Detection (2015, unknown): “Hence, these works operate in a domain adaptation setting [12], [20], [29].”
Incremental Evolving Domain Adaptation (2016, methodology): “struct data representation in which the discrepancy between source and target distributions is minimized by means of a linear projection [8], [9], [10] or a kernel function [7].”
Person Re-Identification by Camera Correlation Aware Feature Augmentation (2017, background): “In a broader context, our crossview adaptation for re-id is related to but different vitally from domain adaptation [63], [64], [65], [66], [67], [68], [69], [70].”
A Review of Domain Adaptation without Target Labels (2019, background): “Every point along that path could generate an intermediate domain [129], [130].”