Unsupervised Multi-Target Domain Adaptation: An Information Theoretic Approach

Abstract

Unsupervised domain adaptation (uDA) models focus on pairwise adaptation settings where there is a single, labeled, source and a single target domain. However, in many real-world settings one seeks to adapt to multiple, but somewhat similar, target domains. Applying pairwise adaptation approaches to this setting may be suboptimal, as they fail to leverage shared information among multiple domains. In this work we propose an information theoretic approach for domain adaptation in the novel context of multiple target domains with unlabeled instances and one source domain with labeled instances. Our model aims to find a shared latent space common to all domains, while simultaneously accounting for the remaining private, domainspecific factors. Disentanglement of shared and private information is accomplished using a unified information-theoretic approach, which also serves to establish a stronger link between the latent representations and the observed data. The resulting model, accompanied by an efficient optimization algorithm, allows simultaneous adaptation from a single source to multiple target domains. We test our approach on three challenging publicly-available datasets, showing that it outperforms several popular domain adaptation methods.

Deep visual unsupervised domain adaptation for classification tasks: a survey

“…Liu et al [128] presented a unified feature disentanglement network (UFDN), which learns deep disentangled features for image translation and manipulates image outputs in the multi‐domain scheme. Gholami et al [129] proposed a multi target domain adaptation information theoretic approach (MTDA‐ITA), which makes a solid relationship between the hidden feature spaces and the source data, they utilised a unified approach for disentangling the shared and private knowledge.…”

Artificial Intelligence Approach for Target Classification: A State of The Art

“…According to the data, the performance of support vector machines is of the same order, or even better, then that of a neural network or of a Gaussian mixture model. Also, the timescale characteristics [24], the modulation domain [25], basic function neural networks [26], Rihaczek distribution and Hough transform [27], which is a pattern recognition technique invented in 1959 by Paul Hough, subject to a patent, and used in the processing of digital images. The simplest application can detect lines present in an image, but modifications can be made to this technique to detect other geometric shapes; it is the generalized Hough transform developed by Richard Duda and Peter Hart in 1972 [28][29][30], frequency estimation [28], pulse repetition interval [31], twodimensional bispectrum [32], etc.…”

Adversarial Training Based Multi-Source Unsupervised Domain Adaptation for Sentiment Analysis

“…Here, the two most important points are how to acquire a strong system which can benefit every f j and obtain w j i according to the relations between target instances and source domains. For the first point, we employ an effective multi-task architecture named shared-private model (Bousmalis et al 2016;Liu, Qiu, and Huang 2017;Gholami et al 2018); For the second point, we introduce a weighting scheme to estimate w j i by utilizing the discriminator as a probability distribution estimator. As Figure 1 illustrates, this framework includes K domain-specific extractors E pj K j=1 , a shared feature extractor E s , a classifier C, and a discriminator D.…”