Multi‐transfer: Transfer learning with multiple views and multiple sources

Abstract

Transfer learning, which aims to help learning tasks in a target domain by leveraging knowledge from auxiliary domains, has been demonstrated to be effective in different applications such as text mining, sentiment analysis, and so on. In addition, in many real‐world applications, auxiliary data are described from multiple perspectives and usually carried by multiple sources. For example, to help classify videos on Youtube, which include three perspectives: image, voice and subtitles, one may borrow data from Flickr, Last.FM and Google News. Although any single instance in these domains can only cover a part of the views available on Youtube, the piece of information carried by them may compensate one another. If we can exploit these auxiliary domains in a collective manner, and transfer the knowledge to the target domain, we can improve the target model building from multiple perspectives. In this article, we consider this transfer learning problem as Transfer Learning with Multiple Views and Multiple Sources. As different sources may have different probability distributions and different views may compensate or be inconsistent with each other, merging all data in a simplistic manner will not give an optimal result. Thus, we propose a novel algorithm to leverage knowledge from different views and sources collaboratively, by letting different views from different sources complement each other through a co‐training style framework, at the same time, it revises the distribution differences in different domains. We conduct empirical studies on several real‐world datasets to show that the proposed approach can improve the classification accuracy by up to 8% against different kinds of state‐of‐the‐art baselines.

Multitask Learning for Emotionally Analyzing Sexual Abuse Disclosures

“…Multitask Learning Frameworks for learning representations across two different sources within the same domain follow multitask learning (Caruana, 1997). The ability to utilize knowledge from various sources compensates for missing data and complements existing meta-data (Tan et al, 2013;Ding et al, 2014), thus allowing for effective sharing of task-invariant features (Caruana, 1997;Zhang and Wang, 2016;Zhang et al, 2018). MTL has been utilized for name error recognition (Cheng et al, 2015), tagging-chunking (Collobert et al, 2011), machine translation (Luong et al, 2015) and relation extraction (Gupta et al, 2016).…”

Reducing the effect of sample bias for small data sets with double‐weighted support vector transfer regression

“…These approaches can be expressed as instance based or feature based. Instance‐based transfer approaches (such that a portion of sample points from the source domain data can be used in the target domain) have been created, which use boosting (Dai, Yang, Xue, & Yu, 2007; Pardoe & Stone, 2010), multiple input sources (Tan, Zhong, Xiang, & Yang, 2014), and reweighting approaches based on covariate shift setting (Cortes & Mohri, 2014; Gretton et al., 2009; Huang et al., 2007; Sugiyama, Nakajima, Kashima, Buenau, & Kawanabe, 2008). In feature‐based transfer approaches, the source and target data are mapped into a space where the shared information from both data can be applied to the target domain (Argyriou, Evgeniou, & Pontil, 2007, 2008).…”

Fuzzy Multiple-Source Transfer Learning

“…These algorithms reduce negative transfers by increasing the number of sources. Tan et al [28] presented a novel algorithm to leverage knowledge from different views and sources collaboratively by letting different views from different sources complement each other through a co-training style framework to reduce the differences in distribution across different domains. Beyond transferring the source data, Zhuang et al [29] discovered a more powerful feature representation of the data when transferring knowledge from multiple source domains to the target domain.…”