Chang WangChang Wang, Sridhar mahadevanSridhar mahadevan
Problem Introduction Problem Definition Who cares Previous work & challenges Contribution A glance at methods
Example problem Input: Three collections of documents in English (sufficient labels) Italian (sufficient labels) Arabic (few labels). Target: Assign labels to the Arabic documents. A way: find a common feature space for 3 domains Shared labels, (sports, military) No shared documents. (no instance correspondence) No words translations are available.
English docs Italian Docs Arabic docs Shared label set: {sports, military} No corresponding instances or words Question : Can we construct a common feature space so that we can use English docs and Italian docs to help classify Arabic docs? docword1word2…label 102sports 220military sports docparola1parola2…etichetta 120sports 202military sports doc1 كلمة 2 كلمة … ملصق 102sports 220military 310? 401?
English docs Italian docs Arabic docs docword1word2…label 102sports 220military sports docparola1parola2…etichetta 120sports 202military sports doc1 كلمة 2 كلمة … ملصق 102sports 220military 310? 401? docfeature1feature2…label 102sports 220military sports military sports military 1110? 1201? Common feature space
Given K input datasets in different domains, with different features, but all of the datasets shared the same label set. Source domain have sufficient labeled instances. Target domain have few labeled instances. Question: Can we construct a common feature space? So all instances in different domain can be mapped to the same feature space, so that we can perform learning task?
Source k Source 1 Target Common feature space Learning
Search engine classify docs, rank docs, find docs topics Businessman Customer clustering Biologist Match protein
Target domain have little labels No instance correspondence Source domain and target domain have different feature space
Most work assumes that the source domain and the target domain have the same features. Manifold regularization Do not leverage source domain information Transfer learning based on manifold alignment: use both label and unlabeled instance to learn mapping require small amount of instance correspondence.
Transfer learning perspective Can work on different feature space Cope with multiple input domain Can combine with existing domain adaption methods Manifold alignment perspective Need no instance correspondence Use label to learn alignment
English docs Italian docs Arabic docs docword1word2…label 102sports 220military sports docparola1parola2…etichetta 120sports 202military sports doc1 كلمة 2 كلمة … ملصق 102sports 220military 310? 401? docfeature1feature2…label 102sports 220military sports military sports military 1110? 1201? Common feature space