Outline Morning program Preliminaries Modeling user behavior Semantic matching Learning to rank Afternoon program Entities Generating responses Recommender systems Industry insights Q & A 157
Entities Entities are polysemic “Finding entities” has multiple meanings. Entities can be I nodes in knowledge graphs, I mentions in unstructured texts or queries, I retrievable items characterized by texts. 158
Outline Morning program Preliminaries Modeling user behavior Semantic matching Learning to rank Afternoon program Entities Knowledge graph embeddings Entity mentions in unstructured text Entity finding Generating responses Recommender systems Industry insights Q & A 159
Entities Knowledge graphs 1997 Beyoncé Knowles start date member of Destjny’s Child member of end date Kelly Rowland 2005 member of Michelle Williams Triples (beyonc´ e knowles, member of, destinys child) (kelly rowland, member of, destinys child) (michelle williams, member of, destinys child) (destinys child, start date, 1997) (destinys child, end date, 2005) ... Nice overview on using knowledge bases in IR: [Dietz et al., 2017] 160
Entities Knowlegde graphs Tasks Datatsets I Link prediction WordNet Predict the missing h or t for a triple (h, r, t) (car, hyponym, vehicle) Rank entities by score. Metrics: Freebase/DBPedia (steve jobs, founder of, apple) I Mean rank of correct entity I Hits@10 I Triple classification Predict if (h, r, t) is correct. Metric: accuracy. I Relation fact extraction from free text Use knowledge base as weak supervision for extracting new triples. Suppose some IE system gives us (steve jobs, ‘‘was the initiator of’’, apple) , then we want to predict the founder of relation. 161
Entities Knowlegde graphs Knowledge graph embeddings I TransE [Bordes et al., 2013] I TransH [Wang et al., 2014] I TransR [Lin et al., 2015] 162
Entities TransE “Translation intuition” For a triple (h, l, t) : ~ h + ~ l ≈ ~ t . t i t j l l h i h j 163
Entities TransE “Translation intuition” For a triple (h, l, t) : ~ h + ~ l ≈ ~ t . distance functjon positjve negatjve examples examples [Bordes et al., 2013] 164
Entities TransE t i t j t i “Translation intuition” r?? t j For a triple (h, l, t) : ~ h + ~ l ≈ ~ t . r?? r r h i h i h j How about: t i I one-to-many relations? t j I many-to-many relations? r?? I many-to-one relations? r?? h i h j 165
Entities TransH [Wang et al., 2014] 166
Entities TransH distance functjon [Wang et al., 2014] 167
Entities TransH i.e., translatjon vector d r Constraints is in the hyperplane sofu constraints [Wang et al., 2014] 168
Entities TransR Use di ff erent embedding spaces for entities and relations I 1 entity space I multiple relation spaces I perform translation in appropriate relation space [Lin et al., 2015] 169
Entities TransR [Lin et al., 2015] 170
Entities TransR Relatjons: R d Entjtjes: R k M r = projectjon matrix: k * d Constraints: [Lin et al., 2015] 171
Entities Challenges I How about time? E.g., some relations hold from a certain date, until a certain date. I New entities/relationships I Finding synonymous relationships/duplicate entities (2005, end date, destinys child) (destinys child, disband, 2005) (destinys child, last performance, 2005) I Evaluation Link prediction? Relation classification? Is this fair? As in, is this even possible in all cases (for a human without any world knowledge)? 172
Entities Resources: toolkits + knowledge bases Source Code KB2E : https://github.com/thunlp/KB2E [Lin et al., 2015] TransE : https://everest.hds.utc.fr/ Knowledge Graphs I Google Knowledge Graph google.com/insidesearch/features/search/knowledge.html I Freebase freebase.com I GeneOntology geneontology.org I WikiLinks code.google.com/p/wiki-links 173
Outline Morning program Preliminaries Modeling user behavior Semantic matching Learning to rank Afternoon program Entities Knowledge graph embeddings Entity mentions in unstructured text Entity finding Generating responses Recommender systems Industry insights Q & A 174
Entities Entity mentions Recognition Detect mentions within unstructured text (e.g., query). Linking Link mentions to knowledge graph entities. Utilization Use mentions to improve search. 175
Entities Named entity recognition B−ORG O B−MISC O y B-ORG O B-MISC O O O B-MISC O O h EU rejects German call to boycott British lamb . x rejects EU German call Task vanilla RNN 176
Entities Named entity recognition I A Unified Architecture for Natural Language Processing: Deep Neural Networks with Multitask Learning [Collobert and Weston, 2008] I Natural Language Processing (Almost) from Scratch [Collobert et al., 2011] Feed-forward language model architecture for Learning a single model to solve multiple NLP di ff erent NLP tasks. Taken from [Collobert and tasks. Taken from [Collobert and Weston, Weston, 2008]. 2008]. 177
Entities Named entity recognition B−ORG O B−MISC O CRF CRF CRF forward backward call rejects German EU BI-LSTM-CRF model [Huang et al., 2015] 178
Entities Entity disambiguation I Learn representations for documents and entities. I Optimize a distribution of candidate entities given a document using (a) cross entropy or (b) pairwise loss. Learn initial document representation in Learn similarity between document and entity unsupervised pre-training stage. Taken from representations using supervision. Taken from [He et al., 2013]. [He et al., 2013]. 179
Entities Entity linking Learn representations for the context, the mention, the entity (using surface words) and the entity class. Uses pre-trained word2vec embeddings. Taken from [Sun et al., 2015]. 180
Entities Entity linking Encode Wikipedia descriptions, linked mentions in Wikipedia and fine-grained entity types. All representations are optimized jointly. Taken from [Gupta et al., 2017]. 181
Entities Entity linking A single mention phrase refers to various entities. Multi-Prototype Mention Embedding model that learns multiple sense embeddings for each mention by jointly modeling words from textual contexts and entities derived from a KB. Taken from [Cao et al., 2017]. 182
Entities Improving search using linked entities Attention-based ranking model for word-entity duet. Learn a similarity between query and document entities. Resulting model can be used to obtain ranking signal. Taken from [Xiong et al., 2017a]. 183
Outline Morning program Preliminaries Modeling user behavior Semantic matching Learning to rank Afternoon program Entities Knowledge graph embeddings Entity mentions in unstructured text Entity finding Generating responses Recommender systems Industry insights Q & A 184
Entities Entity finding Task definition Rank entities satisfying a topic described by a few query terms. Not just document search — (a) topics do not typically correspond to entity names, (b) average textual description much longer than typical document. Di ff erent instantiations of the task within varying domains: I Wikipedia: INEX Entity Ranking Track [de Vries et al., 2007, Demartini et al., 2008, 2009, 2010] (lots of text, knowledge graph, revisions) I Enterprise search: expert finding [Balog et al., 2006, 2012] (few entities, abundance of text per entity) I E-commerce: product ranking [Rowley, 2000] (noisy text, customer preferences) 185
Entities Semantic Expertise Retrieval [Van Gysel et al., 2016] I Expert finding is a particular entity retrieval task where there is a lot of text. I Learn representations of words and entities such that n-grams extracted from a document predict the correct expert. Taken from slides of Van Gysel et al. [2016]. 186
Entities Semantic Expertise Retrieval [Van Gysel et al., 2016] (cont’d) I Expert finding is a particular entity retrieval task where there is a lot of text. I Learn representations of words and entities such that n-grams extracted from a document predict the correct expert. Taken from slides of Van Gysel et al. [2016]. 187
Entities Regularities in Text-based Entity Vector Spaces [Van Gysel et al., 2017c] To what extent do entity representation models, trained only on text, encode structural regularities of the entity’s domain? Goal : give insight into learned entity representations. I Clusterings of experts correlate somewhat with groups that exist in the real world. I Some representation methods encode co-authorship information into their vector space. I Rank within organizations is learned (e.g., Professors > PhD students) as senior people typically have more published works. 188
Entities Latent Semantic Entities [Van Gysel et al., 2016] I Learn representations of e-commerce products and query terms for product search. I Tackles learning objective scalability limitations from previous work. I Useful as a semantic feature within a Learning To Rank model in addition to a lexical matching signal. Taken from slides of Van Gysel et al. [2016]. 189
Recommend
More recommend
