Alexander Panchenko Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources
Overview Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 2/54
Making induced senses interpretable [Panchenko et al., 2017b, Panchenko et al., 2017c] Linking induced word senses to lexical resources [Panchenko, 2016, Faralli et al., 2016, Panchenko et al., 2017a, Biemann et al., 2018] Overview Overview Inducing word sense representations : word sense embeddings via retrofjtting [Pelevina et al., 2016, Remus & Biemann, 2018]; inducing synsets [Ustalov et al., 2017b, Ustalov et al., 2017a, Ustalov et al., 2018] inducing semantic classes [Panchenko et al., 2018] Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 3/54
Linking induced word senses to lexical resources [Panchenko, 2016, Faralli et al., 2016, Panchenko et al., 2017a, Biemann et al., 2018] Overview Overview Inducing word sense representations : word sense embeddings via retrofjtting [Pelevina et al., 2016, Remus & Biemann, 2018]; inducing synsets [Ustalov et al., 2017b, Ustalov et al., 2017a, Ustalov et al., 2018] inducing semantic classes [Panchenko et al., 2018] Making induced senses interpretable [Panchenko et al., 2017b, Panchenko et al., 2017c] Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 3/54
Overview Overview Inducing word sense representations : word sense embeddings via retrofjtting [Pelevina et al., 2016, Remus & Biemann, 2018]; inducing synsets [Ustalov et al., 2017b, Ustalov et al., 2017a, Ustalov et al., 2018] inducing semantic classes [Panchenko et al., 2018] Making induced senses interpretable [Panchenko et al., 2017b, Panchenko et al., 2017c] Linking induced word senses to lexical resources [Panchenko, 2016, Faralli et al., 2016, Panchenko et al., 2017a, Biemann et al., 2018] Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 3/54
Inducing word sense representations Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 4/54
Inducing word sense representations Word vs sense embeddings Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 5/54
Inducing word sense representations Word vs sense embeddings Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 6/54
Inducing word sense representations Related work Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 7/54
Inducing word sense representations Related work: knowledge-based AutoExtend [Rothe & Schütze, 2015] * image is reproduced from the original paper Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 8/54
V N C p Y Z X p p z i x i p y ij z i x i wk w i j k z i – a hidden variable: a sense index of word x i in context C ; – a meta-parameter controlling number of senses. See also : [Neelakantan et al., 2014] and [Li and Jurafsky, 2015] Inducing word sense representations Related work: knowledge-free Adagram [Bartunov et al., 2016] Multiple vector representations θ for each word: Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 9/54
See also : [Neelakantan et al., 2014] and [Li and Jurafsky, 2015] Inducing word sense representations Related work: knowledge-free Adagram [Bartunov et al., 2016] Multiple vector representations θ for each word: V N C ∞ p ( Y , Z , β | X , α, θ ) = ∏ ∏ p ( β wk | α ) ∏ [ p ( z i | x i , β ) ∏ p ( y ij | z i , x i , θ )] , w =1 i =1 j =1 k =1 z i – a hidden variable: a sense index of word x i in context C ; α – a meta-parameter controlling number of senses. Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 9/54
Inducing word sense representations Related work: knowledge-free Adagram [Bartunov et al., 2016] Multiple vector representations θ for each word: V N C ∞ p ( Y , Z , β | X , α, θ ) = ∏ ∏ p ( β wk | α ) ∏ [ p ( z i | x i , β ) ∏ p ( y ij | z i , x i , θ )] , w =1 i =1 j =1 k =1 z i – a hidden variable: a sense index of word x i in context C ; α – a meta-parameter controlling number of senses. See also : [Neelakantan et al., 2014] and [Li and Jurafsky, 2015] Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 9/54
Inducing word sense representations Related work: word sense induction Word sense induction (WSI) based on graph clustering : [Lin, 1998] [Pantel and Lin, 2002] [Widdows and Dorow, 2002] Chinese Whispers [Biemann, 2006] [Hope and Keller, 2013] Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 10/54
Inducing word sense representations Related work: Chinese Whispers#1 * source of the image: http://ic.pics.livejournal.com/blagin_anton/33716210/2701748/2701748_800.jpg Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 11/54
Inducing word sense representations Related work: Chinese Whispers#2 Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 12/54
Inducing word sense representations Related work: Chinese Whispers#2 Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 13/54
Inducing word sense representations Related work: Chinese Whispers#2 Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 14/54
Inducing word sense representations Sense embeddings using retrofjtting RepL4NLP@ACL’16 [Pelevina et al., 2016], LREC’18 [Remus & Biemann, 2018] Prior methods: Induce inventory by clustering of word instances Use existing sense inventories Our method: Input: word embeddings Output: word sense embeddings Word sense induction by clustering of word ego-networks Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 15/54
Inducing word sense representations Sense embeddings using retrofjtting From word embeddings to sense embeddings 1 2 Calculate Word Learning Word Vectors Similarity Graph Word Vectors Text Corpus Word Similarity Graph 3 4 Pooling of Word Vectors Word Sense Induction Sense Inventory Sense Vectors Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 16/54
Inducing word sense representations Sense embeddings using retrofjtting Word sense induction using ego-network clustering Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 17/54
table#0 leftmost#0, column#1, tableau#1, indent#1, bracket#3, pointer#0, footer#1, cursor#1, diagram#0, grid#0 table#1 pile#1, stool#1, tray#0, basket#0, bowl#1, bucket#0, box#0, cage#0, saucer#3, mirror#1, pan#1, lid#0 Inducing word sense representations Sense embeddings using retrofjtting Neighbours of Word and Sense Vectors Vector Nearest Neighbors table tray, bottom, diagram, bucket, brackets, stack, basket, list, parenthesis, cup, saucer, pile, playfjeld, bracket, pot, drop-down, cue, plate Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 18/54
Inducing word sense representations Sense embeddings using retrofjtting Neighbours of Word and Sense Vectors Vector Nearest Neighbors table tray, bottom, diagram, bucket, brackets, stack, basket, list, parenthesis, cup, saucer, pile, playfjeld, bracket, pot, drop-down, cue, plate table#0 leftmost#0, column#1, tableau#1, indent#1, bracket#3, pointer#0, footer#1, cursor#1, diagram#0, grid#0 table#1 pile#1, stool#1, tray#0, basket#0, bowl#1, bucket#0, box#0, cage#0, saucer#3, mirror#1, pan#1, lid#0 Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 18/54
Inducing word sense representations Sense embeddings using retrofjtting Word Sense Disambiguation 1 Context extraction : use context words around the target word 2 Context fjltering : based on context word’s relevance for disambiguation 3 Sense choice in context : maximise similarity between a context vector and a sense vector Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 19/54
Inducing word sense representations Sense embeddings using retrofjtting Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 20/54
Inducing word sense representations Sense embeddings using retrofjtting Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 21/54
Inducing word sense representations Sense embeddings using retrofjtting Jan 11, 2018 Inducing Interpretable Word Senses for WSD and Enrichment of Lexical Resources, Alexander Panchenko 22/54
Recommend
More recommend
