paper-with-me

홈 › Papers

Towards Understanding Linear Word Analogies

2018-10-11 · ACL 2019 7 · Kawin Ethayarajh, David Duvenaud, Graeme Hirst

A surprising property of word vectors is that word analogies can often be solved with vector arithmetic. However, it is unclear why arithmetic operators correspond to non-linear embedding models such as skip-gram with negative sampling (SGNS). We provide a formal explanation of this phenomenon without making the strong assumptions that past theories have made about the vector space and word distribution. Our theory has several implications. Past work has conjectured that linear substructures exist in vector spaces because relations can be represented as ratios; we prove that this holds for SGNS. We provide novel justification for the addition of SGNS word vectors by showing that it automatically down-weights the more frequent word, as weighting schemes do ad hoc. Lastly, we offer an information theoretic interpretation of Euclidean distance in vector spaces, justifying its use in capturing word dissimilarity.

📄 PDF Abstract BibTeX arXiv:1810.04882

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Word Embeddings, Analogies, and Machine Learning: Beyond king - man + woman = queen

2016-12-01 · COLING 2016 12 · Aleks Drozd, R, Anna Gladkova, Satoshi Matsuoka

Solving word analogies became one of the most popular benchmarks for word embeddings on the assumption that linear relations between word pairs (such as \textit{king}:\textit{man} :: \textit{woman}:\textit{queen}) are in…

BIG-bench Machine LearningMorphological AnalysisRetrievalWord Embeddings+1

Analogies Explained: Towards Understanding Word Embeddings

2019-01-28 · Carl Allen, Timothy Hospedales

Word embeddings generated by neural network methods such as word2vec (W2V) are well known to exhibit seemingly linear behaviour, e.g. the embeddings of analogy "woman is to queen as man is to king" approximately describe…

Word Embeddings

Cross-lingual Word Analogies using Linear Transformations between Semantic Spaces

2018-07-11 · Tomáš Brychcín, Stephen Eugene Taylor, Lukáš Svoboda

We generalize the word analogy task across languages, to provide a new intrinsic evaluation method for cross-lingual semantic spaces. We experiment with six languages within different language families, including English…

Temporal Word Analogies: Identifying Lexical Replacement with Diachronic Word Embeddings

2017-07-01 · ACL 2017 7 · Terrence Szymanski

This paper introduces the concept of temporal word analogies: pairs of words which occupy the same semantic space at different points in time. One well-known property of word embeddings is that they are able to effective…

ArticlesDiachronic Word EmbeddingsWord Embeddings

ANALOGICAL -- A Novel Benchmark for Long Text Analogy Evaluation in Large Language Models

2023-05-08 · Thilini Wijesiriwardene, Ruwan Wickramarachchi, Bimal G. Gajera, Shreeyash Mukul Gowaikar 외

Over the past decade, analogies, in the form of word-level analogies, have played a significant role as an intrinsic measure of evaluating the quality of word embedding methods such as word2vec. Modern large language mod…

NegationSentence