Contextual correlates of synonymy

By H. Rubenstein, John B. Goodenough. 1965

Experimentol corroboration was obtained for the hypothesis that the proportion of words common to the contexts of word A and to the contexts of word B is a function of the degree to which A and B are similar in meaning. The tests were carried out for variously defined contexts. The shapes of the functions, however, indicate that similarity of context is reliable as criterion only for detecting pairs of words that are very similar in meaning.

Subfields: Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

1. Mohit Bansal
2. Graham Neubig
3. Christopher D. Manning
4. Noah A. Smith
5. Luke S. Zettlemoyer
6. Mirella Lapata
7. Richard Socher
8. Iryna Gurevych

Citations in the corpus

• GloVe: Global Vectors for Word Representation (2014, background): “These include WordSim-353 (Finkelstein et al., 2001), MC (Miller and Charles, 1991), RG (Rubenstein and Goodenough, 1965), SCWS (Huang et al., 2012), and RW (Luong et al., 2013).”
• WiC: the Word-in-Context Dataset for Evaluating Context-Sensitive Meaning Representations (2018, methodology): “The Stanford Contextual Word Similarity (SCWS) dataset (Huang et al., 2012) comprises 2003 word pairs and is analogous to standard word similarity datasets, such as RG-65 (Rubenstein and Goodenough, 1965) and Sim-Lex ( Hill et al., 2015), in which the task is to automatically estimate the semantic s”
• Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings (2016, background): “Second, the vector differences between words in embeddings have been shown to represent relationships between words [32, 26].”
• Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data (2020, background): “And of course it is well established that vector-space representations of words pick up word classes, both syntactic (POS, e.g. Lin et al., 2015) and semantic (lexical similarity, e.g. Rubenstein and Goodenough, 1965; Mikolov et al., 2013).”
• Learning Gender-Neutral Word Embeddings (2018, neutral): “, 2001), RG-65 (Rubenstein and Goodenough, 1965), MTurk-287 (Radinsky et al.”
• Speech2Vec: A Sequence-to-Sequence Framework for Learning Word Embeddings from Speech (2018, positive): “We used 13 benchmarks [30] to measure word similarity, including WS-353 [31], WS-353-REL [32], WS-353-SIM , MC-30 [33], RG-65 [34], Rare-Word [35], MEN [36], MTurk-287 [37], MTurk-771 [38], YP-130 [31], SimLex-999 [39], Verb-143 [40], and SimVerb-3500 [41].”
• SimLex-999: Evaluating Semantic Models With (Genuine) Similarity Estimation (2014, background): “85 (Rubenstein and Goodenough, 1965).”
• Strategies for Pre-training Graph Neural Networks (2019, background): “Our goal is to pre-train a GNN so that it maps nodes appearing in similar structural contexts to nearby embeddings (Rubenstein & Goodenough, 1965; Mikolov et al., 2013).”
• Word Representations via Gaussian Embedding (2014, methodology): “We evaluate the embeddings on seven different standard word similarity benchmarks (Rubenstein & Goodenough, 1965; Szumlanski et al., 2013; Hill et al., 2014; Miller & Charles, 1991; Bruni et al., 2014; Yang & Powers, 2006; Finkelstein et al., 2001).”
• Don’t count, predict! A systematic comparison of context-counting vs. context-predicting semantic vectors (2014, background): “The classic data set of Rubenstein and Goodenough (1965) (rg) consists of 65 noun pairs.”
• Cheap and Fast – But is it Good? Evaluating Non-Expert Annotations for Natural Language Tasks (2008, methodology): “…Pearson correlation; (Miller and Charles, 1991) found a 0.97 correlation of the annotations of 38 subjects with the annotations given by 51 subjects in (Rubenstein and Goodenough, 1965), and a following study (Resnik, 1999) with 10 subjects found a 0.958 correlation with (Miller and Charles, 1991).”
• Multimodal Distributional Semantics (2013, background): “There are several data sets to assess how well a DSM captures human intuitions about semantic relatedness, such as the Rubenstein and Goodenough set (Rubenstein & Goodenough , 1965) and WordSim353 (Finkelstein, Gabrilovich, Matias, Rivlin, Solan, Wolfman, & Ruppin, 2002).”
• Improving Vector Space Word Representations Using Multilingual Correlation (2014, background): “The second and third benchmarks are the RG-65 (Rubenstein and Goodenough, 1965) and the MC-30 (Miller and Charles, 1991) datasets that contain 65 and 30 pairs of nouns respectively and have been given similarity rankings by humans.”
• Distributional Memory: A General Framework for Corpus-Based Semantics (2010, neutral): “Our first challenge comes from the classic data set of Rubenstein and Goodenough (1965), consisting of 65 noun pairs rated by 51 subjects on a 0–4 similarity scale.”
• Jointly optimizing word representations for lexical and sentential tasks with the C-PHRASE model (2015, background): “The only parameters of the model are the word vectors; specifically, no extra parameters are needed for composition (composition models such as the one presented in Socher et al. (2012) require an extra parameter matrix for each word in the vocabulary, and even leaner models such as the one of Gueva”
• A Survey of Cross-lingual Word Embedding Models (2017, methodology): “The most commonly used ones of these human judgement datasets are: a) the RG dataset (Rubenstein & Goodenough, 1965); b) the MC dataset (Miller & Charles, 1991); c) the WordSim-353 dataset (Finkelstein, Gabrilovich, Matias, Rivlin, Solan, Wolfman, & Ruppin, 2002), a superset of MC; and d) the…”
• All-but-the-Top: Simple and Effective Postprocessing for Word Representations (2017, methodology): “For this experiment, we use seven standard datasets: the first published RG65 dataset (Rubenstein and Goodenough, 1965); the widely used WordSim-353 (WS) dataset (Finkelstein et al.”