Learning Discriminative Projections for Text Similarity Measures

By Wen-tau Yih, Kristina Toutanova, John C. Platt, Christopher Meek. 2011

265 citations recorded by the source corpus.

Subfields: Information Retrieval & Search, 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 Information Retrieval & Search

In Natural Language Processing

Citations in the corpus

Dense Passage Retrieval for Open-Domain Question Answering (2020, methodology): “The trick of in-batch negatives has been used in the full batch setting (Yih et al., 2011) and more recently for mini-batch (Gillick et al., 2019; Hen-derson et al., 2017).”
Few-shot Learning with Retrieval Augmented Language Models (2022, methodology): “These usually follow a dual-encoder architecture (Yih et al., 2011; Huang et al., 2013; Shen et al., 2014), where queries and passages are encoded independently as vectors, and relevance is computed using the inner product or Euclidean distance.”
Learning Dense Representations for Entity Retrieval (2019, background): “Dual encoder architectures have a long history (Bromley et al., 1994; Chopra et al., 2005; Yih et al., 2011), including for retrieval (Gillick et al., 2018).”
End-to-End Training of Multi-Document Reader and Retriever for Open-Domain Question Answering (2021, background): “There are two broad classes of neural retrievers based on the number of embeddings computed for a document: dual encoders (Yih et al., 2011, Lee et al., 2019) and multivector encoders (Khattab and Zaharia, 2020, Luan et al., 2021).”
Learning Dense Representations of Phrases at Scale (2020, background): “Beyond words, dense representations of many different granularities of text such as sentences (Le and Mikolov, 2014; Kiros et al., 2015) or documents (Yih et al., 2011) have been explored.”
Convolutional Neural Networks for Sentence Classification (2014, background): “…much of the work with deep learning meth-ods has involved learning word vector representations through neural language models (Bengio et al., 2003; Yih et al., 2011; Mikolov et al., 2013) and performing composition over the learned word vectors for classiﬁcation (Collobert et al., 2011).”
Towards Universal Paraphrastic Sentence Embeddings (2015, methodology): “We tune over the strategy in our experiments.”
Learning Continuous Phrase Representations for Translation Modeling (2014, methodology): “The projection matrix is learned on parallel training data using the S2Net algorithm (Yih et al. 2011).”
Multilingual Models for Compositional Distributed Semantics (2014, methodology): “Related to our work, Yih et al. (2011) proposed S2Nets to learn joint embeddings of tf-idf vectors for comparable documents.”
End-to-End Retrieval in Continuous Space (2018, background): “This work and its descendants (Yih et al., 2011; Hu et al., 2014, etc.) refer to the models as siamese networks because two similar objects are encoded by two copies of the same network (all parameters are shared).”
End-to-End Training of Neural Retrievers for Open-Domain Question Answering (2021, background): “7 Related Work (Yih et al., 2011) proposed a discriminative approach to train a retriever by learning dense representations of query and context documents based on word frequency.”
Siamese Recurrent Architectures for Learning Sentence Similarity (2016, neutral): “Siamese neural networks been proposed for a number of language-related metric learning tasks (Yih et al. 2011; Chen and Salman 2011), but to our knowledge, recurrent connections remain largely unexplored in this context.”
Multi-Task Retrieval for Knowledge-Intensive Tasks (2021, background): “…tokens explicitly and can potentially map para-phrases of completely different tokens to close vectors, performance of early dense retrieval methods was often inferior to term-matching approaches, except when large labelled data is available (Yih et al., 2011; Gao et al., 2011; Huang et al., 2013).”
Convolutional 2D Knowledge Graph Embeddings (2017, methodology): “Several convolutional models have been proposed in natural language processing (NLP) for solving a variety of tasks, including semantic parsing (Yih et al. 2011), sentence classification (Kim 2014), search query retrieval (Shen et al.”
Siamese CBOW: Optimizing Word Embeddings for Sentence Representations (2016, neutral): “Yih et al. (2011) also propose a siamese architecture.”
Polarity Inducing Latent Semantic Analysis (2012, methodology): “Along this line of work, Yih et al. (2011) proposed a Siamese neural network approach called S2Net , which tunes the projection matrix directly through gradient descent, and has shown to outperform other methods in several tasks.”
Sketch-based 3D shape retrieval using Convolutional Neural Networks (2015, methodology): “Recently, the Siamese network has been applied to text clas-siﬁcation [28] and speech feature classiﬁcation [4].”