Aggregating Continuous Word Embeddings for Information Retrieval

By Stéphane Clinchant, Florent Perronnin. 2013

101 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

Word Representations via Gaussian Embedding (2014, methodology): “Fitting Gaussian mixture models on embeddings has been done in order to apply Fisher kernels to entire documents (Clinchant & Perronnin, 2013b;a).”
Relevance-based Word Embedding (2017, background): “Subsequently, Clinchant and Per-ronnin [5] proposed Fisher Vector (FV), a document representation framework based on continuous word embeddings, which aggregates a non-linear mapping of word vectors into a document-level representation.”
An Introduction to Neural Information Retrieval (2018, unknown): “Non-linear combinations of term vectors—such as using Fisher Kernel Framework [Clinchant and Perronnin, 2013]—have also been explored, as well as other families of aggregate functions of which AWE has been shown to be a special case [Zamani and Croft, 2016b].”
Semantic Matching by Non-Linear Word Transportation for Information Retrieval (2016, methodology): “[5] generated document representations using the Fisher kernel framework.”
Word Embeddings Go to Italy: A Comparison of Models and Training Datasets (2015, background): “Word embeddings have empirically proved to be helpful in the solution on several tasks that span from NLP [17,21] to information retrieval [3,8].”
Neural information retrieval: at the end of the early years (2017, methodology): “TREC 1–2 Ad-hoc (AP 88–89) Clinchant and Perronnin (2013), Zamani and Croft (2016a, b), Zuccon et al.”
Retrieving Textual Evidence for Knowledge Graph Facts (2019, methodology): “Different aggregation strategies including Fisher Vectors are also used to build document vectors [4,25].”
Embedding-based Query Language Models (2016, background): “Clinchant and Perronnin [5] proposed Fisher Vector (FV), a document representation framework based on continuous word embeddings, which aggregates the nonlinear mapping of word vectors into a document-level representation.”
Query Expansion Using Word Embeddings (2016, methodology): “Document embeddings [5, 2] and word-embeddings-based translation models [8, 21] were also used for retrieval.”
Estimating Embedding Vectors for Queries (2016, background): “The Fisher Vector (FV) [8] is a document representation framework based on continuous word embeddings, that aggregates a non-linear mapping of word embedding vectors into a document-level representation.”
Learning Continuous Word Embedding with Metadata for Question Retrieval in Community Question Answering (2015, methodology): “According to the framework of FK (Clinchant and Perronnin, 2013; Sanchez et al., 2013; Zhang et al., 2014b),”
Seeping Semantics: Linking Datasets Using Word Embeddings for Data Discovery (2018, unknown): “For example, in [16], the authors fit a probabilistic model over the corpus of word embeddings and then apply the Fisher kernel framework [17] to combine multiple word embeddings into a single fixeddimensional one.”