Probabilistic models of information retrieval based on measuring the divergence from randomness

By Gianni Amati, C. J. Van Rijsbergen. 2002

968 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems, Information Retrieval & Search, Probabilistic & Bayesian Methods

How researchers see this paper

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

In Artificial Intelligence

In Information Systems

In Information Retrieval & Search

In Probabilistic & Bayesian Methods

Citations in the corpus

Relevance-based Word Embedding (2017, methodology): “…estimation (NCE) [12] which hypothesizes that we can achieve a good model by only differentiating the data from noise via a logistic regression model. main concept in NCE is similar to those proposed in the divergence from randomness model [3] and the divergence minimization feedback model [44].”
Anserini (2018, background): “• PL2 [4] is a representative ranking model from the divergence from randomness framework.”
Aggregating Continuous Word Embeddings for Information Retrieval (2013, methodology): “…on LSI word embeddings signiﬁcantly outperforms LSI, it is out-performed by strong IR baselines such as Diver-gence From Randomness (DFR) models (Amati and Rijsbergen, 2002) or Language Models (Ponte and Croft, 1998 These results are not surprising as it has been shown experimentally in many…”
To Clean or Not to Clean (2018, methodology): “The DFR model 2] contains a detailed discussion about these components).”
Modeling Diverse Relevance Patterns in Ad-hoc Retrieval (2018, methodology): “Specifically, in the local matching layer, the dimension of the spatial GRU is set to 2 which is tuned in [1, 2, 3, 4].”
TOPSIG: topology preserving document signatures (2011, methodology): “These are JelineckMercer (LMJM) [22], DLH13 [12], Divergence from Randomness [1], and Bose-Einstein [1].”
Teaching a New Dog Old Tricks: Resurrecting Multilingual Retrieval Using Zero-Shot Learning (2019, background): “Historically, ad-hoc retrieval systems have been primarily designed, trained, and evaluated on English corpora (e.g., [1, 5, 6, 23] ).”
Medical Question Answering for Clinical Decision Support (2016, neutral): “5) or Dirichlet ( µ =2 , 000) smoothing [39], respectively; and DFR considered the Divergence from Randomness framework [1] with an inverse expected document frequency model for information content, a Bernoulli-process normalization of information gain, and Zipfian term frequency normalization.”
Inferring Query Performance Using Pre-retrieval Predictors (2004, methodology): “For example, the optimal setting for the so-called normalisation 2 in Amati & van Rijsbergen’s probabilistic framework is query-dependent [ 3 ].”
A formal study of information retrieval heuristics (2004, methodology): “Third, we will apply these constraints to many other retrieval models proposed in the literature [1] and different smoothing methods for language models as well [20].”
Graph-of-word and TW-IDF: new approach to ad hoc IR (2013, methodology): “In particular, we highlight vector space model (TF-IDF [30]), probabilistic (BM25 [25]) and language modeling (Dirichlet prior [33]) approaches and the divergence from randomness framework (PL2 [1]).”
Diagnostic Evaluation of Information Retrieval Models (2011, methodology): “…space model (pivoted normalization [Singhal et al. 1996a]), the classic probabilistic retrieval model (Okapi [Robertson and Walker 1994]), the language modeling approach (Dirichlet prior [Zhai and Lafferty 2001a]), and the divergence from randomness approach (PL2 [Amati and Rijsbergen 2002]).”
Statistical Language Models for Information Retrieval: A Critical Review (2008, background): “Authorization does not extend to other kinds of copying, such as that for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale.”
A Markov random field model for term dependencies (2005, neutral): “There is a rich history of statistical models for information retrieval, including the binary independence model (BIM), language modeling [16], inference network model [23], and the divergence from randomness model [1], amongst others [4].”
Overview of Touché 2022: Argument Retrieval (2022, positive): “Additionally, we provided the collection of 868,655 passages expanded with queries generated using the docT5query model [57].”