Discriminative models for information retrieval

By Ramesh Nallapati. Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, 2004

333 citations recorded by the source corpus.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/3add8c6c36af4755cecaf994bda0e5a202035811

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

Citations in the corpus

LETOR: Benchmark Dataset for Research on Learning to Rank for Information Retrieval (2007, neutral): “…|í µí±| í µí± í µí± ∈í µí±⋂í µí± Normalized tf [1] L4 log í µí± í µí± í µí± ,í µí± |í µí±| + 1 í µí± í µí± ∈í µí±⋂í µí± SIGIR feature [23] L5 log |í µí° ¶| í µí±í µí± (í µí± í µí± ) í µí± í µí± ∈í µí±⋂í µí± Inverse doc frequency (idf) [1] L6 log log |í µí° ¶| í µí±í µí± (í…”
A discriminative model approach for accurate duplicate bug report retrieval (2010, methodology): “We make use of the recent advances in information retrieval community that uses a classiﬁer to retrieve similar documents from a collection[15].”
dipIQ: Blind Image Quality Assessment by Learning-to-Rank Discriminable Image Pairs (2017, methodology): “Nallapati [72] formulated L2R as a classification problem and investigated the use of maximum entropy and support vector machines (SVMs) to classify each instance into two classes—relevant or irrelevant.”
Listwise approach to learning to rank: theory and algorithm (2008, background): “The pointwise approach (Nallapati, 2004) transforms ranking into regression or classiﬁcation on single objects.”
Predicting a Correct Program in Programming by Example (2015, neutral): “The pointwise approaches transform the ranking problem to a classification [20] or ordinal regression problem [4].”
Learning to Rank for Information Retrieval and Natural Language Processing (2011, neutral): “Learning to Rank Methods • Pointwise Approach – Subset Ranking [Cossock and Zhang, 2006]: Regression – SVM [Nallapati, 2004]: Binary Classification Using SVM – McRank [Li et al 2007]: Multi-Class Classification Using Boosting Tree – Prank [Crammer and Singer 2002]: Ordinal Regression Using Perceptro”
IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models (2017, background): “Generally speaking, there are three types of loss functions designed for learning to rank in web search, namely, pointwise [31], pairwise [3] and listwise [6].”
Learning to Rank by Optimizing NDCG Measure (2009, background): “In the simplest form, the so-called pointwise approaches, ranking can be treated as classiﬁ-cation or regression by learning the numeric rank value of documents as an absolute quantity [3, 4].”
UMass at TREC 2004: Novelty and HARD (2004, methodology): “Of the runs that UMass submitted, several runs involved the use of support vector machines for passage or document retrieval [26].”
Linear feature-based models for information retrieval (2007, methodology): “We first compare a linear feature-based model trained by directly maximizing mean average precision, Nallapati’s SVM model (Nallapati, 2004), and a language modeling system (Ponte and Croft, 1998).”
Learning to efficiently rank (2010, methodology): “This has given rise to a steady stream of eﬀectiveness-centric models, such as language models for information retrieval [20], the BM25 model [21], numerous term proximity models [16, 9, 23, 3], and learning to rank [12, 18, 6, 15].”
A Theoretical Analysis of NDCG Type Ranking Measures (2013, background): “Although early results of learning to rank often reduce ranking problem to classiﬁcation or regression [15, 18, 23, 26, 5], recently there is evidence that learning a ranking function by optimizing a ranking measure such as NDCG is a promising approach [33, 36].”
AdaRank: a boosting algorithm for information retrieval (2007, methodology): “All the existing methods used for document retrieval [2, 3, 8, 13, 16, 20] are designed to optimize loss functions loosely related to the IR performance measures, not loss functions directly based on the measures.”
Discriminative Information Retrieval for Question Answering Sentence Selection (2017, positive): “We generalize Yao et al.’s intuition by casting the problem as an instance of classiﬁcation-based retrieval (Robertson and Sp¨arck Jones, 1976), formalized as a discriminative retrieval model (Cooper et al., 1992; Gey, 1994; Nallapati, 2004) allowing for the use of NLP features.”
A Markov random field model for term dependencies (2005, methodology): “We note that our model shares some similarities, such as the exponential form, with the maximum entropy model presented by Nallapati [14], but is fundamentally different.”
Adapting ranking SVM to document retrieval (2006, background): “In another approach, document retrieval is formalized as a “learning to rank” problem in which documents are mapped into several ordered categories (ranks).”
The Simplest Thing That Can Possibly Work: Pseudo-Relevance Feedback Using Text Classification (2019, methodology): “As an early example, Nallapati [16] used logistic regression and SVMs for document ranking with features based on tf , idf , and statistics derived from their combination.”
Learning to rank for information retrieval (2009, neutral): “In recent years, learning to rank has become a very hot research direction in IR, and a large number of learning-to-rank algorithms have been proposed, such as [9, 13, 14, 16, 17, 26, 29, 33, 34, 47, 49, 59, 63, 73, 78, 90, 97, 99, 102, 114, 119, 122, 129, 134, 136].”
Learning to rank: from pairwise approach to listwise approach (2007, neutral): “The standard features in document retrieval (Nallapati, 2004) are extracted for each query-document pair.”