Active learning for ranking through expected loss optimization

Long, Bo; Chapelle, Olivier; Zhang, Ya; Chang, Yi; Zheng, Zhaohui; Tseng, Belle L.

doi:10.1145/1835449.1835495

Cited by 63 publications

(64 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most of the other ranking algorithms such as Rank SVM [9] and Rank Boost [7] suggests to add the most relevant pairs of documents to the training set, the document's predicted relevance scores are very close under the current ranking models. In the terms of binary relevance, greedy algorithm [1] is proposed which selects the document which differentiates two different ranking systems in terms of average precision. The comparison of relevant document selection methodologies in learning to rank are found in [8].…”

Section: Learning To Rank Surveymentioning

confidence: 99%

Ranking Models and Learning to Rank: A Survey

2015

IJSR

View full text Add to dashboard Cite

Current era is of fastest information retrieval. There are lots of research methodologies which are arising to give the fastest and correct result set. This paper mainly focuses on survey of the ranking models and learning to rank technique for giving the effective and efficient information retrieval. Learning to rank is flattering progressively more trendy research area in machine learning. The problem of ranking aims to encourage an ordering or inclination of relation among a set of instances in the input space. Learning to rank for information retrieval has gained a lot of interest in the recent years as ranking is the central problem in many information retrieval applications, like document retrieval, multimedia retrieval, text summarization, collaborative filtering, question answering and online advertising machine translation etc. The tremendously large size of the Web documents makes it usually impracticable for the common users for finding their desired information by surfing the net. As a result, effective and efficient information retrieval is being more important and also search engine (information retrieval system) has turned out to be a vital tool for people to locate their needed information.

show abstract

Section: Learning To Rank Surveymentioning

confidence: 99%

Ranking Models and Learning to Rank: A Survey

2015

IJSR

View full text Add to dashboard Cite

show abstract

“…Margin-based selection criteria seek pairs of instances whose estimated ranks are nearest under the current model [5,37], while others seek examples expected to most influence the ranking function [9] or minimize expected loss [23]. We explore the suitability of margin-based criteria for attribute training, and we propose a new formulation that accounts for diversity.…”

Section: Diversity In Active Learningmentioning

confidence: 99%

Beyond Comparing Image Pairs: Setwise Active Learning for Relative Attributes

Liang

Grauman

2014

2014 IEEE Conference on Computer Vision and Pattern Recognition

View full text Add to dashboard Cite

show abstract

“…However, there is relatively little work on active learning for ranking tasks. One notable exception is [15], who use the notion of Expected Loss Optimization (ELO). Another work in this area is [4], whose aim was to identify the most interesting substances for drug screening using a minimum number of tests.…”

Section: Related Work In Other Scientific Areasmentioning

confidence: 99%