Advances in Information Retrieval: 36th European Conference by Maarten de Rijke, Tom Kenter, Arjen P. de Vries, ChengXiang

By Maarten de Rijke, Tom Kenter, Arjen P. de Vries, ChengXiang Zhai, Franciska de Jong, Kira Radinsky, Katja Hofmann

This publication constitutes the court cases of the thirty sixth ecu convention on IR examine, ECIR 2014, held in Amsterdam, The Netherlands, in April 2014.

The 33 complete papers, 50 poster papers and 15 demonstrations provided during this quantity have been rigorously reviewed and chosen from 288 submissions. The papers are equipped within the following topical sections: evaluate, suggestion, optimization, semantics, aggregation, queries, mining social media, electronic libraries, potency, and knowledge retrieval concept. additionally incorporated are three instructional and four workshop presentations.

Show description

Read or Download Advances in Information Retrieval: 36th European Conference on IR Research, ECIR 2014, Amsterdam, The Netherlands, April 13-16, 2014. Proceedings PDF

Similar human-computer interaction books

Computer Vision for Human-Machine Interaction

Contemporary advances within the box of computing device imaginative and prescient are resulting in novel and radical adjustments within the approach we engage with pcs. it is going to quickly be attainable to allow a working laptop or computer associated with a video digital camera to become aware of the presence of clients, tune faces, fingers and arms in genuine time, and research expressions and gestures.

Plan, Activity, and Intent Recognition: Theory and Practice

Plan popularity, task reputation, and motive reputation jointly mix and unify innovations from consumer modeling, computer imaginative and prescient, clever consumer interfaces, human/computer interplay, self reliant and multi-agent platforms, normal language realizing, and desktop studying. Plan, job, and rationale acceptance explains the the most important position of those options in a wide selection of functions together with: .

Online Social Media Analysis and Visualization

This edited quantity addresses the tremendous demanding situations of adapting on-line Social Media (OSM) to constructing examine equipment and purposes. the subjects disguise producing lifelike social community topologies, understanding of consumer actions, subject and pattern iteration, estimation of person attributes from their social content material, habit detection, mining social content material for universal traits, choosing and rating social content material assets, construction friend-comprehension instruments, and so forth.

Additional info for Advances in Information Retrieval: 36th European Conference on IR Research, ECIR 2014, Amsterdam, The Netherlands, April 13-16, 2014. Proceedings

Example text

In [1], Amati and van Rijsbergen proposed the Divergence from Randomness (DFR) framework to construct probabilistic term weighting schemes. The framework was also inspired by Harter’s 2-Poisson model [11] and further refined the notion of eliteness using semantic information theory and Popper’s notion of information content. A DFR model is typically composed of two divergence functions (referred to as P rob1 to characterise the randomness of a term, and P rob2 to model the risk of using the term as a document descriptor [1]) and a normalisation function.

Our methods beat the Condorcet method on all four of the test collections that we tried, and beat Soboroff’s random-voting method on three of four. For incomplete judgments, we find that with a limited judging budget, a few hundred documents for example, RTC will be the better choice. However, the high computational complexity of RTC may make it impractical for interactive use with large collections. If more documents are judged, a few thousand for example, our results indicate that using EM’s learning framework with Hedge’s loss function would be a good choice.

Nuray and Can’s Condorcet method used system document rankings to rank documents by predicted relevance, and then assigned binary relevance to documents up to some cutoff [11]. Hauff et al. compared all the above methods across 16 different test collections (from TREC and elsewhere), finding Soboroff et al’s random-voting method best on nine collections, and Nuray and Can’s Condorcet method best on six [8]. Hosseini et al. used the EM framework to solve the problem of acquiring relevance judgements for Book Search tasks through crowdsourcing when no true relevance labels are available [9].

Download PDF sample

Rated 4.60 of 5 – based on 47 votes