Advances in Information Retrieval: Recent Research from the by W. Bruce Croft (auth.), W. Bruce Croft (eds.)

By W. Bruce Croft (auth.), W. Bruce Croft (eds.)

The heart for clever info Retrieval (CIIR) used to be shaped within the machine technological know-how division ofthe college ofMassachusetts, Amherst in 1992. The center aid for the heart got here from a countrywide technological know-how Foun- tion State/Industry/University Cooperative examine Center(S/IUCRC) furnish, even supposing there have been a colossal info retrieval (IR) study workforce for over 10 years sooner than that furnish. Thebasic aim ofthese facilities is to mix simple study, utilized examine, and expertise move. The CIIR has been winning in each one of those parts, in that it has produced over 270 examine papers, has been interested by many profitable executive and collaborations, and has had an important position in high-visibility websites and start-ups. because of those efforts, the CIIR has develop into recognized the world over as one of many major study teams within the region of knowledge retrieval. The CIIR specializes in examine that leads to better and effective entry and discovery in huge, heterogeneous, disbursed, textual content and multimedia databases. The scope of the paintings that's performed within the CIIR is vast and is going considerably past “traditional” components of data retrieval reminiscent of retrieval versions, cross-lingual seek, and automated question growth. The examine contains either low-level platforms matters akin to the layout of protocols and architectures for allotted seek, in addition to extra human-centered issues equivalent to consumer interface layout, visualization and knowledge mining with textual content, and multimedia retrieval.

Show description

Read Online or Download Advances in Information Retrieval: Recent Research from the Center for Intelligent Information Retrieval PDF

Best research books

Advances in Down Syndrome Research

"Advances in Down Syndrome study” represents up to date study in numerous components of Down Syndrome (DS). a brand new promising animal version of DS is pronounced and this opens new possibilities to review pathomechanisms and pharmacological methods because it is greater than tough to hold out reviews in people and the scientific positive factors are hugely variable.

Progress in Drug Research / Fortschritte der Arzneimittelforschung / Progrès des recherches pharmaceutiques

Quantity 39 of "Progress in Drug examine" comprises 8 reports and a few of the indexes which facilitate its use and determine the con­ nection with the former volumes. The articles during this quantity care for polyamines as tumor markers; the histamine (H3)-receptor and its function as a normal regulatory method; with cytokines as immunosti­ mulants and immunosuppressants in addition to capability anticancer brokers; with fibrinolysis and medical use of thrombolytic brokers; and with catechol O-methyltransferase and its inhibitors, that are of capability curiosity within the therapy of Parkinson's affliction.

Additional info for Advances in Information Retrieval: Recent Research from the Center for Intelligent Information Retrieval

Sample text

This example shows that the inference net framework can be used for most of the combination processes in Figure 1 . 1. Individual classifiers are built using representation nodes and combination operators, and the output of those classifiers is combined using other operators. e. new operators), new types of representation nodes, and different techniques for computing initial probabilities for representation nodes. The probabilities associated with the query node propositions are computed from the probabilities associated with representation nodes.

4) Desiderata for a Concept of Weight of Evidence Good elucidates three simple, natural desiderata for the formalization of the notion of weight of evidence (Good, 1989; Good, 1983b). 1. 5) For example, let us suppose that a document is evaluated with respect to the query, Clinton impeachment proceedings and we discover the term Hyde in the document: h e = = document is relevant to the query Hyde occurs in the document This first criterion states that the weight of evidence provided by finding Hyde in the document should be some function of the likelihood of finding Hyde in a document that is relevant to the query and the (for this example, 42 ADVANCES IN INFORMATION RETRIEVAL presumably lower) likelihood of finding Hyde in a document that is not relevant to the query.

1995a), assume that because the goal of the system is to rank documents, the parts of a probabilistic formula that are constant for a given query (such as prior probabilities) can be ignored. In addition, ad-hoc (but effective) formulas are used to calculate parts of the document scores. This means that the numbers produced by these systems are not probabilities. This is a significant disadvantage when it comes to combining the output of a system with the output of other systems. Systems with compatible outputs, and accurate probability estimates, will produce the best combinations assuming they are independent.

Download PDF sample

Rated 4.66 of 5 – based on 33 votes