
Web search solved%3F all result rankings the same%3F

Active Learning for Building a Corpus of Questions for Parsing.

Finding support sentences for entities.

Entity summarization of news articles.

Caching search engine results over incremental indices.

Early exit optimizations for additive machine learned ranking systems.
Some commercial web search engines rely on sophisticated machine learning systems for ranking web documents. Due to very large collection sizes and tight constraints on query response times, online efficiency of these learning systems forms a bottleneck. An important problem in such systems is to speedup the ranking process without sacrificing much from the quality of results. In this paper, we propose optimization strategies that allow short-circuiting score computations in additive learning systems. The strategies are evaluated over a state-of-the-art machine learning system and a large, real-life query log, obtained from Yahoo!. By the proposed strategies, we are able to speedup the score computations by more than four times with almost no loss in result quality.
Caching search engine results over incremental indices.

Ad%2Dhoc object retrieval in the web of data.

Structure of morphologically expanded queries A genetic algorithm approach.

Introduction.
The use of monads to structure functional programs is described. Monads provide a convenient framework for simulating e ects found in other languages, such as global state, exception handling, output, or non-determinism. Three case studies are looked at in detail: how monads ease the modi cation of a simple evaluator; how monads act as the basis of a datatype of arrays subject to in-place update; and how monads can be used to build parsers.  
Using Semantic Relatedness and Word Sense Disambiguation for %28CL%29IR.

Elhuyar%2DIXA Semantic Relatedness and Cross%2DLingual Passage Retrieval.

Company%2DOriented Extractive Summarization of Financial News.

Semi%2Dparametric and Non%2Dparametric Term Weighting for Information Retrieval.
Abstract. Most of the previous research on term weighting for information retrieval has focused on developing specialized parametric term weighting functions. Examples include T F.IDF vector-space formulations, BM25, and language modeling weighting. Each of these term weighting functions takes on a specific parametric form. While these weighting functions have proven to be highly effective, they impose strict constraints on the functional form of the term weights. Such constraints may possibly degrade retrieval effectiveness. In this paper we propose two new classes of term weighting schemes that we call semi-parametric and nonparametric weighting. These weighting schemes make fewer assumptions about the underlying term weights and allow the data to speak for itself. We argue that these robust weighting schemes have the potential to be significantly more effective compared to existing parametric schemes, especially with the growing amount of training data becoming available. 1
Investigating the Semantic Gap through Query Log Analysis.
Abstract. Significant efforts have focused in the past years on bringing large amounts of metadata online and the success of these efforts can be seen by the impressive number of web sites exposing data in RDFa or RDF/XML. However, little is known about the extent to which this data fits the needs of ordinary web users with everyday information needs. In this paper we study what we perceive as the semantic gap between the supply of data on the Semantic Web and the needs of web users as expressed in the queries submitted to a major Web search engine. We perform our analysis on both the level of instances and ontologies. First, we first look at how much data is actually relevant to Web queries and what kind of data is it. Second, we provide a generic method to extract the attributes that Web users are searching for regarding particular classes of entities. This method allows to contrast class definitions found in Semantic Web vocabularies with the attributes of objects that users are interested in. Our findings are crucial to measuring the potential of semantic search, but also speak to the state of the Semantic Web in general. 1
An evaluation of entity and frequency based query completion methods.

Web Search Relevance Ranking.
Abstract. This survey reviews the research related to PageRank computing. Components of a PageRank vector serve as authority weights for web pages independent of their textual content, solely based on the hyperlink structure of the web. PageRank is typically used as a web search ranking component. This defines the importance of the model and the data structures that underly PageRank processing. Computing even a single PageRank is a difficult computational task. Computing many PageRanks is a much more complex challenge. Recently, significant effort has been invested in building sets of personalized PageRank vectors. PageRank is also used in many diverse applications other than ranking. We are interested in the theoretical foundations of the PageRank formulation, in the acceleration of PageRank computing, in the effects of particular aspects of web graph structure on the optimal organization of computations, and in PageRank stability. We also review alternative models that lead to authority indices similar to PageRank and the role of such indices in applications other than web search. We also discuss linkbased search personalization and outline some aspects of PageRank infrastructure from associated measures of convergence to link preprocessing. 1.
Learning to Rank Answers on Large Online QA Collections.
This work describes an answer ranking engine for non-factoid questions built using a large online community-generated question-answer collection (Yahoo! Answers). We show how such collections may be used to effectively set up large supervised learning experiments. Furthermore we investigate a wide range of feature types, some exploiting NLP processors, and demonstrate that using them in combination leads to considerable improvements in accuracy. 1
The Probabilistic Relevance Framework BM25 and Beyond.

Learning to Rank Answers on Large Online QA Collections.
This work describes an answer ranking engine for non-factoid questions built using a large online community-generated question-answer collection (Yahoo! Answers). We show how such collections may be used to effectively set up large supervised learning experiments. Furthermore we investigate a wide range of feature types, some exploiting NLP processors, and demonstrate that using them in combination leads to considerable improvements in accuracy. 1
Query Clauses and Term Independence.

Exploiting Semantic Annotations in Information Retrieval.
There is an increasing amount of structure on the Web as a result of modern Web languages, user tagging and annotation, and emerging robust NLP tools. These meaningful, semantic, annotations hold the promise to significantly enhance information access, by enhancing the depth of analysis of today’s systems. Currently, we have only started exploring the possibilities and only begin to understand how these valuable semantic cues can be put to fruitful use. Unleashing the potential of semantic annotations requires us to think outside the box, by combining the insights of natural language processing (NLP) to go beyond bags of words, the insights of databases (DB) to use structure efficiently even when aggregating over millions of records, the insights of information retrieval (IR) in effective goal-directed search and evaluation, and the insights of knowledge management (KM) to get grips on the greater whole. The Workshop aims to bring together researchers from these different disciplines and work together on one of the greatest challenges in the years to come. The desired result of the workshop will be concrete insight into the potential of semantic annotations, and in concrete steps to take this research forward; synchronize related research happening in NLP, DB, IR, and KM, in ways that combine the strengths of each discipline; and have a lively, interactive workshop were everyone contributes and that inspires attendees to think “outside the box.”
Semantically Annotated Snapshot of the English Wikipedia.
This paper describes SW1, the first version of a semantically annotated snapshot of the English Wikipedia. In recent years Wikipedia has become a valuable resource for both the Natural Language Processing (NLP) community and the Information Retrieval (IR) community. Although NLP technology for processing Wikipedia already exists, not all researchers and developers have the computational resources to process such a volume of information. Moreover, the use of different versions of Wikipedia processed differently might make it difficult to compare results. The aim of this work is to provide easy access to syntactic and semantic annotations for researchers of both NLP and IR communities by building a reference corpus to homogenize experiments and make results comparable. These resources, a semantically annotated corpus and a “entity containment ” derived graph, are licensed under the GNU Free Documentation License and available from
Exploiting Morphological Query Structure Using Genetic Optimisation.

Towards Semantic Search.
Todays web search engines try to offer services for finding various information in addition to simple web pages, like showing locations or answering simple fact queries. Understanding the association of named entities and documents is one of the key steps towards such semantic search tasks. This paper addresses the ranking of entities and models it in a graph-based relevance propagation framework. In particular we study the problem of expert finding as an example of an entity ranking task. Entity containment graphs are introduced that represent the relationship between text fragments on the one hand and their contained entities on the other hand. The paper shows how these graphs can be used to propagate relevance information from the pre-ranked text fragments to their entities. We use this propagation framework to model existing approaches to expert finding based on the entity’s indegree and extend them by recursive relevance propagation based on a probabilistic random walk over the entity containment graphs. Experiments on the TREC expert search task compare the retrieval performance of the different graph and propagation models. 1.
Inferring the most important types of a query a semantic approach.

Near%2DTerm Prospects for Semantic Technologies.

Learning to Tag and Tagging to Learn A Case Study on Wikipedia.

Learning to Tag and Tagging to Learn A Case Study on Wikipedia.

Exploiting semantic annotations in information retrieval ESAIR %2708.

Ranking very many typed entities on wikipedia.

Predictive user click models based on click%2Dthrough history.

Hits on the web how does it compare%3F

On rank%2Dbased effectiveness measures and optimization.
ranking functions; optimization Many current retrieval models and scoring functions contain free parameters which need to be set – ideally, optimized. The process of optimization normally involves some training corpus of the usual document-query-relevance judgement type, and some choice of measure that is to be optimized. The paper proposes a way to think about the process of exploring the space of parameter values, and how moving around in this space might be expected to affect different measures. One result, concerning local optima, is demonstrated for a range of rank-based evaluation measures. 1 Parameters and optimization This paper addresses some basic features of many of the measures of retrieval effectiveness in current or past use. The objective is to understand aspects of their behaviour, with particular reference to the optimization of parameters of a retrieval model or ranking function. This is a theoretical paper.
Optimisation methods for ranking functions with multiple parameters.
parameters
The Role of Machine Learning in Information Retrieval Research.

Learning What to Talk about in Descriptive Games.
Text generation requires a planning module to select an object of discourse and its properties. This is specially hard in descriptive games, where a computer agent tries to describe some aspects of a game world. We propose to formalize this problem as a Markov Decision Process, in which an optimal message policy can be defined and learned through simulation. Furthermore, we propose back-off policies as a novel and effective technique to fight state dimensionality explosion in this framework. 1
Microsoft Cambridge at TREC 14 Enterprise Track.
mbination and 3) The tuning and ranking framework we used this year.  To calculate BM25F [4] we first calculate a normalised term frequency for each field:   x d,f,t :=  x d,f,t  (1 +B f (  l d,f  l f  -  1))    # {SUBJECT,  BODY, QUOTED} indicates the field type, x d,f,t is the term frequency of term t in the field type f of document d, l d,f is the length of that field, and l f is the average field length for that field type. B f is a field-dependant parameter similar to the B parameter in BM25. In particular, if B f = 0 there is no normalisation and if B f = 1 the frequency is completely normalised w.r.t. the average field length.  These term frequencies can then be combined in a linearly weighted sum to obtain the final term pseudo-frequency:   x d,t =    W f     x d,f,t (2) with weight parameters W f . This is then used in the usual BM25 saturating function. This leads the following ranking function, which we refer to as BM25F:  BM25F (d) :=    t#q#d   x d,t  K 1 +  x d,t    t (3)
Relevance weighting for query independent evidence.

Microsoft Cambridge at TREC 14 Enterprise Track.
mbination and 3) The tuning and ranking framework we used this year.  To calculate BM25F [4] we first calculate a normalised term frequency for each field:   x d,f,t :=  x d,f,t  (1 +B f (  l d,f  l f  -  1))    # {SUBJECT,  BODY, QUOTED} indicates the field type, x d,f,t is the term frequency of term t in the field type f of document d, l d,f is the length of that field, and l f is the average field length for that field type. B f is a field-dependant parameter similar to the B parameter in BM25. In particular, if B f = 0 there is no normalisation and if B f = 1 the frequency is completely normalised w.r.t. the average field length.  These term frequencies can then be combined in a linearly weighted sum to obtain the final term pseudo-frequency:   x d,t =    W f     x d,f,t (2) with weight parameters W f . This is then used in the usual BM25 saturating function. This leads the following ranking function, which we refer to as BM25F:  BM25F (d) :=    t#q#d   x d,t  K 1 +  x d,t    t (3)
Simple BM25 extension to multiple weighted fields.
This paper describes a simple way of adapting the BM25 ranking formula to deal with structured documents. In the past it has been common to compute scores for the individual fields (e.g. title and body) independently and then combine these scores (typically linearly) to arrive at a final score for the document. We highlight how this approach can lead to poor performance by breaking the carefully constructed non-linear saturation of term frequency in the BM25 function. We propose a much more intuitive alternative which weights term frequencies before the nonlinear term frequency saturation function is applied. In this scheme, a structured document with a title weight of two is mapped to an unstructured document with the title content repeated twice. This more verbose unstructured document is then ranked in the usual way. We demonstrate the advantages of this method with experiments on Reuters Vol1 and the TREC dotGov collection.
Parsimonious language models for information retrieval.
We systematically investigate a new approach to estimating the parameters of language models for information retrieval, called parsimonious language models. Parsimonious language models explicitly address the relation between levels of language models that are typically used for smoothing. As such, they need fewer (non-zero) parameters to describe the data. We apply parsimonious models at three stages of the retrieval process: 1) at indexing time; 2) at search time; 3) at feedback time. Experimental results show that we are able to build models that are significantly smaller than standard models, but that still perform at least as well as the standard approaches.
Microsoft Cambridge at TREC 13 Web and Hard Tracks.
 
Bayesian extension to the language model for ad hoc information retrieval.
We propose a Bayesian extension to the ad-hoc Language Model. Many smoothed estimators used for the multinomial query model in ad-hoc Language Models (including Laplace and Bayes-smoothing) are approximations to the Bayesian predictive distribution. In this paper we derive the full predictive distribution in a form amenable to implementation by classical IR models, and then compare it to other currently used estimators. In our experiments the proposed model outperforms Bayes-smoothing, and its combination with linear interpolation smoothing outperforms all other estimators.
Microsoft Cambridge at TREC%2D12 HARD track.
 
The Perceptron Algorithm with Uneven Margins.
 
Microsoft Cambridge at TREC 2002 Filtering Track.
this paper. For these comparisons we consider only the msPUMb model. We note that these results are better than those submitted originally: after submission we discovered an error in our data normalisation procedure; after correcting it the performance of all models was increased
Information Retrieval Algorithms and Heuristics.

Learning for Sequence Extraction Tasks.
We consider the application of machine learning techniques for sequence modeling to Information Retrieval (IR) and surface Information Extraction (IE) tasks. We introduce a generic sequence model and show how it can be used for dealing with different tasks. Taking into account the sequential nature of texts allows for a finer analysis than what is usually done in IR with static text representations. The task we are focusing on is the retrieval and labeling of texts passages, also known as highlighting and surface information extraction. We describe different implementations of our model based on Hidden Markov Models and Neural Networks. Experiments are performed using the MUC6 corpus from the information extraction community.
Microsoft Cambridge at TREC%2D10 Filtering and Web Tracks.
This report is concerned with the Adaptive Filtering and Web tracks. There are separate reports in this volume [1, 2] on the Microsoft Research Redmond participation in QA track and the Microsoft Research Beijing participation in the Web track. Two runs
Learning for Sequence Extraction Tasks.
We consider the application of machine learning techniques for sequence modeling to Information Retrieval (IR) and surface Information Extraction (IE) tasks. We introduce a generic sequence model and show how it can be used for dealing with different tasks. Taking into account the sequential nature of texts allows for a finer analysis than what is usually done in IR with static text representations. The task we are focusing on is the retrieval and labeling of texts passages, also known as highlighting and surface information extraction. We describe different implementations of our model based on Hidden Markov Models and Neural Networks. Experiments are performed using the MUC6 corpus from the information extraction community.
Sequence Models for Automatic Highlighting and Surface Information Extraction.
With the increase of textual information available electronically, we assist to a great diversification of the demands on Information Retrieval (IR) and Information Extraction (IE) systems. In this paper we apply Machine Learning techniques of sequence analysis to the tasks of highlighting and labeling text with respect to an information extraction task. Specifically, dynamic probability models are used. Like IR systems, they use little semantics, are fully trainable and do not require any knowledge representation of the domain. Unlike IR approaches, documents are considered as a dynamic sequence of words. Furthermore, additional word information is naturally included in the representation. Models are evaluated on a sub-task of the MUC6 Scenario Template corpus. When morpho-syntactic word information is introduced into the representation, an increase in performances is observed. 1
Coupled Hierarchical IR and Stochastic Models for Surface Information Extraction.
We present in this paper a combination of Machine Learning based Information Retrieval (IR) techniques and stochastic  language modelling in a hierarchical system that extracts surface information from text. At the lowest level  of this hierarchy, documents and paragraphs are successively routed with IR techniques. At the top level, a stochastic  language model extracts the most relevant phrases, and labels the type of information they contain. The approach  and preliminary results are demonstrated on a subset of the MUC-6 Scenario Templates task.
Multiple Multivariate Regression and Global Optimization in a Large Scale Thermodynamical Application.
We describe a large scale real-world application of neural networks for the  modelization of heat radiation emitted by a source and observed through the atmosphere.
