Abstract In this paper we investigate unsupervised name transliteration using comparable corpora,... more Abstract In this paper we investigate unsupervised name transliteration using comparable corpora, corpora where texts in the two languages deal in some of the same topics---and therefore share references to named entities---but are not translations of each other. We present two distinct methods for transliteration, one approach using an unsupervised phonetic transliteration method, and the other using the temporal distribution of candidate pairs.
Background Conserved protein sequence motifs are short stretches of amino acid sequence patterns ... more Background Conserved protein sequence motifs are short stretches of amino acid sequence patterns that potentially encode the function of proteins. Several sequence pattern searching algorithms and programs exist foridentifying candidate protein motifs at the whole genome level. However, amuch needed and importanttask is to determine the functions of the newly identified protein motifs.
Abstract In this work, we study a new text mining problem of discovering named entities with temp... more Abstract In this work, we study a new text mining problem of discovering named entities with temporally correlated bursts of mention counts in multiple multilingual Web news streams. Mining named entities with temporally correlated bursts of mention counts in multilingual text streams has many interesting and important applications, such as identification of the latent events, attracting the attention of on-line media in different countries, and valuable linguistic knowledge in the form of transliterations.
A common task in many applications is to find persons who are knowledgeable about a given topic (... more A common task in many applications is to find persons who are knowledgeable about a given topic (ie, expert finding). In this paper, we propose and develop a general probabilistic framework for studying expert finding problem and derive two families of generative models (candidate generation models and topic generation models) from the framework. These models subsume most existing language models proposed for expert finding.
Abstract Given the experimental nature of information retrieval, progress critically depends on a... more Abstract Given the experimental nature of information retrieval, progress critically depends on analyzing the errors made by existing retrieval approaches and understanding their limitations. Our research explores various hypothesized reasons for hard topics in TREC-8 ad hoc task, and shows that the bad performance is partially due to the existence of highly distracting sub-collections that can dominate the overall performance.
Abstract: In this thesis, we study the importance of background knowledge in relation extraction ... more Abstract: In this thesis, we study the importance of background knowledge in relation extraction systems. We not only demonstrate the benefits of leveraging background knowledge to improve the systems' performance but also propose a principled framework that allows one to effectively incorporate knowledge into statistical machine learning models for relation extraction.
Abstract A common task in many text mining applications is to generate a multi-faceted overview o... more Abstract A common task in many text mining applications is to generate a multi-faceted overview of a topic in a text collection. Such an overview not only directly serves as an informative summary of the topic, but also provides a detailed view of navigation to different facets of the topic. Existing work has cast this problem as a categorization problem and requires training examples for each facet.
ABSTRACT The exponential growth of data demands scalable and adaptable infrastructures for indexi... more ABSTRACT The exponential growth of data demands scalable and adaptable infrastructures for indexing and searching a huge amount of data sources with high accuracy and efficiency. Existing centralized search engines are not scalable and suffer from single-point-offailures. The recent work on P2P index construction partitions the document vectors either randomly or statically, making it difficult to tradeoff between search efficiency and accuracy.
Abstract This paper proposes a new model for information retrieval through the use of an author's... more Abstract This paper proposes a new model for information retrieval through the use of an author's pass information. This model assumes that if an author writes more about a particular topic, then the document that he/she have written should be more insightful than the other similarly relevant documents. Therefore, those documents written by experienced authors should be ranked higher. This paper describes an application of this model to retrieving messages from a mailing list or a newsgroup environment.
Abstract A user's web search history contains many valuable search patterns. In this paper, we st... more Abstract A user's web search history contains many valuable search patterns. In this paper, we study search patterns that represent a user's long-lasting and exploratory search interests. By focusing on long-lastingness and exploratoriness, we are able to discover search patterns that are most useful for recommending new and relevant information to the user. Our approach is based on language modeling and clustering, and specifically designed to handle web search logs.
Abstract Query spelling correction is a crucial component of modern search engines. Existing meth... more Abstract Query spelling correction is a crucial component of modern search engines. Existing methods in the literature for search query spelling correction have two major drawbacks. First, they are unable to handle certain important types of spelling errors, such as concatenation and splitting.
Abstract Collaborative filtering is a very useful general technique for exploiting the preference... more Abstract Collaborative filtering is a very useful general technique for exploiting the preference patterns of a group of users to predict the utility of items to a particular user. Previous research has studied several probabilistic graphic models for collaborative filtering with promising results.
Abstract UIUC participated in the HARD track in TREC 2004 and focused on the evaluation of a new ... more Abstract UIUC participated in the HARD track in TREC 2004 and focused on the evaluation of a new method for identifying variable-length passages using HMMs. Most existing approaches to passage retrieval rely on pre-segmentation of documents, but the optimal boundaries of a relevant passage depends on both the query and the document. Our new method aims at determining or improving the boundaries of a relevant passage based on both the query and topical coherence in the document.
Current search engines heavily emphasize on direct querying which tends to work well only for sim... more Current search engines heavily emphasize on direct querying which tends to work well only for simple information needs such as navigational queries. However, direct querying may not support complex information needs such as exploratory search well [4, 11] since users' interactions are mainly limited to submitting a query, viewing results, and reformulating queries [1]. As a complementary way of information seeking with querying, browsing can be very useful for exploratory search or information foraging [5].
Abstract As the amount of textual information grows explosively in various kinds of business syst... more Abstract As the amount of textual information grows explosively in various kinds of business systems, it becomes more and more desirable to analyze both structured data records and unstructured text data simultaneously. Although online analytical processing (OLAP) techniques have been proven very useful for analyzing and mining structured data, they face challenges in handling text data. On the other hand, probabilistic topic models are among the most effective approaches to latent topic analysis and mining on text data.
Abstract As more and more content is published and consumed online, it is imperative to know if a... more Abstract As more and more content is published and consumed online, it is imperative to know if an information nugget found on the Web is trustworthy or not. This is especially important for online medical information as it affects the most vulnerable group of users looking for medical help online.
Abstract This thesis presents a new general probabilistic framework for text retrieval based on B... more Abstract This thesis presents a new general probabilistic framework for text retrieval based on Bayesian decision theory. In this framework, queries and documents are modeled using statistical language models, user preferences are modeled through loss functions, and retrieval is cast as a risk minimization problem.
ABSTRACT In this paper, an online news recommender system for the popular social network, Faceboo... more ABSTRACT In this paper, an online news recommender system for the popular social network, Facebook, is described. This system provides daily newsletters for communities on Facebook. The system fetches the news articles and filters them based on the community description to prepare the daily news digest. Explicit survey feedback from the users show that most users found the application useful and easy to use. They also indicated that they could get some community specific articles that they would not have got otherwise.
Link-based ranking algorithms are among the most important techniques to improve web search. In p... more Link-based ranking algorithms are among the most important techniques to improve web search. In particular, the PageRank algorithm has been successfully used in the Google search engine and has been attracting much attention recently. However, we find that PageRank has a ���zeroone gap��� problem which, to the best of our knowledge, has not been addressed in any previous work.
Abstract In this paper we investigate unsupervised name transliteration using comparable corpora,... more Abstract In this paper we investigate unsupervised name transliteration using comparable corpora, corpora where texts in the two languages deal in some of the same topics---and therefore share references to named entities---but are not translations of each other. We present two distinct methods for transliteration, one approach using an unsupervised phonetic transliteration method, and the other using the temporal distribution of candidate pairs.
Background Conserved protein sequence motifs are short stretches of amino acid sequence patterns ... more Background Conserved protein sequence motifs are short stretches of amino acid sequence patterns that potentially encode the function of proteins. Several sequence pattern searching algorithms and programs exist foridentifying candidate protein motifs at the whole genome level. However, amuch needed and importanttask is to determine the functions of the newly identified protein motifs.
Abstract In this work, we study a new text mining problem of discovering named entities with temp... more Abstract In this work, we study a new text mining problem of discovering named entities with temporally correlated bursts of mention counts in multiple multilingual Web news streams. Mining named entities with temporally correlated bursts of mention counts in multilingual text streams has many interesting and important applications, such as identification of the latent events, attracting the attention of on-line media in different countries, and valuable linguistic knowledge in the form of transliterations.
A common task in many applications is to find persons who are knowledgeable about a given topic (... more A common task in many applications is to find persons who are knowledgeable about a given topic (ie, expert finding). In this paper, we propose and develop a general probabilistic framework for studying expert finding problem and derive two families of generative models (candidate generation models and topic generation models) from the framework. These models subsume most existing language models proposed for expert finding.
Abstract Given the experimental nature of information retrieval, progress critically depends on a... more Abstract Given the experimental nature of information retrieval, progress critically depends on analyzing the errors made by existing retrieval approaches and understanding their limitations. Our research explores various hypothesized reasons for hard topics in TREC-8 ad hoc task, and shows that the bad performance is partially due to the existence of highly distracting sub-collections that can dominate the overall performance.
Abstract: In this thesis, we study the importance of background knowledge in relation extraction ... more Abstract: In this thesis, we study the importance of background knowledge in relation extraction systems. We not only demonstrate the benefits of leveraging background knowledge to improve the systems' performance but also propose a principled framework that allows one to effectively incorporate knowledge into statistical machine learning models for relation extraction.
Abstract A common task in many text mining applications is to generate a multi-faceted overview o... more Abstract A common task in many text mining applications is to generate a multi-faceted overview of a topic in a text collection. Such an overview not only directly serves as an informative summary of the topic, but also provides a detailed view of navigation to different facets of the topic. Existing work has cast this problem as a categorization problem and requires training examples for each facet.
ABSTRACT The exponential growth of data demands scalable and adaptable infrastructures for indexi... more ABSTRACT The exponential growth of data demands scalable and adaptable infrastructures for indexing and searching a huge amount of data sources with high accuracy and efficiency. Existing centralized search engines are not scalable and suffer from single-point-offailures. The recent work on P2P index construction partitions the document vectors either randomly or statically, making it difficult to tradeoff between search efficiency and accuracy.
Abstract This paper proposes a new model for information retrieval through the use of an author's... more Abstract This paper proposes a new model for information retrieval through the use of an author's pass information. This model assumes that if an author writes more about a particular topic, then the document that he/she have written should be more insightful than the other similarly relevant documents. Therefore, those documents written by experienced authors should be ranked higher. This paper describes an application of this model to retrieving messages from a mailing list or a newsgroup environment.
Abstract A user's web search history contains many valuable search patterns. In this paper, we st... more Abstract A user's web search history contains many valuable search patterns. In this paper, we study search patterns that represent a user's long-lasting and exploratory search interests. By focusing on long-lastingness and exploratoriness, we are able to discover search patterns that are most useful for recommending new and relevant information to the user. Our approach is based on language modeling and clustering, and specifically designed to handle web search logs.
Abstract Query spelling correction is a crucial component of modern search engines. Existing meth... more Abstract Query spelling correction is a crucial component of modern search engines. Existing methods in the literature for search query spelling correction have two major drawbacks. First, they are unable to handle certain important types of spelling errors, such as concatenation and splitting.
Abstract Collaborative filtering is a very useful general technique for exploiting the preference... more Abstract Collaborative filtering is a very useful general technique for exploiting the preference patterns of a group of users to predict the utility of items to a particular user. Previous research has studied several probabilistic graphic models for collaborative filtering with promising results.
Abstract UIUC participated in the HARD track in TREC 2004 and focused on the evaluation of a new ... more Abstract UIUC participated in the HARD track in TREC 2004 and focused on the evaluation of a new method for identifying variable-length passages using HMMs. Most existing approaches to passage retrieval rely on pre-segmentation of documents, but the optimal boundaries of a relevant passage depends on both the query and the document. Our new method aims at determining or improving the boundaries of a relevant passage based on both the query and topical coherence in the document.
Current search engines heavily emphasize on direct querying which tends to work well only for sim... more Current search engines heavily emphasize on direct querying which tends to work well only for simple information needs such as navigational queries. However, direct querying may not support complex information needs such as exploratory search well [4, 11] since users' interactions are mainly limited to submitting a query, viewing results, and reformulating queries [1]. As a complementary way of information seeking with querying, browsing can be very useful for exploratory search or information foraging [5].
Abstract As the amount of textual information grows explosively in various kinds of business syst... more Abstract As the amount of textual information grows explosively in various kinds of business systems, it becomes more and more desirable to analyze both structured data records and unstructured text data simultaneously. Although online analytical processing (OLAP) techniques have been proven very useful for analyzing and mining structured data, they face challenges in handling text data. On the other hand, probabilistic topic models are among the most effective approaches to latent topic analysis and mining on text data.
Abstract As more and more content is published and consumed online, it is imperative to know if a... more Abstract As more and more content is published and consumed online, it is imperative to know if an information nugget found on the Web is trustworthy or not. This is especially important for online medical information as it affects the most vulnerable group of users looking for medical help online.
Abstract This thesis presents a new general probabilistic framework for text retrieval based on B... more Abstract This thesis presents a new general probabilistic framework for text retrieval based on Bayesian decision theory. In this framework, queries and documents are modeled using statistical language models, user preferences are modeled through loss functions, and retrieval is cast as a risk minimization problem.
ABSTRACT In this paper, an online news recommender system for the popular social network, Faceboo... more ABSTRACT In this paper, an online news recommender system for the popular social network, Facebook, is described. This system provides daily newsletters for communities on Facebook. The system fetches the news articles and filters them based on the community description to prepare the daily news digest. Explicit survey feedback from the users show that most users found the application useful and easy to use. They also indicated that they could get some community specific articles that they would not have got otherwise.
Link-based ranking algorithms are among the most important techniques to improve web search. In p... more Link-based ranking algorithms are among the most important techniques to improve web search. In particular, the PageRank algorithm has been successfully used in the Google search engine and has been attracting much attention recently. However, we find that PageRank has a ���zeroone gap��� problem which, to the best of our knowledge, has not been addressed in any previous work.
Uploads