Visual Contextual Advertising Bringing Textual Advertisements to Images.
Advertising in the case of textual Web pages has been studied extensively by many researchers. However, with the increasing amount of multimedia data such as image, audio and video on the Web, the need for recommending advertisement for the multimedia data is becoming a reality. In this paper, we address the novel problem of visual contextual advertising, which is to directly advertise when users are viewing images which do not have any surrounding text. A key challenging issue of visual contextual advertising is that images and advertisements are usually represented in image space and word space respectively, which are quite different with each other inherently. As a result, existing methods for Web page advertising are inapplicable since they represent both Web pages and advertisement in the same word space. In order to solve the problem, we propose to exploit the social Web to link these two feature spaces together. In particular, we present a unified generative model to integrate advertisements, words and images. Specifically, our solution combines two parts in a principled approach: First, we transform images from a image feature space to a word space utilizing the knowledge from images with annotations from social Web. Then, a language model based approach is applied to estimate the relevance between transformed images and advertisements. Moreover, in this model, the probability of recommending an advertisement can be inferred efficiently given an image, which enables potential applications to online advertising. 1.
Click Prediction for Product Search on C2C Web Sites.

Predicting Product Duration for Adaptive Advertisement.

Text%2DAided Image Classification Using Labeled Text from Web to Help Image Classification.

Text%2DAided Image Classification Using Labeled Text from Web to Help Image Classification.

Transfer learning for behavioral targeting.

Learning with Positive and Unlabeled Examples Using Topic%2DSensitive PLSA.

Knowledge transfer for cross domain learning to rank.

Multi%2Dtask learning for learning to rank in web search.

Proceedings of the 2nd ACM Workshop on Social Web Search and Mining

Multi%2Dtask learning for learning to rank in web search.

EigenTransfer a unified framework for transfer learning.
This paper proposes a general framework, called EigenTransfer, to tackle a variety of transfer learning problems, e.g. cross-domain learning, self-taught learning, etc. Our basic idea is to construct a graph to represent the target transfer learning task. By learning the spectra of a graph which represents a learning task, we obtain a set of eigenvectors that reflect the intrinsic structure of the task graph. These eigenvectors can be used as the new features which transfer the knowledge from auxiliary data to help classify target data. Given an arbitrary non-transfer learner (e.g. SVM) and a particular transfer learning task, EigenTransfer can produce a transfer learner accordingly for the target transfer learning task. We apply EigenTransfer on three different transfer learning tasks, cross-domain learning, cross-category learning and self-taught learning, to demonstrate its unifying ability, and show through experiments that EigenTransfer can greatly outperform several representative non-transfer learners. 1.
Global ranking by exploiting user clicks.
It is now widely recognized that user interactions with search results can provide substantial relevance information on the documents displayed in the search results. In this paper, we focus on extracting relevance information from one source of user interactions, i.e., user click data, which records the sequence of documents being clicked and not clicked in the result set during a user search session. We formulate the problem as a global ranking problem, emphasizing the importance of the sequential nature of user clicks, with the goal to predict the relevance labels of all the documents in a search session. This is distinct from conventional learning to rank methods that usually design a ranking model defined on a single document; in contrast, in our model the relational information among the documents as manifested by an aggregation of user clicks is exploited to rank all the documents jointly. In particular, we adapt several sequential supervised learning algorithms, including the conditional random field (CRF), the sliding window method and the recurrent sliding window method, to the global ranking problem. Experiments on the click data collected from a commercial search engine demonstrate that our methods can outperform the baseline models for search results re-ranking.
Efficient query expansion for advertisement search.

Dataplorer a scalable search engine for the data web.

Web%2Dscale classification with naive bayes.
Traditional Naive Bayes Classifier performs miserably on web-scale taxonomies. In this paper, we investigate the reasons behind such bad performance. We discover that the low performance are not completely caused by the intrinsic limitations of Naive Bayes, but mainly comes from two largely ignored problems: contradiction pair problem and discriminative evidence cancelation problem. We propose modifications that can alleviate the two problems while preserving the advantages of Naive Bayes. The experimental results show our modified Naive Bayes can significantly improve the performance on real web-scale taxonomies.
Enhancing diversity
This report discusses the practices which have been used or recommended for increasing the degree of diversity between redundant implementations of software or software-based systems. Its purpose is to give useful indications for designers, project managers and safety/reliability assessors in deciding about how great an advantage should be expected from the use of these practices, in absolute and in comparative terms. Existing knowledge does not allow one to state any strong general recommendations, but it is possible to improve on the intuitive justifications usually given for these various practices. This report clarifies the ways the various practices are conjectured to aid system reliability, the factors that should affect their efficacy, and thus, for a practitioner, the aspects of a specific project situation that need to be considered to inform decisions. Thus this report
User language model for collaborative personalized search.

Iterative Reinforcement Cross%2DDomain Text Classification.

Knowledge Transferring Via Implicit Link Analysis.

Self%2Dtaught clustering.
This paper focuses on a new clustering task, called self-taught clustering. Self-taught clustering is an instance of unsupervised transfer learning, which aims at clustering a small collection of target unlabeled data with the help of a large amount of auxiliary unlabeled data. The target and auxiliary data can be different in topic distribution. We show that even when the target data are not sufficient to allow effective learning of a high quality feature representation, it is possible to learn the useful features with the help of the auxiliary data on which the target data can be clustered effectively. We propose a co-clustering based self-taught clustering algorithm to tackle this problem, by clustering the target and auxiliary data simultaneously to allow the feature representation from the auxiliary data to influence the target data through a common set of features. Under the new data representation, clustering on the target data can be improved. Our experiments on image clustering show that our algorithm can greatly outperform several state-of-the-art clustering methods when utilizing irrelevant unlabeled auxiliary data. 1.
Spectral domain%2Dtransfer learning.
Traditional spectral classification has been proved to be effective in dealing with both labeled and unlabeled data when these data are from the same domain. In many real world applications, however, we wish to make use of the labeled data from one domain (called in-domain) to classify the unlabeled data in a different domain (out-of-domain). This problem often happens when obtaining labeled data in one domain is difficult while there are plenty of labeled data from a related but different domain. In general, this is a transfer learning problem where we wish to classify the unlabeled data through the labeled data even though these data are not from the same domain. In this paper, we formulate this domain-transfer learning problem under a novel spectral classification framework, where the objective function is introduced to seek consistency between the in-domain supervision and the out-of-domain intrinsic structure. Through optimization of the cost function, the label information from the in-domain data is effectively transferred to help classify the unlabeled data from the out-of-domain. We conduct extensive experiments to evaluate our method and show that our algorithm achieves significant improvements on classification performance over many state-of-the-art algorithms.
Translated Learning Transfer Learning across Different Feature Spaces.
This paper investigates a new machine learning strategy called translated learning. Unlike many previous learning tasks, we focus on how to use labeled data from one feature space to enhance the classification of other entirely different learning spaces. For example, we might wish to use labeled text data to help learn a model for classifying image data, when the labeled images are difficult to obtain. An important aspect of translated learning is to build a “bridge ” to link one feature space (known as the “source space”) to another space (known as the “target space”) through a translator in order to migrate the knowledge from source to target. The translated learning solution uses a language model to link the class labels to the features in the source spaces, which in turn is translated to the features in the target spaces. Finally, this chain of linkages is completed by tracing back to the instances in the target spaces. We show that this path of linkage can be modeled using a Markov chain and risk minimization. Through experiments on the text-aided image classification and cross-language classification tasks, we demonstrate that our translated learning framework can greatly outperform many state-of-the-art baseline methods. 1
Knowledge Supervised Text Classification with No Labeled Documents.

Learning to rank with ties.

Deep classification in large%2Dscale text hierarchies.
Most classification algorithms are best at categorizing the Web documents into a few categories, such as the top two levels in the Open Directory Project. Such a classification method does not give very detailed topic-related class information for the user because the first two levels are often too coarse. However, classification on a large-scale hierarchy is known to be intractable for many target categories with cross-link relationships among them. In this paper, we propose a novel deep-classification approach to categorize Web documents into categories in a large-scale taxonomy. The approach consists of two stages: a search stage and a classification stage. In the first stage, a category-search algorithm is used to acquire the
Topic%2Dbridged PLSA for cross%2Ddomain text classification.
In many Web applications, such as blog classification and newsgroup classification, labeled data are in short supply. It often happens that obtaining labeled data in a new domain is expensive and time consuming, while there may be plenty of labeled data in a related but different domain. Traditional text classification approaches are not able to cope well with learning across different domains. In this paper, we propose a novel cross-domain text classification algorithm which extends the traditional probabilistic latent semantic analysis (PLSA) algorithm to integrate labeled and unlabeled data, which come from different but related domains, into a unified probabilistic model. We call this new model Topic-bridged PLSA, or TPLSA. By exploiting the common topics between two domains, we transfer knowledge across different domains through a topic-bridge to help the text classification in the target domain. A unique advantage of our method is its ability to maximally mine knowledge that can be transferred between domains, resulting in superior performance when compared to other state-of-the-art text classification approaches. Experimental evaluation on different kinds of datasets shows that our proposed algorithm can improve the performance of cross-domain text classification significantly.

Deep classifier automatically categorizing search results into large%2Dscale hierarchies.
Organizing Web search results into hierarchical categories facilitates users ’ browsing through Web search results, especially for ambiguous queries where the potential results are mixed together. Previous methods on search result classification are usually based on pre-training a classification model on some fixed and shallow hierarchical categories, where only the top-two-level categories of a Web taxonomy is used. Such classification methods may be too coarse for users to browse, since most search results would be classified into only two or three shallow categories. Instead, a deep hierarchical classifier must provide many more categories. However, the performance of such classifiers is usually limited because their classification effectiveness can deteriorate rapidly at the third or fourth level of a hierarchy. In this paper, we propose a novel algorithm known as Deep Classifier to classify the search results into detailed hierarchical categories with higher effectiveness than previous approaches. Given the search results in response to a query, the algorithm first prunes a wide-ranged hierarchy into a narrow one with the help of some Web directories. Different strategies are proposed to select the training data by utilizing the hierarchical structures. Finally, a discriminative naïve Bayesian classifier is developed to perform efficient and effective classification. As a result, the algorithm can provide more meaningful and specific class labels for search result browsing than shallow style of classification. We conduct experiments to show that the Deep Classifier can achieve significant improvement over state-of-the-art algorithms. In addition, with sufficient off-line preparation, the efficiency of the proposed algorithm is suitable for online application.

Can chinese web pages be classified with english data source%3F
As the World Wide Web in China grows rapidly, mining knowledge in Chinese Web pages becomes more and more important. Mining Web information usually relies on the machine learning techniques which require a large amount of labeled data to train credible models. Although the number of Chinese Web pages increases quite fast, it still lacks Chinese labeled data. However, there are relatively sufficient English labeled Web pages. These labeled data, though in different linguistic representations, share a substantial amount of semantic information with Chinese ones, and can be utilized to help classify Chinese Web pages. In this paper, we propose an information bottleneck based approach to address this cross-language classification problem. Our algorithm first translates all the Chinese Web pages to English. Then, all the Web pages, including Chinese and English ones, are encoded through an information bottleneck which can allow only limited information to pass. Therefore, in order to retain as much useful information as possible, the common part between Chinese and English Web pages is inclined to be encoded to the same code (i.e. class label), which makes the cross-language classification accurate. We evaluated our approach using the Web pages collected from Open Directory Project (ODP). The experimental results show that our method significantly improves several existing supervised and semi-supervised classifiers.
Semantic Services for Wikipedia.

Transferring Naive Bayes Classifiers for Text Classification.
A basic assumption in traditional machine learning is that the training and test data distributions should be identical. This assumption may not hold in many situations in practice, but we may be forced to rely on a different-distribution data to learn a prediction model. For example, this may be the case when it is expensive to label the data in a domain of interest, although in a related but different domain there may be plenty of labeled data available. In this paper, we propose a novel transfer-learning algorithm for text classification based on an EM-based Naive Bayes classifiers. Our solution is to first estimate the initial probabilities under a distribution Dℓ of one labeled data set, and then use an EM algorithm to revise the model for a different distribution Du of the test data which are unlabeled. We show that our algorithm is very effective in several different pairs of domains, where the distances between the different distributions are measured using the Kullback-Leibler (KL) divergence. Moreover, KL-divergence is used to decide the trade-off parameters in our algorithm. In the experiment, our algorithm outperforms the traditional supervised and semi-supervised learning algorithms when the distributions of the training and test sets are increasingly different.
Boosting for transfer learning.
Traditional machine learning makes a basic assumption: the training and test data should be under the same distribution. However, in many cases, this identicaldistribution assumption does not hold. The assumption might be violated when a task from one new domain comes, while there are only labeled data from a similar old domain. Labeling the new data can be costly and it would also be a waste to throw away all the old data. In this paper, we present a novel transfer learning framework called TrAdaBoost, which extends boosting-based learning algorithms (Freund &amp; Schapire, 1997). TrAdaBoost allows users to utilize a small amount of newly labeled data to leverage the old data to construct a high-quality classification model for the new data. We show that this method can allow us to learn an accurate model using only a tiny amount of new data and a large amount of old data, even when the new data are not sufficient to train a model alone. We show that TrAdaBoost allows knowledge to be effectively transferred from the old data to the new. The effectiveness of our algorithm is analyzed theoretically and empirically to show that our iterative algorithm can converge well to an accurate model.
Co%2Dclustering based classification for out%2Dof%2Ddomain documents.

Bridged Refinement for Transfer Learning.

Adaptive Email Spam Filtering Based on Information Theory.

Exploring in the weblog space by detecting informative and affective articles.

Exploring in the weblog space by detecting informative and affective articles.

Optimizing web search using social annotations.
Folksonomies have shown interesting potential for improving information discovery and exploration. Recent folksonomy systems investigate on the use of tag assignments, which combine Web resources with annotations (tags), and the users that have created the annotations. In this paper, we investigate on the effect of grouping resources, i.e. creating sets of resources, and using this additional structure for search in folksonomies. Our experiments show that the quality of search result ranking can be significantly improved by introducing and exploiting the grouping of resources (onetailed t-Test, level of significance α = 0.05).
A Hierarchical Model of Web Graph.

Image Description Mining and Hierarchical Clustering on Data Records Using HR%2DTree.

IRFCF Iterative Rating Filling Collaborative Filtering Algorithm.

An Empirical Study of Data Smoothing Methods for Memory%2DBased and Hybrid Collaborative Filtering.

A Novel Web Page Categorization Algorithm Based on Block Propagation Using Query%2DLog Information.

Exploiting Rating Behaviors for Effective Collaborative Filtering.

Reinforcing Web%2Dobject Categorization Through Interrelationships.
Abstract. Existing categorization algorithms deal with homogeneous Web objects, and consider interrelated objects as additional features when taking the interrelationships with other types of objects into account. However, focusing on any single aspect of the inter-object relationship is not sufficient to fully reveal the true categories of Web objects. In this paper, we propose a novel categorization algorithm, called the Iterative Reinforcement Categorization Algorithm (IRC), to exploit the full interrelationship between different types of Web objects on the Web, including Web pages and queries. IRC classifies the interrelated Web objects by iteratively reinforcing the individual classification results of different types of objects via their interrelationship. Experiments on a clickthrough-log dataset from the MSN search engine show that, in terms of the F1 measure, IRC achieves a 26.4 % improvement over a pure content-based classification method. It also achieves a 21% improvement over a query-metadata-based method, as well as a 16.4 % improvement on F1 measure over the well-known virtual document-based method. Our experiments show that IRC converges fast enough to be applicable to real world applications.
TSSP Multi%2Dfeatures based reinforcement algorithm to find related papers.

Block%2DBased Language Modeling Approach Towards Web Search.

An Algorithm for Enumerating SCCs in Web Graph.



Using Probabilistic Latent Semantic Analysis for Personalized Web Search.
Abstract. Web users use search engine to find useful information on the Internet. However current web search engines return answer to a query independent of specific user information need. Since web users with similar web behaviors tend to acquire similar information when they submit a same query, these unseen factors can be used to improve search result. In this paper we present an approach that mines these unseen factors from web logs to personalized web search. Our approach is based on probabilistic latent semantic analysis, a model based technique that is used to analyze co-occurrence data. Experimental results on real data collected by MSN search engine show the improvements over traditional web search. 1
Scalable collaborative filtering using cluster%2Dbased smoothing.
Memory-based approaches for collaborative filtering identify the similarity between two users by comparing their ratings on a set of items. In the past, the memory-based approaches have been shown to suffer from two fundamental problems: data sparsity and difficulty in scalability. Alternatively, the model-based approaches have been proposed to alleviate these problems, but these approaches tends to limit the range of users. In this paper, we present a novel approach that combines the advantages of these two kinds of approaches by introducing a smoothing-based method. In our approach, clusters generated from the training data provide the basis for data smoothing and neighborhood selection. As a result, we provide higher accuracy as well as increased efficiency in recommendations. Empirical studies on two datasets (EachMovie and MovieLens) show that our new proposed approach consistently outperforms other state-of-the-art collaborative filtering algorithms.
Exploiting the hierarchical structure for link analysis.
Link analysis algorithms have been extensively used in Web information retrieval. However, current link analysis algorithms generally work on a flat link graph, ignoring the hierarchal structure of the Web graph. They often suffer from two problems: the sparsity of link graph and biased ranking of newly-emerging pages. In this paper, we propose a novel ranking algorithm called Hierarchical Rank as a solution to these two problems, which considers both the hierarchical structure and the link structure of the Web. In this algorithm, Web pages are first aggregated based on their hierarchical structure at directory, host or domain level and link analysis is performed on the aggregated graph. Then, the importance of each node on the aggregated graph is distributed to individual pages belong to the node based on the hierarchical structure. This algorithm allows the importance of linked Web pages to be distributed in the Web page space even when the space is sparse and contains new pages. Experimental results on the.GOV collection of TREC 2003 and 2004 show that hierarchical ranking algorithm consistently outperforms other well-known ranking algorithms, including the PageRank, BlockRank and LayerRank. In addition, experimental results show that link aggregation at the host level is much better than link aggregation at either the domain or directory levels.
Importance%2DBased Web Page Classification Using Cost%2DSensitive SVM.

Interactive Chinese Search Results Clustering for Personalization.

A study on combination of block importance and relevance to estimate page relevance.

Multi%2Dmodel similarity propagation and its application for web image retrieval.

Optimizing web search using web click%2Dthrough data.
Search System
MRSSA an iterative algorithm for similarity spreading over interrelated objects.

IRC An Iterative Reinforcement Categorization Algorithm for Interrelated Web Objects.
Most existing categorization algorithms deal with homogeneous Web data objects, and consider interrelated objects as additional features when taking the interrelationships with other types of objects into account. However, focusing on any single aspects of these interrelationships and objects will not fully reveal their true categories. In this paper, we propose a novel categorization algorithm, the Iterative Reinforcement Categorization algorithm (IRC), to exploit the full interrelationships between the heterogeneous objects on the Web. IRC attempts to classify the interrelated Web objects by iterative reinforcement between individual classification results of different types via the interrelationships. Experiments on a clickthrough log dataset from MSN search engine show that, with the F1 measures, IRC achieves a 26.4 % improvement over a pure content-based classification method, a 21% improvement over a query metadata-based method, and a 16.4 % improvement over a virtual document-based method. Furthermore, our experiments show that IRC converges rapidly. 1.
Microsoft Research Asia at Web Track and Terabyte Track of TREC 2004.

DHT Based Searching Improved by Sliding Window.

Exploiting PageRank at Different Block Level.
Abstract. In recent years, information retrieval methods focusing on the link analysis have been developed; The PageRank and HITS are two typical ones According to the hierarchical organization of Web pages, we could partition the Web graph into blocks at different level, such as page level, directory level, host level and domain level. On the basis of block, we could analyze the different hyperlinks among pages. Several approaches proposed that the intrahyperlink in a host maybe less useful in computing the PageRank. However, there are no reports on how concretely the intra- or inter-hyperlink affects the PageRank. Furthermore, based on different block level, inter-hyperlink and intra-hyperlink can be two relative concepts. Thus which level should be optimal to distinguish the intra- or inter-hyperlink? And how the ratio set between the intra-hyperlink and inter-hyperlink could ultimately improve performance of the PageRank algorithm? In this paper, we analyze the link distribution at the different block level and evaluate the importance of the intra- and interhyperlink to PageRank on the TREC Web Track data set. Experiment shows that, if we set the block at host level and the ratio of the weight between the intra-hyperlink and inter-hyperlink is 1:4, the retrieval could achieve the best performance. 1
Multi%2Dtype Features Based Web Document Clustering.
Abstract. Clustering has been demonstrated as a feasible way to explore the contents of document collection and organize search engine results. For this task, many features of Web page, such as content, anchor text, URL, hyperlink etc, can be exploited and different results can be obtained. We expect to provide a unified and even better result for end users. Some work have studied how to use several types of features together to perform clustering. Most of them focus on ensemble method or combination of similarity. In this paper, we propose a novel algorithm: Multi-type Features based Reinforcement Clustering (MFRC). This algorithm does not use a unique combine score for all feature spaces, but uses the intermediate clustering result in one feature space as additional information to gradually enhance clustering in other spaces. Finally a consensus can be achieved by such mutual reinforcement. And the experimental results show that MFRC also provides some performance improvement. 1
Optimizing Web Search Using Spreading Activation on the Clickthrough Data.

Similarity spreading a unified framework for similarity calculation of interrelated objects.
In many Web search applications, similarities between objects of one type (say, queries) can be affected by the similarities between their interrelated objects of another type (say, Web pages), and vice versa. We propose a novel framework called similarity spreading to take account of the interrelationship and improve the similarity calculation. Experiment results show that the proposed framework can significantly improve the accuracy of the similarity measurement of the objects in a search engine. Categories &amp; Subject Descriptors:
Similarity spreading a unified framework for similarity calculation of interrelated objects.
In many Web search applications, similarities between objects of one type (say, queries) can be affected by the similarities between their interrelated objects of another type (say, Web pages), and vice versa. We propose a novel framework called similarity spreading to take account of the interrelationship and improve the similarity calculation. Experiment results show that the proposed framework can significantly improve the accuracy of the similarity measurement of the objects in a search engine. Categories &amp; Subject Descriptors:
TSSP A Reinforcement Algorithm to Find Related Papers.
Content analysis and citation analysis are two common methods in recommending system. Compared with content analysis, citation analysis can discover more implicitly related papers. However, the citation-based methods may introduce more noise in citation graph and cause topic drift. Some work combine content with citation to improve similarity measurement. The problem is that the two features are not used to reinforce each other to get better result. To solve the problem, we propose a new algorithm, Topic Sensitive Similarity Propagation (TSSP), to effectively integrate content similarity into similarity propagation. TSSP has two parts: citation context based propagation and iterative reinforcement. First, citation contexts provide clues for which papers are topic related to and filter out less irrelevant citations. Second, iteratively integrating content and citation similarity enable them to reinforce each other during the propagation. The experimental results of a user study show TSSP outperforms other algorithms in almost all cases. 1.
Implicit link analysis for small web search.

User Access Pattern Enhanced Small Web Search.

Log Mining to Improve the Performance of Site Search.
Despite of the popularity of global search engines, people still suffer from low accuracy of site search. The primary reason lies in the difference of link structures and data scale between global Web and website, which leads&#039; to failures of traditional re-ranking methods&#039; such as HITS, PageRank and DirectHit. This&#039; paper proposes a novel re-ranking method based on user logs within websites. With the help of website taxonomy, we mine for generalized association rules and abstract access patterns of different levels&#039;. Mining results&#039; are subsequently used to re-rank the retrieved pages. One of  the advantages of our mining algorithm is that it resolves the diversity problem of user&#039;s&#039; access behavior and discovers&#039; general patterns. Experiment shows that the proposed method outperforms keyword-based method by 15% and DirectHit by 13% respectively.
