Early assessment, with the help of machine learning methods, can aid clinicians in optimizing the diagnosis and treatment process, allowing patients to receive critical treatment time. Due to the advantages of effective information organization and interpretable reasoning, knowledge graph-based methods have become one of the most widely used machine learning algorithms for this task. However, due to a lack of effective organization and use of multi-granularity and temporal information, current knowledge graph-based approaches are hard to fully and comprehensively exploit the information contained in medical records, restricting their capacity to make superior quality diagnoses. To address these challenges, we examine and study disease diagnosis applications in-depth, and propose a novel disease diagnosis framework named FIT-Graph. With novel medical multi-grained evolutionary graphs, FIT-Graph efficiently organizes the extracted information from various granularities and time stages, maximizing the retention of valuable information for disease inference and ensuring the comprehensiveness and validity of the final disease inference. We compare FIT-Graph with two real-world clinical datasets from cardiology and respiratory departments with the baseline. The experimental results show that its effect is better than the baseline model, and the baseline performance of the task is improved by about 5% in multiple indices.

While matrix factorisation models are ubiquitous in large scale recommendation and search, real time application of such models requires inner product computations over an intractably large set of item factors. In this manuscript we present a novel framework that exploits structural properties of sparse vectors, using the inverted index representation, to significantly reduce the run time computational cost of factorisation models. We develop techniques that use geometry aware permutation maps on a tessellated unit sphere to obtain high dimensional sparse embeddings for latent factors with sparsity patterns related to angular closeness of the original latent factors. We also design several efficient and deterministic realisations within this framework and demonstrate with experiments that our techniques lead to faster run time operation with minimal loss of accuracy.

Factorization Machines (FM) have been recognized as an effective learning paradigm for incorporating complex relations to improve item recommendation in recommender systems. However, one open issue of FM lies in its factorized representation (latent factors) for each feature in the observed feature space, a characteristic often resulting in a large parameter space. Therefore, training FM (in other words, learning a large number of parameters in FM) is a computationally expensive task. Our work targets to improve the scalability of FM by building it in a distributed environment. We propose a new system framework that integrates Parameter Server (PS) with the Map/Reduce (MR) framework. In addition to the data parallelism achieved via MR, our framework particularly benefits from PS for model parallelism, a critical characteristic for learning with a large number of parameters in FM. We further address two specific challenges in our system, namely, communication cost and parameter update collision. Through both offline and online experiments on recommendation tasks, we demonstrate that the proposed system framework succeeds in scaling up FM for very large datasets, while it also maintains competitive performance on recommendation quality compared to alternative baselines.

Transfer learning, which leverages knowledge from source domains to enhance learning ability in a target domain, has been proven effective in various applications. One major limitation of transfer learning is that the source and target domains should be directly related. If there is little overlap between the two domains, performing knowledge transfer between these domains will not be effective. Inspired by human transitive inference and learning ability, whereby two seemingly unrelated concepts can be connected by a string of intermediate bridges using auxiliary concepts, in this paper we study a novel learning problem: Transitive Transfer Learning (abbreviated to TTL). TTL is aimed at breaking the large domain distances and transfer knowledge even when the source and target domains share few factors directly. For example, when the source and target domains are documents and images respectively, TTL could use some annotated images as the intermediate domain to bridge them. To solve the TTL problem, we propose a learning framework to mimic the human learning process. The framework is composed of an intermediate domain selection component and a knowledge transfer component. Extensive empirical evidence shows that the framework yields state-of-the-art classification accuracies on several classification data sets.

Content recommendation systems are typically based on one of the following paradigms: user based customization, or recommendations based on either collaborative filtering or low rank matrix factorization methods, or with systems that impute user interest profiles based on content browsing behavior and retrieve items similar to the interest profiles. All of these systems have a distinct disadvantage, namely data sparsity and cold-start on items or users. Furthermore, very few content recommendation solutions explicitly model the wealth of information in implicit negative feedback from the users. In this paper, we propose a hybrid solution that makes use of a latent factor model to infer user interest vectors. The hybrid approach enables us to overcome both the data sparsity and cold-start problems. Our proposed method is learned purely on implicit user feedback, both positive and negative. Exploiting the information in the negative feedback allows the user profiles generated to be discriminative. We also provide a Map/Reduce framework based implementation that enables scaling our solution to real-world recommendation problems. We demonstrate the efficacy of our proposed approach with both offline experiments and A/B tests on live traffic on Yahoo properties.

Time-sync video tagging aims to automatically generate tags for each video shot. It can improve the user's experience in previewing a video's timeline structure compared to traditional schemes that tag an entire video clip. In this paper, we propose a new application which extracts time-sync video tags by automatically exploiting crowdsourced comments from video websites such as Nico Nico Douga, where videos are commented on by online crowd users in a time-sync manner. The challenge of the proposed application is that users with bias interact with one another frequently and bring noise into the data, while the comments are too sparse to compensate for the noise. Previous techniques are unable to handle this task well as they consider video semantics independently, which may overfit the sparse comments in each shot and thus fail to provide accurate modeling. To resolve these issues, we propose a novel temporal and personalized topic model that jointly considers temporal dependencies between video semantics, users' interaction in commenting, and users' preferences as prior knowledge. Our proposed model shares knowledge across video shots via users to enrich the short comments, and peels off user interaction and user bias to solve the noisy-comment problem. Log-likelihood analyses and user studies on large datasets show that the proposed model outperforms several state-of-the-art baselines in video tagging quality. Case studies also demonstrate our model's capability of extracting tags from the crowdsourced short and noisy comments.

Music emotion recognition, which aims to automatically recognize the affective content of a piece of music, has become one of the key components of music searching, exploring, and social networking applications. Although researchers have given more and more attention to music emotion recognition studies, the recognition performance has come to a bottleneck in recent years. One major reason is that experts' labels for music emotion are mostly song-level, while music emotion usually varies within a song. Traditional methods have considered each song as a single instance and have built models based on song-level features. However, they ignored the dynamics of music emotion and failed to capture accurate emotion-feature correlations. In this paper, we model music emotion recognition as a novel multi-label multi-layer multi-instance multi-view learning problem: music is formulated as a hierarchical multi-instance structure (e.g., song-segment-sentence) where multiple emotion labels correspond to at least one of the instances with multiple views of each layer. We propose a Hierarchical Music Emotion Recognition model (HMER) -- a novel hierarchical Bayesian model using sentence-level music and lyrics features. It captures music emotion dynamics with a song-segment-sentence hierarchical structure. HMER also considers emotion correlations between both music segments and sentences. Experimental results show that HMER outperforms several state-of-the-art methods in terms of $F_1$ score and mean average precision.

Many internet companies, such as Yahoo, Facebook, Google and Twitter, rely on content recommendation systems to deliver the most relevant content items to individual users through personalization. Delivering such personalized user experiences is believed to increase the long term engagement of users. While there has been a lot of progress in designing effective personalized recommender systems, by exploiting user interests and historical interaction data through implicit (item click) or explicit (item rating) feedback, directly optimizing for users' satisfaction with the system remains challenging. In this paper, we explore the idea of using item-level dwell time as a proxy to quantify how likely a content item is relevant to a particular user. We describe a novel method to compute accurate dwell time based on client-side and server-side logging and demonstrate how to normalize dwell time across different devices and contexts. In addition, we describe our experiments in incorporating dwell time into state-of-the-art learning to rank techniques and collaborative filtering models that obtain competitive performances in both offline and online settings.

Heterogeneous transfer learning has been proposed as a new learning strategy to improve performance in a target domain by leveraging data from other heterogeneous source domains where feature spaces can be different across different domains. In order to connect two different spaces, one common technique is to bridge feature spaces by using some co-occurrence data. For example, annotated images can be used to build feature mapping from words to image features, and then applied on text-to-image knowledge transfer. However, in practice, such co-occurrence data are often from Web, e.g. Flickr, and generated by users. That means these data can be sparse and contain personal biases. Directly building models based on them may fail to provide reliable bridge. To solve these aforementioned problems, in this paper, we propose a novel algorithm named Mixed-Transfer. It is composed of three components, that is, a cross domain harmonic function to avoid personal biases, a joint transition probability graph of mixed instances and features to model the heterogeneous transfer learning problem, a random walk process to simulate the label propagation on the graph and avoid the data sparsity problem. We conduct experiments on 171 real-world tasks, showing that the proposed approach outperforms four state-of-the-art heterogeneous transfer learning algorithms.

One important challenge in social network analysis is how to model users’ distance as a single measure. We propose to model this distance by simultaneously exploring users’ profile attributes and local network structures. Due to the sparsity of data, where each user may interact with just a few people and only a few users provide their profile information, it is typically difficult to learn effective distance measures for any individual network. One important observation is that, people nowadays engage in multiple social networks, such as Facebook, Twitter, etc., where auxiliary knowledge from related networks can help alleviate the data sparsity problem. Nonetheless, due to the network differences, borrowing knowledge directly does not work well. Instead, we propose an adaptive metric learning framework. The basic idea is to exploit knowledge from related networks collectively through embedding and employ boosting-based techniques to eliminate irrelevant attributes. We evaluate the adaptive user distance measure on link prediction problem - an important social modeling task. Empirical studies demonstrate that the proposed approach significantly improves the link-prediction precision over state-of-the-art metric learning and link prediction approaches on two large-scale social networking datasets significantly.

Accurate prediction of user behaviors is important for many social media applications, including social marketing, personalization, and recommendation. A major challenge lies in that although many previous works model user behavior from only historical behavior logs, the available user behavior data or interactions between users and items in a given social network are usually very limited and sparse (e.g., ⩾ 99.9% empty), which makes models overfit the rare observations and fail to provide accurate predictions. We observe that many people are members of several social networks in the same time, such as Facebook, Twitter, and Tencent’s QQ. Importantly, users’ behaviors and interests in different networks influence one another. This provides an opportunity to leverage the knowledge of user behaviors in different networks by considering the overlapping users in different networks as bridges, in order to alleviate the data sparsity problem, and enhance the predictive performance of user behavior modeling. Combining different networks “simply and naively” does not work well. In this article, we formulate the problem to model multiple networks as “adaptive composite transfer” and propose a framework called ComSoc . ComSoc first selects the most suitable networks inside a composite social network via a hierarchical Bayesian model, parameterized for individual users. It then builds topic models for user behavior prediction using both the relationships in the selected networks and related behavior data. With different relational regularization, we introduce different implementations, corresponding to different ways to transfer knowledge from composite social relations. To handle big data, we have implemented the algorithm using Map/Reduce. We demonstrate that the proposed composite network-based user behavior models significantly improve the predictive accuracy over a number of existing approaches on several real-world applications, including a very large social networking dataset from Tencent Inc.

Recommender systems, especially the newly launched ones, have to deal with the data-sparsity issue, where little existing rating information is available. Recently, transfer learning has been proposed to address this problem by leveraging the knowledge from related recommender systems where rich collaborative data are available. However, most previous transfer learning models assume that entity-correspondences across different systems are given as input, which means that for any entity (e.g., a user or an item) in a target system, its corresponding entity in a source system is known. This assumption can hardly be satisfied in real-world scenarios where entity-correspondences across systems are usually unknown, and the cost of identifying them can be expensive. For example, it is extremely difficult to identify whether a user A from Facebook and a user B from Twitter are the same person. In this paper, we propose a framework to construct entity correspondence with limited budget by using active learning to facilitate knowledge transfer across recommender systems. Specifically, for the purpose of maximizing knowledge transfer, we first iteratively select entities in the target system based on our proposed criterion to query their correspondences in the source system. We then plug the actively constructed entity-correspondence mapping into a general transferred collaborative-filtering model to improve recommendation quality. We perform extensive experiments on real world datasets to verify the effectiveness of our proposed framework for this cross-system recommendation problem.

The study of users' social behaviors has gained much research attention since the advent of various social media such as Facebook, Renren and Twitter. A major kind of applications is to predict a user's future activities based on his/her historical social behaviors. In this paper, we focus on a fundamental task: to predict a user's future activity levels in a social network, e.g. weekly activeness, active or inactive. This problem is closely related to Social Customer Relationship Management (Social CRM). Compared to traditional CRM, the three properties: user diversity, social influence, and dynamic nature of social networks, raise new challenges and opportunities to Social CRM. Firstly, the user diversity property implies that a global predictive model may not be precise for all users. On the other hand, historical data of individual users are too sparse to build precisely personalized models. Secondly, the social influence property suggests that relationships between users can be embedded to further boost prediction results on individual users. Finally, the dynamical nature of social networks means that users' behaviors may keep changing over time. To address these challenges, we develop a personalized and social regularized time-decay model for user activity level prediction. Experiments on the social media Renren validate the effectiveness of our proposed model compared with some baselines including traditional supervised learning methods and node classification methods in social networks.

Crowdsourcing is an effective method for collecting labeled data for various data mining tasks. It is critical to ensure the veracity of the produced data because responses collected from different users may be noisy and unreliable. Previous works solve this veracity problem by estimating both the user ability and question difficulty based on the knowledge in each task individually. In this case, each single task needs large amounts of data to provide accurate estimations. However, in practice, budgets provided by customers for a given target task may be limited, and hence each question can be presented to only a few users where each user can answer only a few questions. This data sparsity problem can cause previous approaches to perform poorly due to the overfitting problem on rare data and eventually damage the data veracity. Fortunately, in real-world applications, users can answer questions from multiple historical tasks. For example, one can annotate images as well as label the sentiment of a given title. In this paper, we employ transfer learning, which borrows knowledge from auxiliary historical tasks to improve the data veracity in a given target task. The motivation is that users have stable characteristics across different crowdsourcing tasks and thus data from different tasks can be exploited collectively to estimate users' abilities in the target task. We propose a hierarchical Bayesian model, TLC (Transfer Learning for Crowdsourcing), to implement this idea by considering the overlapping users as a bridge. In addition, to avoid possible negative impact, TLC introduces task-specific factors to model task differences. The experimental results show that TLC significantly improves the accuracy over several state-of-the-art non-transfer-learning approaches under very limited budget in various labeling tasks.

Modeling the dynamics of online social networks over time not only helps us understand the evolution of network structures and user behaviors, but also improves the performance of other analysis tasks, such as link prediction and community detection. Nowadays, users engage in multiple networks and form a "composite social network" by considering common users as the bridge. State-of-the-art network-dynamics analysis is performed in isolation for individual networks, but users' interactions in one network can influence their behaviors in other networks, and in an individual network, different types of user interactions also affect each other. Without considering the influences across networks, one may not be able to model the dynamics in a given network correctly due to the lack of information. In this paper, we study the problem of modeling the dynamics of composite networks, where the evolution processes of different networks are jointly considered. However, due to the difference in network properties, simply merging multiple networks into a single one is not ideal because individual evolution patterns may be ignored and network differences may bring negative impacts. The proposed solution is a nonparametric Bayesian model, which models each user's common latent features to extract the cross-network influences, and use network-specific factors to describe different networks' evolution patterns. Empirical studies on large-scale dynamic composite social networks demonstrate that the proposed approach improves the performance of link prediction over several state-of-the-art baselines and unfolds the network evolution accurately.

Music emotion recognition (MER) aims to recognize the affective content of a piece of music, which is important for applications such as automatic soundtrack generation and music recommendation. MER is commonly formulated as a supervised learning problem. In practice, except for Pop music, there is little labeled data in most genres. In addition, emotion is genre specific in music and thus the labeled data of Pop music cannot be used for other genres. In this paper, we aim to solve the genre-specific MER problem by exploiting two kinds of auxiliary data: unlabeled songs and social tags. However, using these two kinds of data effectively is a non-trivial task, e.g. tags are noisy and therefore cannot be treated as fully trustworthy. To build an accurate model with the help from the unlabeled songs and noisy tags, we present SMART, which stands for Semi-Supervised Music Affective Emotion Recognition with Social Tagging, combining of a graph-based semi-supervised learning algorithm with a novel tag refinement method. Experiments on the Million Song Dataset show that our proposed approach, trained with only 10 labeled instances, is as accurate as Support Vector Regression trained with 750 labeled songs.

Collaborative filtering (CF) aims to predict users’ ratings on items according to historical user-item preference data. In many real-world applications, preference data are usually sparse, which would make models overfit and fail to give accurate predictions. Recently, several research works show that by transferring knowledge from some manually selected source domains, the data sparseness problem could be mitigated. However for most cases, parts of source domain data are not consistent with the observations in the target domain, which may misguide the target domain model building. In this paper, we propose a novel criterion based on empirical prediction error and its variance to better capture the consistency across domains in CF settings. Consequently, we embed this criterion into a boosting framework to perform selective knowledge transfer. Comparing to several state-of-the-art methods, we show that our proposed selective transfer learning framework can significantly improve the accuracy of rating prediction on several real-world recommendation tasks.

Transfer learning, which aims to help the learning task in a target domain by leveraging knowledge from auxiliary domains, has been demonstrated to be effective in different applications, e.g., text mining, sentiment analysis, etc. In addition, in many real-world applications, auxiliary data are described from multiple perspectives and usually carried by multiple sources. For example, to help classify videos on Youtube, which include three views/perspectives: image, voice and subtitles, one may borrow data from Flickr, Last.FM and Google News. Although any single instance in these domains can only cover a part of the views available on Youtube, actually the piece of information carried by them may compensate with each other. In this paper, we define this transfer learning problem as Transfer Learning with Multiple Views and Multiple Sources. As different sources may have different probability distributions and different views may be compensate or inconsistent with each other, merging all data in a simplistic manner will not give optimal result. Thus, we propose a novel algorithm to leverage knowledge from different views and sources collaboratively, by letting different views from different sources complement each other through a co-training style framework, while revise the distribution differences in different domains. We conduct empirical studies on several real-world datasets to show that the proposed approach can improve the classification accuracy by up to 8% against different state-of-the-art baselines.

As the popularity of the social media increases, as evidenced in Twitter, Facebook and China's Renren, spamming activities also picked up in numbers and variety. On social network sites, spammers often disguise themselves by creating fake accounts and hijacking normal users' accounts for personal gains. Different from the spammers in traditional systems such as SMS and email, spammers in social media behave like normal users and they continue to change their spamming strategies to fool anti spamming systems. However, due to the privacy and resource concerns, many social media websites cannot fully monitor all the contents of users, making many of the previous approaches, such as topology-based and content-classification-based methods, infeasible to use. In this paper, we propose a novel method for spammer detection in social networks that exploits both social activities as well as users' social relations in an innovative and highly scalable manner. The proposed method detects spammers following collective activities based on users' social actions and relations. We have empirically tested our method on data from Renren.com, which is the largest social network in China, and demonstrated that our new method can improve the detection performance significantly.

Accurate prediction of user behaviors is important for many social media applications, including social marketing, personalization and recommendation, etc. A major challenge lies in that, the available behavior data or interactions between users and items in a given social network are usually very limited and sparse (e.g., >= 99.9% empty). Many previous works model user behavior from only historical user logs. We observe that many people are members of several social networks in the same time, such as Facebook, Twitter and Tencent's QQ. Importantly, their behaviors and interests in different networks influence one another. This gives us an opportunity to leverage the knowledge of user behaviors in different networks, in order to alleviate the data sparsity problem, and enhance the predictive performance of user modeling. Combining different networks "simply and naively" does not work well. Instead, we formulate the problem to model multiple networks as "composite network knowledge transfer". We first select the most suitable networks inside a composite social network via a hierarchical Bayesian model, parameterized for individual users, and then build topic models for user behavior prediction using both the relationships in the selected networks and related behavior data. To handle big data, we have implemented the algorithm using Map/Reduce. We demonstrate that the proposed composite network-based user behavior model significantly improve the predictive accuracy over a number of existing approaches on several real world applications, such as a very large social-networking dataset from Tencent Inc.

Matrix factorization (MF) has been demonstrated to be one of the most competitive techniques for collaborative filtering. However, state-of-the-art MFs do not consider contextual information, where ratings can be generated under different environments. For example, users select items under various situations, such as happy mood vs. sad, mobile vs. stationary, movies vs. book, etc. Under different contexts, the preference of users are inherently different. The problem is that MF methods uniformly decompose the rating matrix, and thus they are unable to factorize for different contexts. To amend this problem and improve recommendation accuracy, we introduce a “hierarchical” factorization model by considering the local context when performing matrix factorization. The intuition is that: as ratings are being generated from heterogeneous environments, certain user and item pairs tend to be more similar to each other than others, and hence they ought to receive more collaborative information from each other. To take the contextual information into consideration, the proposed “contextual collaborative filtering” approach splits the rating matrix hierarchically by grouping similar users and items together, and factorizes each sub-matrix locally under different contexts. By building an ensemble model, the approach further avoids over-fitting with less parameter tuning. We analyze and demonstrate that the proposed method is a model-averaging gradient boosting model, and its error rate can be bounded. Experimental results show that it outperforms three state-of-the-art algorithms on a number of real-world datasets (Movie-Lens, Netflix, etc). The source code and datasets are available for download.

State-of-the-art learning algorithms accept data in feature vector format as input. Examples belonging to different classes may not always be easy to separate in the original feature space. One may ask: can transformation of existing features into new space reveal significant discriminative information not obvious in the original space? Since there can be infinite number of ways to extend features, it is impractical to first enumerate and then perform feature selection. Second, evaluation of discriminative power on the complete dataset is not always optimal. This is because features highly discriminative on subset of examples may not necessarily be significant when evaluated on the entire dataset. Third, feature construction ought to be automated and general, such that, it doesn't require domain knowledge and its improved accuracy maintains over a large number of classification algorithms. In this paper, we propose a framework to address these problems through the following steps: (1) divide-conquer to avoid exhaustive enumeration; (2) local feature construction and evaluation within subspaces of examples where local error is still high and constructed features thus far still do not predict well; (3) weighting rules based search that is domain knowledge free and has provable performance guarantee. Empirical studies indicate that significant improvement (as much as 9% in accuracy and 28% in AUC) is achieved using the newly constructed features over a variety of inductive learners evaluated against a number of balanced, skewed and high-dimensional datasets. Software and datasets are available from the authors.

When labeled examples are limited and difficult to obtain, transfer learning employs knowledge from a source domain to improve learning accuracy in the target domain. However, the assumption made by existing approaches, that the marginal and conditional probabilities are directly related between source and target domains, has limited applicability in either the original space or its linear transformations. To solve this problem, we propose an adaptive kernel approach that maps the marginal distribution of targetdomain and source-domain data into a common kernel space, and utilize a sample selection strategy to draw conditional probabilities between the two domains closer. We formally show that under the kernel-mapping space, the difference in distributions between the two domains is bounded; and the prediction error of the proposed approach can also be bounded. Experimental results demonstrate that the proposed method outperforms both traditional inductive classifiers and the state-of-the-art boosting-based transfer algorithms on most domains, including text categorization and web page ratings. In particular, it can achieve around 10 % higher accuracy than other approaches for the text categorization problem. The source code and datasets are available from the authors.

When the number of labeled examples is limited, traditional supervised feature selection techniques often fail due to sample selection bias or unrepresentative sample problem. To solve this, semi-supervised feature selection techniques exploit the statistical information of both labeled and unlabeled examples in the same time. However, the results of semi-supervised feature selection can be at times unsatisfactory, and the culprit is on how to effectively use the unlabeled data. Quite different from both supervised and semi-supervised feature selection, we propose a ldquohybridrdquoframework based on graph models. We first apply supervised methods to select a small set of most critical features from the labeled data. Importantly, these initial features might otherwise be missed when selection is performed on the labeled and unlabeled examples simultaneously. Next,this initial feature set is expanded and corrected with the use of unlabeled data. We formally analyze why the expected performance of the hybrid framework is better than both supervised and semi-supervised feature selection. Experimental results demonstrate that the proposed method outperforms both traditional supervised and state-of-the-art semi-supervised feature selection algorithms by at least 10% inaccuracy on a number of text and biomedical problems with thousands of features to choose from. Software and dataset is available from the authors.