Yin Zhu’s research while affiliated with Hong Kong University of Science and Technology and other places

The recent advancements in online social networks and mobile devices have provided valuable data sources to track users’ smartphone adoption, i.e., the usage of smartphones over time. An incisive understanding of users’ smartphone adoption can benefit many useful applications, ranging from user behavior understanding to targeted marketing. This paper studies smartphone adoption prediction in social networks by leveraging the wisdom of an online world. A critical challenge along this line is to identify the key factors that underline people’s adoption behaviors and distinguish the relative contribution of each factor. Specifically, we model the final smartphone status of each user as a result of three influencing factors: the social influence factor, the homophily factor, and the personal factor. We further develop a supervised model that takes all three factors for smartphone adoption and at the same time learns the relative contribution of each factor from the data. Experimental results on a large real world dataset demonstrate the effectiveness of our proposed model.

Transfer learning uses relevant auxiliary data to help the learning task in a target domain where labeled data is usually insufficient to train an accurate model. Given appropriate auxiliary data, researchers have proposed many transfer learning models. How to find such auxiliary data, however, is of little research so far. In this paper, we focus on the problem of auxiliary data retrieval, and propose a transfer learning framework that effectively selects helpful auxiliary data from an open knowledge space (e.g. The World Wide Web). Because there is no need of manually selecting auxiliary data for different target domain tasks, we call our framework Source Free Transfer Learning (SFTL). For each target domain task, SFTL framework iteratively queries for the helpful auxiliary data based on the learned model and then updates the model using the retrieved auxiliary data. We highlight the automatic constructions of queries and the robustness of the SFTL framework. Our experiments on 20NewsGroup dataset and a Google search snippets dataset suggest that the framework is capable of achieving comparable performance to those state-of-the-art methods with dedicated selections of auxiliary data. Copyright © 2014, Association for the Advancement of Artificial Intelligence.

With the growing popularity of location-based social networks, vast amount of user check-in histories have been accumulated. Based on such historical data, predicting a user’s next check-in place is of much interest recently. There is, however, little study on the limit of predictability of this task and its correlation with users’ demographics. These studies can give deeper insight to the prediction task and bring valuable insights to the design of new prediction algorithms. In this paper, we carry out a thorough study on the limit of check-in location predictability, i.e., to what extent the next locations are predictable, in the presence of special properties of check-in traces. Specifically, we begin with estimating the entropy of an individual check-in trace and then leverage Fano’s inequality to transform it to predictability. Extensive analysis has then been performed on two large-scale check-in datasets from Jiepang and Gowalla with 36M and 6M check-ins, respectively. As a result, we find 25% and 38% potential predictability respectively. Finally, the correlation analysis between predictability and users’ demographics has been performed. The results show that the demographics, such as gender and age, are significantly correlated with location predictability.

We present in this paper our winning solution to Dedicated Task 1 in Nokia Mobile Data Challenge (MDC). MDC Task 1 is to infer the semantic category of a place based on the smartphone sensing data obtained at that place. We approach this task in a standard supervised learning setting: we extract discriminative features from the sensor data and use state-of-the-art classifiers (SVM, Logistic Regression and Decision Tree Family) to build classification models. We have found that feature engineering, or in other words, constructing features using human heuristics, is very effective for this task. In particular, we have proposed a novel feature engineering technique, Conditional Feature (CF), a general framework for domain-specific feature construction. In total, we have generated 2,796,200 features and in our final five submissions we use feature selection to select 100 to 2000 features. One of our key findings is that features conditioned on fine-granularity time intervals, e.g. every 30 min, are most effective. Our best 10-fold CV accuracy on training set is 75.1% by Gradient Boosted Trees, and the second best accuracy is 74.6% by L1-regularized Logistic Regression. Besides the good performance, we also report briefly our experience of using F# language for large-scale (∼70 GB raw text data) conditional feature construction.

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.

We present the ideas and methodologies that we used to address the KDD Cup 2013 challenge on author-paper identification. We firstly formulate the problem as a personalized ranking task and then propose to solve the task through a supervised learning framework. The key point is to eliminate those incorrectly assigned papers of a given author based on existing records. We choose Gradient Boosted Tree as our main classifier. Through our exploration we conclude that the most critical factor to achieve our results is the effective feature engineering. In this paper, we formulate this process as a unified framework that constructs features based on contextual information and combines machine learning techniques with human intelligence. Besides this, we suggest several strategies to parse authors' names, which improve the prediction results significantly. Divide-conquer based model building as well as the model averaging techniques also benefit the prediction precision.

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.

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 dis-guise themselves by creating fake accounts and hijack-ing normal users' accounts for personal gains. Different from the spammers in traditional systems such as SMS and email, spammers in social media behave like nor-mal users and they continue to change their spamming strategies to fool anti-spamming systems. However, due to the privacy and resource concerns, many social me-dia websites cannot fully monitor all the contents of users, making many of the previous approaches, such as topology-based and content-classification-based meth-ods, infeasible to use. In this paper, we propose a Su-pervised Matrix Factorization method with Social Reg-ularization (SMFSR) for spammer detection in social networks that exploits both social activities as well as users' social relations in an innovative and highly scal-able manner. The proposed method detects spammers collectively based on users' social actions and social re-lations. We have empirically tested our method on data from Renren.com, which is one of the largest social net-works in China, and demonstrated that our new method can improve the detection performance significantly.

In today's world, we have increasingly sophisticated means to record the movement of humans and other moving objects in the form of trajectory data. These data are being accumulated at an extremely fast rate. As a result, knowledge discovery from these data for recognizing activities has become an important problem. The discovered activity patterns can help us understand people's lives, analyze traffic in a large city and study social networks among people. Trajectory-based activity recognition builds upon some fundamental functions of location estimation and machine learning, and can provide new insights on how to infer high-level goals and objectives from low-level sensor readings. In this chapter, we survey the area of trajectory-based activity recognition. We start from research in location estimation from sensors for obtaining the trajectories. We then review trajectory-based activity recognition research. We classify the research work on trajectory-based activity recognition into several broad categories, and systematically summarize existing work as well as future works in light of the categorization.

Transfer learning as a new machine learning paradigm has gained increasing attention lately. In situations where the training data in a target domain are not sufficient to learn predictive models effectively, transfer learning leverages auxiliary source data from other related source domains for learning. While most of the existing works in this area only focused on using the source data with the same structure as the target data, in this paper, we push this boundary further by proposing a heterogeneous transfer learning framework for knowledge transfer between text and images. We observe that for a target-domain classification problem, some annotated images can be found on many social Web sites, which can serve as a bridge to transfer knowledge from the abundant text documents available over the Web. A key question is how to effectively transfer the knowledge in the source data even though the text can be arbitrarily found. Our solution is to enrich the representation of the target images with semantic concepts extracted from the auxiliary source data through a novel matrix factorization method. By using the latent semantic features generated by the auxiliary data, we are able to build a better integrated image classifier. We empirically demonstrate the effectiveness of our algorithm on the Caltech-256 image dataset.