We consider the problem of learning a structured multi-task regression, where the output consists of multiple responses that are related by a graph and the correlated response variables are dependent on the common inputs in a sparse but synergistic manner. Previous methods such as l1/l2-regularized multi-task regression assume that all of the output variables are equally related to the inputs, although in many real-world problems, outputs are related in a complex manner. In this paper, we propose graph-guided fused lasso (GFlasso) for structured multi-task regression that exploits the graph structure over the output variables. We introduce a novel penalty function based on fusion penalty to encourage highly correlated outputs to share a common set of relevant inputs. In addition, we propose a simple yet efficient proximal-gradient method for optimizing GFlasso that can also be applied to any optimization problems with a convex smooth loss and the general class of fusion penalty defined on arbitrary graph structures. By exploiting the structure of the non-smooth ''fusion penalty'', our method achieves a faster convergence rate than the standard first-order method, sub-gradient method, and is significantly more scalable than the widely adopted second-order cone-programming and quadratic-programming formulations. In addition, we provide an analysis of the consistency property of the GFlasso model. Experimental results not only demonstrate the superiority of GFlasso over the standard lasso but also show the efficiency and scalability of our proximal-gradient method.

Variable selection in high dimensional space has challenged many contemporary statistical problems from many frontiers of scientific disciplines. Recent technology advance has made it possible to collect a huge amount of covariate information such as microarray, proteomic and SNP data via bioimaging technology while observing survival information on patients in clinical studies. Thus, the same challenge applies to the survival analysis in order to understand the association between genomics information and clinical information about the survival time. In this work, we extend the sure screening procedure Fan and Lv (2008) to Cox's proportional hazards model with an iterative version available. Numerical simulation studies have shown encouraging performance of the proposed method in comparison with other techniques such as LASSO. This demonstrates the utility and versatility of the iterative sure independent screening scheme.

Programs to solve so-called constraint problems are complex pieces of software which require many design decisions to be made more or less arbitrarily by the implementer. These decisions affect the performance of the finished solver significantly [13]. Once a design decision has been made, it cannot easily be reversed, although a different decision may be more appropriate for a particular problem. We investigate using machine learning to make these decisions automatically depending on the problem to solve with the alldifferent constraint as an example. Our system is capable of making nontrivial, multilevel decisions that improve over always making a default choice.

This work presents a novel learning method in the context of embodied artificial intelligence and self-organization, which has as few assumptions and restrictions as possible about the world and the underlying model. The learning rule is derived from the principle of maximizing the predictive information in the sensorimotor loop. It is evaluated on robot chains of varying length with individually controlled, non-communicating segments. The comparison of the results shows that maximizing the predictive information per wheel leads to a higher coordinated behavior of the physically connected robots compared to a maximization per robot. Another focus of this paper is the analysis of the effect of the robot chain length on the overall behavior of the robots. It will be shown that longer chains with less capable controllers outperform those of shorter length and more complex controllers. The reason is found and discussed in the information-geometric interpretation of the learning process.

Aiming at accurate recognition of relations, we introduce local alignment kernels and explore various possibilities of using them for this task. We give a definition of a local alignment (LA) kernel based on the Smith-Waterman score as a sequence similarity measure and proceed with a range of possibilities for computing similarity between elements of sequences. We show how distributional similarity measures obtained from unlabeled data can be incorporated into the learning task as semantic knowledge. Our experiments suggest that the LA kernel yields promising results on various biomedical corpora outperforming two baselines by a large margin. Additional series of experiments have been conducted on the data sets of seven general relation types, where the performance of the LA kernel is comparable to the current state-of-the-art results.

In many collaborative systems, researchers are interested in creating representative user profiles. In this paper, we are particularly interested in using social labeling and automatic keyword extraction techniques for generating user profiles. Social labeling is a process in which users manually tag other users with keywords. Automatic keyword extraction is a technique that selects the most salient words to represent a user’s contribution. We apply each of these two profile generation methods to highly active Wikipedia editors and their contributions, and compare the results. We found that profiles generated through social labeling matches the profiles generated via automatic keyword extraction, and vice versa. The results suggest that user profiles generated from one method can be used as a seed or bootstrapping proxy for the other method.

Personal photos and their associated metadata reveal different aspects of our lives and, when shared online, let others have an idea about us. Automating the extraction of personal information is an arduous task but it contributes to better understanding and serving users. Here we present methods for analyzing textual metadata associated to Flickr photos that unveil users’ home location and gender. We test our techniques on a sample of 30,000 people coming from six different countries, allowing us to compare results across cultures and point out similarities and differences.

Social media sites derive their value by providing a popular and dependable community for participants to engage, share, and interact. This community value and related services like search and advertising are threatened by spammers, content polluters, and malware disseminators. In an effort to preserve community value and ensure long-term success, we present a prototype system for automatically detecting and profiling destructive users in social media. We described the architecture of the system - inspired by the "broken windows" theory embraced by law enforcement - the results and insights gained from a preliminary study conducted to determine the efficacy of our approach, and a discussion of our ongoing research.

We propose a relational ranking model for learning to tag images in social media sharing systems. This model learns to associate a ranked list of tags to unlabeled images, by considering simultaneously content information (visual or textual) and relational information among the images. It is able to handle implicit relations like content similarities, and explicit ones like friendship or authorship. The model itself is based on a transductive algorithm thats learns from both labeled and unlabeled data. Experiments on a real corpus extracted from Flickr show the effectiveness of this model.

As microblogging grows in popularity, services like Twitter are coming to support information gathering needs above and beyond their traditional roles as social networks. But most users’ interaction with Twitter is still primarily focused on their social graphs, forcing the often inappropriate conflation of “people I follow” with “stuff I want to read.” We characterize some information needs that the current Twitter interface fails to support, and argue for better representations of content for solving these challenges. We present a scalable implementation of a partially supervised learning model (Labeled LDA) that maps the content of the Twitter feed into dimensions. These dimensions correspond roughly to substance, style, status, and social characteristics of posts. We characterize users and tweets using this model, and present results on two information consumption oriented tasks.