Online discovery of similarity mappings

We consider the problem of choosing, sequentially, a map which assigns elements of a set A to a few elements of a set B. On each round, the algorithm suffers some cost associated with the chosen assignment, and the goal is to minimize the cumulative loss of these choices relative to the best map on the entire sequence. Even though the offline problem of finding the best map is provably hard, we show that there is an equivalent online approximation algorithm, Randomized Map Prediction (RMP), that is efficient and performs nearly as well. While drawing upon results from the "Online Prediction with Expert Advice" setting, we show how RMP can be utilized as an online approach to several standard batch problems. We apply RMP to online clustering as well as online feature selection and, surprisingly, RMP often outperforms the standard batch algorithms on these problems.

Open problems in the security of learning

Machine learning has become a valuable tool for detecting and preventing malicious activity. However, as more applications employ machine learning techniques in adversarial decision-making situations, increasingly powerful attacks become possible against machine learning systems. In this paper, we present three broad research directions towards the end of developing truly secure learning. First, we suggest that finding bounds on adversarial influence is important to understand the limits of what an attacker can and cannot do to a learning system. Second, we investigate the value of adversarial capabilities-the success of an attack depends largely on what types of information and influence the attacker has. Finally, we propose directions in technologies for secure learning and suggest lines of investigation into secure techniques for learning in adversarial environments. We intend this paper to foster discussion about the security of machine learning, and we believe that the research directions we propose represent the most important directions to pursue in the quest for secure learning.

Learning to act in uncertain environments : technical perspective

The problem of decision making in an uncertain environment arises in many diverse contexts: deciding whether to keep a hard drive spinning in a net-book; choosing which advertisement to post to a Web site visitor; choosing how many newspapers to order so as to maximize profits; or choosing a route to recommend to a driver given limited and possibly out-of-date information about traffic conditions. All are sequential decision problems, since earlier decisions affect subsequent performance; all require adaptive approaches, since they involve significant uncertainty. The key issue in effectively solving problems like these is known as the exploration/exploitation trade-off: If I am at a cross-roads, when should I go in the most advantageous direction among those that I have already explored, and when should I strike out in a new direction, in the hopes I will discover something better?

Learning domain ontology for tag recommendation

Recently, user tagging systems have grown in popularity on the web. The tagging process is quite simple for ordinary users, which contributes to its popularity. However, free vocabulary has lack of standardization and semantic ambiguity. It is possible to capture the semantics from user tagging and represent those in a form of ontology, but the application of the learned ontology for recommendation making has not been that flourishing. In this paper we discuss our approach to learn domain ontology from user tagging information and apply the extracted tag ontology in a pilot tag recommendation experiment. The initial result shows that by using the tag ontology to re-rank the recommended tags, the accuracy of the tag recommendation can be improved.

p53-independent NOXA induction overcomes apoptotic resistance of malignant melanomas

Once melanoma metastasizes, no effective treatment modalities prolong survival in most patients. This notorious refractoriness to therapy challenges investigators to identify agents that overcome melanoma resistance to apoptosis. Whereas many survival pathways contribute to the death-defying phenotype in melanoma, a defect in apoptotic machinery previously highlighted inactivation of Apaf-1, an apoptosome component engaged after mitochondrial damage. During studies involving Notch signaling in melanoma, we observed a gamma-secretase tripeptide inhibitor (GSI; z-Leu-Leu-Nle-CHO), selected from a group of compounds originally used in Alzheimer's disease, induced apoptosis in nine of nine melanoma lines. GSI only induced G2-M growth arrest (but not killing) in five of five normal melanocyte cultures tested. Effective killing of melanoma cells by GSI involved new protein synthesis and a mitochondrial-based pathway mediated by up-regulation of BH3-only members (Bim and NOXA). p53 activation was not necessary for up-regulation of NOXA in melanoma cells. Blocking GSI-induced NOXA using an antisense (but not control) oligonucleotide significantly reduced the apoptotic response. GSI also killed melanoma cell lines with low Apaf-1 levels. We conclude that GSI is highly effective in killing melanoma cells while sparing normal melanocytes. Direct enhancement of BH3-only proteins executes an apoptotic program overcoming resistance of this lethal tumor. Identification of a p53-independent apoptotic pathway in melanoma cells, including cells with low Apaf-1, bypasses an impediment to current cytotoxic therapy and provides new targets for future therapeutic trials involving chemoresistant tumors.

Query expansion using term relationships in language models for information retrieval

Language Modeling (LM) has been successfully applied to Information Retrieval (IR). However, most of the existing LM approaches only rely on term occurrences in documents, queries and document collections. In traditional unigram based models, terms (or words) are usually considered to be independent. In some recent studies, dependence models have been proposed to incorporate term relationships into LM, so that links can be created between words in the same sentence, and term relationships (e.g. synonymy) can be used to expand the document model. In this study, we further extend this family of dependence models in the following two ways: (1) Term relationships are used to expand query model instead of document model, so that query expansion process can be naturally implemented; (2) We exploit more sophisticated inferential relationships extracted with Information Flow (IF). Information flow relationships are not simply pairwise term relationships as those used in previous studies, but are between a set of terms and another term. They allow for context-dependent query expansion. Our experiments conducted on TREC collections show that we can obtain large and significant improvements with our approach. This study shows that LM is an appropriate framework to implement effective query expansion.

Modeling key compromise impersonation attacks on group key exchange protocols

Two-party key exchange (2PKE) protocols have been rigorously analyzed under various models considering different adversarial actions. However, the analysis of group key exchange (GKE) protocols has not been as extensive as that of 2PKE protocols. Particularly, an important security attribute called key compromise impersonation (KCI) resilience has been completely ignored for the case of GKE protocols. Informally, a protocol is said to provide KCI resilience if the compromise of the long-term secret key of a protocol participant A does not allow the adversary to impersonate an honest participant B to A. In this paper, we argue that KCI resilience for GKE protocols is at least as important as it is for 2PKE protocols. Our first contribution is revised definitions of security for GKE protocols considering KCI attacks by both outsider and insider adversaries. We also give a new proof of security for an existing two-round GKE protocol under the revised security definitions assuming random oracles. We then show how to achieve insider KCIR in a generic way using a known compiler in the literature. As one may expect, this additional security assurance comes at the cost of an extra round of communication. Finally, we show that a few existing protocols are not secure against outsider KCI attacks. The attacks on these protocols illustrate the necessity of considering KCI resilience for GKE protocols.

Expressive interactions : tablet usability for young mobile learners

In this paper we examine the usability of tablets for students in middle school in the context of mobile environmental education. Our study focuses on the expressive qualities of three input methods – text, audio and drawing – and the extent to which these methods support on-task behaviour. In our study 28 small groups of children were given iPads and asked to record ecological observations from around their schoolyard. The effectiveness of the devices and their core utility for expressive, on-task data capture is assessed.

Modeling player-like behaviour for game Al design

The design of artificial intelligence in computer games is an important component of a player's game play experience. As games are becoming more life-like and interactive, the need for more realistic game AI will increase. This is particularly the case with respect to AI that simulates how human players act, behave and make decisions. The purpose of this research is to establish a model of player-like behavior that may be effectively used to inform the design of artificial intelligence to more accurately mimic a player's decision making process. The research uses a qualitative analysis of player opinions and reactions while playing a first person shooter video game, with recordings of their in game actions, speech and facial characteristics. The initial studies provide player data that has been used to design a model of how a player behaves.

Stomp : an interactive platform for people with intellectual disabilities

For people with intellectual disabilities there are significant barriers to inclusion in socially cooperative endeavours. This paper investigates the effectiveness of Stomp, a tangible user interface (TUI) designed to provide new participatory experiences for people with intellectual disability. Results from an observational study reveal the extent to which the Stomp system supports social and physical interaction. The tangible, spatial and embodied qualities of Stomp result in an experience that does not rely on the acquisition of specific competencies before interaction and engagement can occur.

Designing an immersive environment for public use

Bystander is a multi-user, immersive, interactive environment intended for public display in a museum or art gallery. It is designed to make available heritage collections in novel and culturally responsible ways. We use its development as a case study to examine the role played in that process by a range of tools and techniques from participatory design traditions. We describe how different tools were used within the design process, specifically: the ways in which the potential audience members were both included and represented; the prototypes that have been constructed as a way of envisioning how the final work might be experienced; and how these tools have been brought together in ongoing designing and evaluation. We close the paper with some reflections on the extension of participatory commitments into still-emerging areas of technology design that prioritise the design of spaces for human experience and reflective interaction.

Engaging new digital locals with interactive urban screens to collaboratively improve the city

Local governments struggle to engage time poor and seemingly apathetic citizens, as well as the city’s young digital natives, the digital locals. This project aims at providing a lightweight, technological contribution towards removing the hierarchy between those who build the city and those who use it. We aim to narrow this gap by enhancing people’s experience of physical spaces with digital, civic technologies that are directly accessible within that space. This paper presents the findings of a design trial allowing users to interact with a public screen via their mobile phones. The screen facilitated a feedback platform about a concrete urban planning project by promoting specific questions and encouraging direct, in-situ, real-time responses via SMS and twitter. This new mechanism offers additional benefits for civic participation as it gives voice to residents who otherwise would not be heard. It also promotes a positive attitude towards local governments and gathers information different from more traditional public engagement tools.

Text mining and probabilistic language modeling for online review spam detecting

In the era of Web 2.0, huge volumes of consumer reviews are posted to the Internet every day. Manual approaches to detecting and analyzing fake reviews (i.e., spam) are not practical due to the problem of information overload. However, the design and development of automated methods of detecting fake reviews is a challenging research problem. The main reason is that fake reviews are specifically composed to mislead readers, so they may appear the same as legitimate reviews (i.e., ham). As a result, discriminatory features that would enable individual reviews to be classified as spam or ham may not be available. Guided by the design science research methodology, the main contribution of this study is the design and instantiation of novel computational models for detecting fake reviews. In particular, a novel text mining model is developed and integrated into a semantic language model for the detection of untruthful reviews. The models are then evaluated based on a real-world dataset collected from amazon.com. The results of our experiments confirm that the proposed models outperform other well-known baseline models in detecting fake reviews. To the best of our knowledge, the work discussed in this article represents the first successful attempt to apply text mining methods and semantic language models to the detection of fake consumer reviews. A managerial implication of our research is that firms can apply our design artifacts to monitor online consumer reviews to develop effective marketing or product design strategies based on genuine consumer feedback posted to the Internet.