Human Centric Security

How does IT change our interactions, and how can more trustworthy IT change them further? This includes studies of on-line trust, ecommerce (business-to-business and business-to-consumer), digital government services, e-conferencing, on-line personae and anonymity, online news, on-line research and the ephemeral nature of information, on-line propaganda, and spam.

The Cloud’s DNA

When conducting research, life scientists rely heavily on clinically annotated specimens, and the most thorough and effective clinical annotations contain information that is found in the electronic health records (EHRs) for the human subjects that are participating in the scientists’ studies. One primary piece of legislation pertinent to electronic health records is the Health Information Portability and Accountability Act (HIPAA, 1996). To protect the privacy of the human subjects, HIPAA dictates differing levels of access to the information found in the EHRs based on the roles that researchers play in a particular study; these levels vary from full access (including protected health information) to very limited (i.e., public) access. In the case of public access, the data must be de-identified based on criteria elucidated in the HIPAA legislation, and some of these criteria are stated in a general fashion to reflect the fluid nature of modern science. Due to these ambiguities, the complex measures that are often necessary to de-identify protected health information, and the risk of litigation and lost reputation, scientists rarely share their de-identified annotated data beyond their current study.

Unfortunately, this lack of sharing negatively impacts the reuse of experimental data beyond its current context, and in turn, this lack of reuse can adversely affect the translational impact of basic life sciences. In contrast to this constricting approach to the management of clinical annotations is the move in computing toward the “Cloud” wherein data are stored for easy retrieval and sharing. In our current study, we are surveying life scientists to ascertain their perceptions of a cloud-based approach to the management of their annotated data.

Health Insurance Portability and Accountability Act of 1996 (HIPAA). (1996). Retrieved July 10, 2009 from http://www.cms.hhs.gov/HIPAAGenInfo/Downloads/HIPAALaw.pdf.

Undergraduate Research Project:

Poster .pdf

The undergraduate student will conduct a comprehensive literature review and perform an analysis of the large data repositories frequently used in the life sciences. There are several large repositories. The Susan B. Komen Virtual Tissue Bank is one example. The Komen Virutal Tissue Bank is the only repository in the world for normal breast tissue and matched serum, plasma and DNA. By studying normal tissue, we accelerate research for the causes and prevention of breast cancer. To more deeply understand the evolution of the disease, it is necessary to compare abnormal, cancerous tissue against normal, healthy tissue. Student research projects include: - Characterization of how these large data repositories handle the sensitivity and privacy of the information they store. - Best practices for designing proteomic, genomic and metabalomic databases to enable data sharing and reuse while managing privacy and security requirements.

Secure, Composable, & Scalable Framework for Trusted Collaborative Computing

Collaborative Computing (CC) is a critical application domain within the Internet environment. A few examples of CC are multi-party computation, collaborative defense, tele-medicine and collaborative decision making. Participants in CC demand confidentiality, privacy, integrity, and controlled sharing of sensitive information. Also, CC environments involve many entities, which are dynamic, heterogeneous, distributed, and can be hostile. Currently, CC uses the Internet as the underlying infrastructure, which by design is not secure and suffers from incessant attacks ranging from eavesdropping to vulnerability exploitation. Hence, it is imperative for the success of CC to require a reliable and secure framework built on top of the Internet to remedy some of its limitations. CC, based on such an underlying framework, can be termed as Trusted Collaborative Computing (TCC). Thus, the long term objective of this research is to develop a framework that will enable TCC. This framework consists of: (1) (group-oriented) secure and anonymous communication, (2) finely-controlled data sharing and (3) secure, composable and scalable integration. The framework will effectively address the underlying challenges of secure communication and guaranteed access, anonymity, composability, interoperability, and scalability.

The core technique in the proposed TCC framework is Access Control Polynomial (ACP) which was just presented at and published in the proceedings of INFOCOM’08, one of the highest international conferences in the networking and security field. The short term yet intensive summer work is to implement and evaluate such an innovative ACP mechanism and related security modules. This work will significantly help the accomplishment of the long term objective and secure the application for external funding.

Assessing the Relationship between Hacking and Various Personality Traits

Surveys indicate that there is an increasing risk of computer intrusion, computer crime and attacks on personal and business information. Computer criminality is a serious problem that affects individuals, businesses, and our nation’s security. The current study has four specific aims. First, we explore whether deviant computer behavior is part of a larger syndrome of deviance. Much research has shown that non-computer-related delinquent/criminal activities, substance use, and early/risky sexual behavior are typically seen in the same individuals and can be considered part of a larger syndrome of deviance. Second, we examine whether the personality profiles of those committing deviant computer behaviors are similar to the profiles obtained from those who engage in more general deviance. Several meta-analyses have demonstrated that interpersonal antagonism (i.e., lack of empathy, oppositionality, grandiosity, and selfishness) and problems with impulse control are the most consistent personality correlates of a variety of antisocial and deviant behavior. Our third aim is to examine a potentially unique correlate of deviant computer behavior—Asperger’s syndrome. Within the past decade, questions are emerging regarding the possibility of there being a link between computer criminality and a disorder known as Asperger syndrome. Finally, our fourth objective is to further validate certain psychometric instruments for use with the “hacker” sub-culture. This project is currently in the preliminary stages of data collection.

Categorization of Various Types of Online Pornography Users & Personality Assessments

Do individuals view, download, and share various types of Internet pornography and are different personality characteristics related to a person’s pornography preference? This research project gathered data from online respondents regarding their use of adult, animal, and child pornography as well as various personality characteristics. Data has been collected and is currently being analyzed.

A Context-Driven Security Framework for Distributed Healthcare Domain

The primary objective of the Personal Health Record (PHR) initiative is to empower users (patients) to control their own private medical information not only in terms of management and access but also allowing the sharing of their information with others in a private, secure, and confidential environment. Generally, disclosure of personal information depends on the circumstances of access including the privacy concerns of the individual patient. In particular, for using EHR/PHR technology the overriding public concern is ensuring security and privacy of their health care information, which is a serious technological challenge for the PHR technology developer. Following are the two key barriers to a wider use of PHR.

(a) Inability of a patient to compose consistent and context-aware disclosure policies for his/her collection of Electronic Health Records (HER). These records can be maintained by various heterogeneous health care and government enterprises. The challenge is to provide an intelligent user-friendly and patient-centered environment empowering the user to control access privileges relevant to various contexts.

(b) Secure and privacy-aware interoperability and data sharing among independent healthcare enterprises. The challenge is how to ensure secure sharing of data among multiple health-care enterprises, with potentially diverse security policies and guarantee privacy-preserving data integration among such enterprises.

The objective of this project is to develop a healthcare prototype on NIST’s Policy Machine (PM) for exhibiting our newly developed context-driven policy framework. For the demo, a PHR multimedia database is being implemented which consists of text, images, audio and video data whereby fine-grained access to individual multimedia objects will be implemented based on the roles across multiple healthcare domains.

Human Factors in Online Security and Privacy

This research focuses on human aspects of online security and privacy assurance. With respect to online security, we have performed task analyses of the procedures required to use different types of authentication methods (e.g., passwords, biometrics, tokens, smart cards) and determined the costs and benefits of the alternative methods. Although passwords are the weakest of the methods, they are the most pervasive and widely accepted form of authentication for many systems. Thus, we have performed experiments designed to identify techniques for improving both the security and memorability of passwords. With respect to privacy assurance, we have performed analyses on Web privacy policies to determine organizations’ privacy and security goals. We also conducted usability tests examining users’ comprehension of privacy policies, factors that influence users’ trust in an organization, and users’ ability to configure privacy agents to check machine-readable policies for an organization’s adherence to specific privacy practices. Because the methods for ensuring security and privacy involve human users, our goal is to improve the interaction between humans and the technical devices and interfaces employed in security- and privacy-related tasks.

Mining Transaction Streams to Infer Semantic Relations

This project focuses on understanding and exploiting information in large-scale, dynamic relational networks. In a growing number of relational domains, the data record temporal sequences of interactions among entities. For example, in social networking sites such as facebook.com, members continuously visit other members’ pages, accessing content and posting comments. These use patterns could be utilized to infer the nature and strength of relationships among members, which may then in turn be exploited to improve personalization efforts, marketing strategies, and system design. Our work will produce the first available data mining tools that can simultaneously exploit both the temporal and relational aspects in streams of transactions. We are developing automated methods to infer high-level semantic relationships (e.g., friend, colleague) among entities from dynamic patterns of low-level transactions (e.g., file transfers, phone calls). We will use these semantic relations to identify and exploit the dependencies among entities, thereby improving the accuracy of predictive models. For example, malfeasance is usually a social phenomenon, communicated and encouraged by the presence of other individuals who also wish to engage in misconduct. Thus, if we know one person is involved in fraudulent activity, then his close contacts have increased likelihood of being engaged in misconduct as well.

Leveraging Social Trust in Building Secure Email Networks

The openness of email systems allows any user to send unsolicited email (or spam). Most current solutions to filtering email spam center on content-based filtering or domain blacklisting approaches, both of which are inaccurate and slow to adapt to the changing face of spam. We propose STAMP, short for Solicitation Token Authenticated Mail Protocol, as a server-side solution to filter unsolicited mail from ever reaching the end-user’s inbox.

STAMP employs distributed access control leveraging the existence of both direct and indirect trust relationships between users. To reduce manual configuration of these trust relationships, STAMP also incorporates scoped broadcast flooding among mail servers to automatically discover paths of trust links through social networks. STAMP is designed for users in high-stakes settings such as enterprises, sales personnel, and defense networks, where it is very important to ensure a spam-free mailbox as it seriously affects their productivity.

Predictive Blacklisting for Detecting Phishing Attacks

Phishing has been an efficient and easy tool for trickery and deception on the Internet costing billions of dollars to users every year. While solutions such as blacklisting in Internet browsers have been effective to some degree, the reliance on exact match of a URL with the blacklist entries makes it easy for attackers to evade and reduce the ability to detect and stop such attacks. In this paper, we observe that several simple modifications such as changing TLDs in existing blacklist entries can result in finding new sources of maliciousness. In particular, we propose five heuristics based on our observations with real blacklists and extensively evaluate them using real-time feeds blacklist feeds. Our approach led to the discovery of more than 18,000 new malign URLs validated through content matching from a set of 6,000 original blacklist entries. Based on the success of these heuristics, we essentially translate the mechanisms used in the heuristics into an approximate matching algorithm that increases the resiliency of blacklists significantly. We propose a system, \system, that breaks up a URL into multiple components that are then matched against individual regular expressions or hash maps formed out of the original set of blacklist entries to assign a normalized score, used to determine whether the entry is malign. We evaluate the efficacy of system using real blacklist URLs from PhishTank and SpamScatter and benign URLs from Yahoo and DMOZ. We find that \system has a false positive rate of 3% rate with a false negative rate of 5%. The false positive rate can be arbitrarily lowered by adjusting the threshold.

Information Security Ethics and Social Issues:  A Moral Development/Constructivist Approach

This project considers ethical and social issues in information assurance and security including privacy, ownership, access and safety, liability and reliability. These issues will be examined using a system-of-systems, inter-disciplinary perspective to include technology, economics, policy, and culture. The system-of-systems approach considers the integration of individual, usually trans-domain, systems into a network of systems that ultimately contribute to social infrastructure. This approach does not necessarily advocate particular tools, methods and practices; rather, the focus promotes a new way of thinking where the interactions among technology, policy, and economics are carefully considered when conceiving grand challenges. This project will consider the emergent and complex co-constitutive, and therefore necessarily uncertain, relationships among technology, policy, economics, and culture with attention to the interacting roles of each in the context of secure information systems design, implementation and use.

Cybertrust is a priority in the information age. Future advances in computing promise substantial benefits for individuals and society; but trust in computing and communications is necessary in order for such benefits to be realized. Cybertrust depends upon software and hardware technologies that people can justifiably rely upon. However, repeatedly research has shown that technical feasibility alone is not sufficient for widespread adoption of a technological innovation. Large-scale adoption of technology is shaped by user acceptance, economics, policy, and organizational practices. The scholarly value of this project is that it will address information assurance and security ethics from these co-constitutive perspectives and by doing so, will advance the conversation on cybertrust in a meaningful way.

Footer

Intranet Site