Visible to the public Biblio

Filters: Keyword is Statistical Learning  [Clear All Filters]
2018-02-14
Hutton, W. J., Dang, Z., Cui, C..  2017.  Killing the password, part 1: An exploratory analysis of walking signatures. 2017 Computing Conference. :808–813.
For over 50 years, the password has been a frequently used, yet relatively ineffective security mechanism for user authentication. The ubiquitous smartphone is a compact suite of sensors, computation, and network connectivity that corporations are beginning to embrace under BYOD (bring your own device). In this paper, we hypothesize that each of us has a unique “walking signature” that a smartphone can recognize and use to provide passive, continuous authentication. This paper describes the exploratory data analysis of a small, cross-sectional, empirical study of users' walking signatures as observed by a smartphone. We then describe an identity management system that could use a walking signature as a means to passively and continuously authenticate a user and manage complex passwords to improve security.
2017-05-18
Nguyen, Anh Tuan, Hilton, Michael, Codoban, Mihai, Nguyen, Hoan Anh, Mast, Lily, Rademacher, Eli, Nguyen, Tien N., Dig, Danny.  2016.  API Code Recommendation Using Statistical Learning from Fine-grained Changes. Proceedings of the 2016 24th ACM SIGSOFT International Symposium on Foundations of Software Engineering. :511–522.

Learning and remembering how to use APIs is difficult. While code-completion tools can recommend API methods, browsing a long list of API method names and their documentation is tedious. Moreover, users can easily be overwhelmed with too much information. We present a novel API recommendation approach that taps into the predictive power of repetitive code changes to provide relevant API recommendations for developers. Our approach and tool, APIREC, is based on statistical learning from fine-grained code changes and from the context in which those changes were made. Our empirical evaluation shows that APIREC correctly recommends an API call in the first position 59% of the time, and it recommends the correct API call in the top five positions 77% of the time. This is a significant improvement over the state-of-the-art approaches by 30-160% for top-1 accuracy, and 10-30% for top-5 accuracy, respectively. Our result shows that APIREC performs well even with a one-time, minimal training dataset of 50 publicly available projects.