Researchers Take Home SIAM SDM Best Research Paper Award
By Josh Quicksall
Work by ISR PhD student, Hemank Lamba, was recently recognized by the Society of Applied and Industrial Mathematics’ International Conference on Data Mining (SDM). Alongside Leman Akoglu (Heinz College), the pair’s paper “Learning On-The-Job to Re-rank Anomalies from Top-1 Feedback” was named Best Research Paper.
Sponsored by the SIAM Activity Group on Data Mining and Analytics and held in cooperation with the American Statistical Association, SDM has established itself as a leading conference in the field of data mining. Emphasizing principled methods with solid mathematical foundations, the event provides a venue for researchers who are addressing these problems to present their work in a peer-reviewed forum.
Lamba and Akoglu’s work examines the current practice in anomaly mining in which a human expert manually verifies anomalies ranked by an anomaly detector. Leveraging this approach, the team developed an always-on anomaly detector which learns from each verification decision made by the human expert. Called OJRank, the detector reduces the false positive rate as well as the overall expert effort. Together, the team’s work shows that this approach achieves statistically significant improvement on both detection precision and human effort over any currently available tool and technique.
Lamba is a past recipient of a Best Student Paper Award at the IEEE/ACM Conference on Advances in Social Network Analysis and Mining. You can learn more about his work by visiting his website.