Data Mining and Machine Learning lab (DMML)
The Data Mining and Machine Learning Lab (DMML) — in the School of Computing and Augmented Intelligence at Arizona State University — is led by Professor Huan Liu. DMML develops computational methods for data mining, machine learning, and social computing; and designs efficient algorithms to enable effective problem-solving in text/web mining and feature selection, with a focus on real-world applications. DMML’s research projects (i) analyze high-dimensional data via feature selection and feature discretization; (ii) incorporate social computing, social media mining, group understanding, and interaction; (iii) integrate multiple data sources to overcome ambiguity and uncertainty, (iv) employ domain knowledge for effective mining and information integration, and (v) assist human experts to develop effective methods of ensemble learning; semi-supervised learning; and active learning with hierarchical classification, subspace clustering, and metadata. Recent investigations apply these research findings to social cyber-attacks (e.g., misinformation campaigns, bot detection), and improved management of personally identifiable information (PII). The DMML team includes 10-12 PhD students at any given time, who regularly receive scholarly awards and continue their careers at research universities, prominent companies, and industry start-ups. The group has produced over 324 conference papers, 137 journal articles, and 23 books including 1 textbook and 9 monographs. Their research works have generated worldwide impact and received more than 98,111 citations.
Motivated by our research on understanding the way users behave and interact on social media platforms and the lack of existing solutions which enable first responders to capture and use the information in social media for disaster response, we have built innovative social media tracking systems, such as TweetTracker, TweetXplorer, and ASU Coordination Tracker (ACT). These systems have helped first responders make informed decisions by incorporating the social signal produced during times of disaster. By bridging the gap between first responders and the information generated on social media we are working towards ASU's challenge of focusing information and technology to produce meaningful change. Our systems have been used by over 200 organizations, including Universities, government agencies, and NGOs. The work done on these tools has earned the lab the ASU President's Award for Innovation, the highest award for research at ASU. The systems developed as part of this project are exemplary of the innovation aspirations of ASU. Our tools are the first of their kind to aid first responders in making sense of the noisy and voluminous data generated on social media. In developing these tools, we have helped first responders to create an informed picture of the events of Hurricane Sandy, Typhoon Haiyan, and several others.
The DMML lab has produced first-class students, many of whom have been recognized for their achievements. Among the members are a Dean's Fellowship award winner, two ASU Faculty Emeriti Fellows, and two SMART scholars. Students have won awards ranging from the 2014 Outstanding Computer Science TA award, two US President's Volunteer Service Awards, and the ASU Graduate Education Dissertation Fellowship. Undergraduate students in the lab have been successful in winning awards pertaining to research, including prestigious awards such as the FURI award, and honorable mention for the CRA Outstanding Undergraduate Research Award. In the last five years, the DMML lab graduated 32 Ph.D. students, 12 of them became professors have gone on to be tenure-track faculty at research universities and other alumni of the lab are working at leading companies and start-ups in the area of data mining and predictive analytics (e.g., LinkedIn, Yahoo, Google, Microsoft, Amazon).
