Massimiliano Albanese
I am now at George Mason University. You will be redirected to my GMU homepage in a few seconds.
I am a Faculty Research Assistant in the University of Maryland Institute for Advanced Computer Studies (UMIACS), and a member of the Laboratory for Computational Cultural Dynamics (LCCD), working with Prof. V.S. Subrahmanian. I received my Ph.D. in Computer Science and Engineering from the University of Naples “Federico II”, where I served as a Graduate Research and Teaching Assistant before joining the University of Maryland in 2006.
Curriculum Vitae RESEARCH
My research interests are in the areas of Modeling and Recognition of Human Activity, Scalable Detection of Cyber Attacks, Multimedia Databases, and Information Extraction. The primary objective of my research across these different areas is to find efficient solutions to the problem of reducing massive collections of raw data to a manageable amount of actionable intelligence.
Countless applications require the ability to mine and summarize information from large data collections. In my Ph.D. thesis, I showed how the same principles can be applied to very diverse media types – text and video – in order to extract information from unstructured data and then present the information to the users in a compact yet effective way, through short stories and video summaries respectively.
Additionally, with today's extensive availability of real-time digital data, many applications, ranging from video surveillance to cyber security, require the ability to monitor a large body of streaming data for the occurrence of activities of interest. I have done extensive work on scalable real-time detection of human activities, in both video and transactional logs, and I am currently leveraging my experience in this field to address scalability issues in the Cyber Security domain.
CURRENT PROJECTS
CSA | When security incidents occur, the top three questions security administrators would ask are: What has happened? Why did it happen? What should I do? Answers to the first two questions form the core of Cyber Situation Awareness. My main contribution to this project consists in providing the capability to answer the first question efficiently. Indeed, the question becomes: What is happening? Attackers can exploit vulnerabilities to incrementally penetrate a network and compromise critical systems. The enormous amount of raw security data involved in the process and the complex interdependencies among vulnerabilities make manual analysis extremely labor-intensive and error-prone. To address this important problem, I proposed an automated framework to manage very large attack graphs and analyze high volumes of incoming alerts to detect occurrences of known attack patterns in real-time. |
| ADETECT | Activity Detection Numerous applications need to model human activities and continuously monitor a body of data for the occurrence of such activities. Data to be monitored might include video streams from surveillance cameras, and logs generated by web applications, transaction processing systems, and intrusion detection systems. Applications might require activity detection to be performed either in real-time, as data is being received, or offline, after a body of data has been acquired. Activities tend to be high-level and can often be executed in many different ways. My research is aimed at developing techniques and algorithms to formally describe what the activities of interest are and identify occurrences of them from a body of data, both online and offline. |
RECENT NEWS
- Manuscript “Scalable Analysis of Attack Scenarios” by M. Albanese, S. Jajodia, A. Pugliese, and V.S. Subrahmanian, accepted for publication at the 16th European Symposium on Research in Computer Security (ESORICS 2011), Leuven, Belgium, September 12-14, 2011.
- Manuscript “Finding Unexplained Activities in Video”, by M. Albanese, C. Molinaro, F. Persia, A. Picariello, and V.S. Subrahmanian, accepted for publication at the 22nd International Joint Conference on Artificial Intelligence (IJCAI 2011), Barcelona, Spain, July 16-22, 2011.
- ATTUNE project on analyzing the use of social networks by terrorist groups selected as one of the finalists for the University of Maryland “2010 Inventions of the Year”.
PAST PROJECTS
| T-REX & ACE | The RDF Extractor & Automatic Coding Engine T-REX is a generic, domain-independent Information Extraction system. Differently from many traditional approaches, T-REX does not rely on domain-specific knowledge or site-specific features, but rather takes as input an RDF schema describing the information to be extracted, along with a training corpus of relevant sentences annotated with the semantic roles of their constituents. The Automatic Coding Engine augments T-REX extraction capabilities with a Prolog-based logic layer which enables the type of reasoning needed to answer more complex questions, and addresses the needs of political scientists and antropologists monitoring terrorist organizations and ethnic groups aroud the world. |
TEACHING
“Give a man a fish; you have fed him for today. Teach a man to fish; and you have fed him for a lifetime” – Author unknown
Fall 2010 | |
Spring 2009 | Information Extraction and Integration (HONR229C) |
| | |
