My Current Research Problem Statement

The Web has grown to be an essential part of our lives, people use the Web to stay up to date on current news, check weather forecasts, book flights, plan vacations, buy and sell goods, express opinions, check cooking recipes, medical diagnoses, book reviews, etc. The successes of web-based transactions and services have made the internet a target of crime.

In the real world, you only have to worry about the criminals who live in your city. But in the online world, you have to worry about criminals who could be on the other side of the planet. Because the Internet has no borders, crime control agencies are finding it extremely difficult to keep up with the rapid growth of online crime. They have limited resources and expertise to investigate online criminal activity.

Mikko Hypponen, Cyber Attacks in 19 Key Essays on How the Internet is Changing our Lives.

Specific Applications to Pharmaceutical Products Anticounterfeiting

Text Mining and Sentiment Analysis 

While the Web and social media platforms are being used by enforcement agencies for counterfeit advisories (information sharing and risk communications), predators and online criminals now use these platforms for sending hoax messages to deceive vulnerable people into purchasing counterfeit products.

All these underscores the need for law enforcement and anti-counterfeiting agencies to do better product counterfeit prediction and prevention. 

Current strategies for detecting, tracing and tracking of counterfeit pharmaceutical products are not suitable for spotting new and emerging attacks. However, they provide a great source of data for knowledge discovery.

Our Contributions

  • figuring out what consumers of pharmaceutical products and other related communities think about different brands. Paper, Presentation Slides (UKCI, Genova, Ph.D. Progression).
  • currently working on the technology aspect of the problem.

Network Analysis

Cybercrime is constantly evolving, it is also becoming clear that the criminals are learning from each other, and possibly forming various kinds of partnerships in the darknet. How then can we:

  • effectively use data to characterise suspicious contents and actors in the cyberspace?
  • identify and measure factors that sustain crime in the cyberspace?
  • understand and disrupt the structure, partnership and dynamics of cyber criminals?

Our Contributions

  • bipartite network modelling and analysing of nodes and community structures in order to infer relationships between actors and resources involved in crime. Paper, Presentation Slide.
  • currently working on the technology aspect of the problem.

Resources and Links

Matt Might

  • great resources on graduate school, programming, productivity, writing, presentation, academia, teaching, and job hunt.

Mirco Musolesi

ACM Online Career Resources for Graduating Students

  • great resources on career websites, professional networking, and salary survey sites.

Must-do’s for PhD students in Machine learning

  • Attend Machine Learning Summer School
  • Participate in an ML competition (e.g., Kaggle, KDD Cup, 1)
  • Periodically reproduce a fundamental ML results from first principles
  • Go deep in an application area
  • Gain exposure on a wide range of application areas

Computer Science Postdocs – Best Practices