Marinka Zitnik

Together with Blaz Zupan we recently published a paper on our new data fusion method in IEEE Transactions on Pattern Analysis and Machine Intelligence.

For most problems in science and engineering we can obtain data sets that describe the observed system from various perspectives and record the behavior of its individual components. Heterogeneous data sets can be collectively mined by data fusion. Fusion can focus on a specific target relation and exploit directly associated data together with contextual data and data about system’s constraints. In the paper we describe a data fusion approach with penalized matrix tri-factorization, called data fusion by matrix factorization (DFMF), that simultaneously factorizes many data matrices to reveal hidden associations. The approach can directly consider any data that can be expressed in a matrix, including those from feature-based representations, ontologies, associations and networks. We demonstrate the utility of DFMF for gene function prediction task with eleven different data sources and for prediction of pharmacologic actions by fusing six data sources. Our data fusion algorithm compares favorably to alternative data integration approaches and achieves higher accuracy than can be obtained from any single data source alone.

Short preprint is available at Arxiv:1307.0803. Full paper is online at IEEE.

New issue of ACM XRDS is here! The focus of this issue is on techniques for natural language processing in the broader sense. You will find interesting stories about how to detect influencers in social media discussions, how to successfully transition from academia to entrepreneurship, and read about the hurdles and opportunities in research of ancient written languages.

My department contributed a short column on exploring news from The New York Times. Techniques of information extraction and natural language processing allow us to search for news articles and to analyze dynamics of published content. News organizations, such as The New York Times, provide programmatic access to their articles to retrieve headlines, abstracts, and links to published multimedia. In the column we use The New York Times Article Search API to demonstrate how to construct search queries that retrieve documents from various news sections and time periods. We also explore the pulse of climate change over the years using data extracted from published news articles.

Recently, I have participated as young researcher in computer science at Heidelberg Laureate Forum. I encourage the reader to check recordings of some of the talks, which are available at official HLF website. If limited by time I recommend at least one of the following talks by Michael Atiyah, Manuel Blum, Wendelin Werner, Vint Cerf, Leslie Lamport, Manjul Bhargava, Daniel Spielman, Efin Zelmanov or John Hopcroft. They are engaging, full of useful tips and strategies, and should be accessible to an interested listener.

Many CS & Math bloggers followed the event, their comments and discussions about laureates' talks can be found at HLF Blogs. Among others, our poster has been highlighted by John D. Cook. Overall, HLF has been an awesome experience for me with many opportunities to network with Turing, Fields, Abel and Nevanlinna laureates and meet other young researchers in computer science and mathematics from around the world.

I have been given an opportunity to join Google Global Planning Committee for Women in Computer Science in an effort to identify ways we can have the greatest impact and reach more women in tech. As member of this committee I will partner with Google to build the community and direct outreach activities for women in computer science. To kick things off, we will have our global meeting at the Grace Hopper Conference in Phoenix, AZ, USA. I am excited to be part of this great program to promote women to excel in computer science and information technology.

Stay tuned, there will be many possibilities to engage with fellow technologists!