[Talk Summary 8] Data-Driven Science of Science

Dr. Ying Ding from School of Informatics and Computing, Indiana University gave the talk "Data-Driven Science of Science" on November, 04, 2016 at School of Information Science, University of Pittsburgh.

In the talk. Dr. Ding presented an overview about Data Science and the current layers of bibliometrics, which are macro level in complex network, meso level in bibliometrics, and micro level in collaboration and team science. Currently, most of research work has been focusing on analyzing data from complex network and bibliometrics. Dr. Ding suggested that collaboration and team science, as a micro level, should be a new trend to have the attention of data scientists.

In addition, Dr. Ding concisely summarized her work related to Data Science which is beyond the bibliometrics. The three following are the main of her research:
  • Data-Driven Discovery: entity metrics, computational hypothesis generation, and digital innovation (e.g. machine reading)
  • Data-Driven Decision Making: understanding scientific career, understanding scientific collaboration, and understanding scientific success and innovation.
  • Data-Driven Knowledge Discovery
In details, she stated some potential, interesting ideas and techniques related to her work. Speaking of keywords, how does a keyword shift from one topic to another topic (or one domain to another domain). How can we increase the strength of relationship between two nodes in a graph in a far distance? Figuring out those kinds of relationship helps us to discover the knowledge which is difficult to be known. In her research, she tried to find patterns in the graph through extracting the same sub-structure of two entities, the semantic relationship, and the gene ontology.

At the end of the talk, Dr. Ding discussed a new potential topic in machine learning, lifelong machine learning. Lifelong learning which learns as humans retains the knowledge gained from the past learning and uses the knowledge to help future learning and problem solving.

135N Bellefield, IS building
University of Pittsburgh


Popular posts from this blog

Influences on Query Reformalution in Collaborative Web Search

[Talk Summary] Machine Learning and Privacy: Friends or Foes?

[Talk Summary 3] Personalized Recommendations using Knowledge Graphs