CMU, Oxford Research to Explore AI and Advanced Analytics

• By Dian Schaffhauser
• 08/30/18

Two universities will be working with international digital media intelligence firm Meltwater on artificial intelligence research. Both Carnegie Mellon University and the University of Oxford have announced projects that use the company's AI platform. Fairhair.ai, as it's called, provides tools for building AI models and doing advanced analytics. It's used by Meltwater itself to identify trends for its clients. (The name of the platform was inspired by Harald Fairhair, the Viking who unified many of the kingdoms of Norway in the 9th century.)

Carnegie Mellon intends to use Fairhair.ai in a graduate AI course and as a resource for the university's data science and AI research community. Oxford is working with Meltwater on a number of research projects. In both cases, the faculty members behind those initiatives serve on Meltwater's scientific advisory board, which has helped to shape development of the platform they're now relying on.

Among those at Carnegie Mellon who are excited about access to Fairhair is Eric Nyberg, who directs the university's Master of Computational Data Science program and serves as a professor in the Language Technologies Institute. Nyberg led a team of students in working with IBM Research from 2007 to 2011 to tune Watson for the Jeopardy! Challenge. He said he plans to use the platform to develop rapid, cost-effective methods for producing specialized question-answering systems for specific information domains.

In January, Nyberg began working with Majd Sakr, a teaching professor in his institution's School of Computer Science, to launch an accelerated-cloud-for-AI project with support from Meltwater. That includes development of "realistic benchmark challenges for natural language processing tasks." For example, one focus will be on entity recognition — classifying objects with proper names as people, organizations or locations. Then the team will create an "open entity recognition challenge" and benchmark possible solutions using Fairhair resources.

This project "will crystallize important advancements in the engineering of cost-effective AI systems," Nyberg said in a statement. "The biggest challenge for current students is how to explore the large space of data, features and models available for developing a particular analytic in order to find an optimal or acceptable solution before they run out of time or computing resources." The framework developed will enable students "to explore this large solution space by providing a systematic approach that teaches cost-effective use of cloud resources to build AI systems."

Nyberg also expects to use Fairhair in his graduate course, "Design and Engineering of Intelligent Information Systems."

Oxford's use of Fairhair spans four research initiatives, according to the university:

• Value added data systems, a five-year research project to scale data science with tools for automating and reducing the expense of "data-wrangling";
• FakeNewsRank, which is working to create a "fake news rank" that will identify news articles by the likelihood that they're fake or not. Fairhair will be used to demonstrate the effectiveness of the developed scoring mechanism on real-world data;
• Realistic data models and query compilation for large-scale probabilistic databases, to be used when estimating the probability that a certain fact, document or source is accurate; and
• DeepReason.ai, a university spinout that uses logical reasoning and database theory in performing "fast and intelligent reasoning" on various data sources.

"These research projects are collectively focused on pushing the boundaries of knowledge graphs, from populating and scaling to processing and reasoning. Meltwater is already leveraging these techniques, together with recent advancements in machine learning, to further strengthen its Fairhair.ai platform," said Georg Gottlob, professor of informatics and a fellow at Oxford's St. John's College. "Together, we have a shared interest to explore what we believe is the future of AI — the combination of machine learning and logical reasoning — and significantly advance the greater data science ecosystem."