Invention Summary:
Data scientists take an enormous mass of messy data points (unstructured and structured) and use math, statistics, and programming to clean, massage, and organize them. Then they apply analytic expertise (e.g. industry knowledge, contextual understanding, skepticism of existing assumptions) to uncover hidden solutions to business challenges. This process of uncovering hidden patterns visualized as plots take a tremendous amount of programming, which is often quite repetitive and tedious. As a result. companies need to hire highly paid, in short supply, data scientists who will need to spend a lot of time performing these tedious tasks.
Rutgers researchers have developed a technology solution referred to as “Boundless Analytics” that addresses this situation. The invention provides a unique approach to data pre-discovery not covered by any data and analytics products on the market today. The patented technology finds data that is 'interesting', a massive challenge when reviewing data over large sets. What is worth a deeper look? The new approach to divide data into slices and measure deviation from baseline will highlight the 'most interesting' data by merging an intuitive and simple interface anchored in a rigorous scientific approach. The invention can also be critical for explanatory AI to reverse engineer models built by deep learning methods. Slices of data generated by Boundless Analytics could provide explanations of decisions made by the deep learner. A demo/prototype is available and inventors are available to talk to interested parties with a demo using their datasets.
Market Applications:
- In many industries, from retail to government to biotech especially ones where there is a lot of data that requires analysis by data scientists.
- Explanatory Artificial Intelligence (Explain a decision of a black box deep learner AI).
Advantages:
- Helps in finding “needles in haystack” – though background process which discovers slices of data and relationships which are worth analyzing further.
- Saves data scientist valuable time by taking away the tedious part of data analysis and automating it.
Intellectual Property & Development Status:
Issued US patent (11,580,118). Available for licensing and/or research collaboration. For any business development and other collaborative partnerships contact marketingbd@research.rutgers.edu.
About the Inventor, Tomasz Imieliński:
Dr. Imieliński is a Professor of Computer Science at Rutgers University. He is the author of a joint paper with Agrawal and Swami, entitled 'Mining Association Rules Between Sets of Items in Large Databases’ that initiated the association rule mining research area and is one of the most cited publications (has over 18,000 citations) in computer science. From 1996 until 2003 Dr. Imieliński served as chairman of Computer Science Department at Rutgers Was a co-founder of Connotate Technologies –and has also held multiple senior-level positions at Ask.com and IAC/Pronto.
