Assessment of End-User Needs Associated with Predictive Innovation Analytics


Through participation in the National Science Foundation Site I-Corps program, Marcia Price pivoted her initial algorithmic objectives to those that better provide predictive innovation analytics that end-users and market participants will value. In addition, the author has better defined the field of predictive innovation analytics

Download paper here.

Learning Concept Embeddings for Efficient Bag-of-Concepts Densification


This paper proposes two neural embedding models in order to learn continuous concept vectors. Once learned, an efficient vector aggregation method to generate fully dense bag-of-concepts representations is proposed. Empirical results on a benchmark dataset for measuring entity semantic relatedness show superior performance over other concept embedding models.

Download paper here.

Mined Semantic Analysis: A New Concept Space Model for Semantic Representation of Textual Data


Mined Semantic Analysis (MSA) represents textual structures as a Bag of Concepts (BoC) where concepts are derived from encyclopedic corpora. MSA uncovers implicit relations between concepts by mining for their associations. Empirical results show competitive performance of MSA compared to prior state-of-the-art methods. 

Download paper here.

Analysis of the USPTO’s Statistical Mapping Between the CPC System and USPC System.


This study concludes that researchers need to be careful when using CPC codes to define technology areas and to compare technology trends over time periods that include the USPTO’s transition from the USPC system to the CPC system. Specifically, if their period of study includes 2013 to 2015, aggregated with any time period before or after those years, patent codes should likely not be used to define the technology area under study. 

Download paper here.

Exploring Emerging Technologies Using Patent Data and Patent Classification


This paper reports preliminary results on discovering emerging new technologies. One technique uses the corpus of US patents and the ontology implicit in the patent examiner's classification manual. It also explores using topic modeling and interactive visualization techniques to find emerging technology trends and validates such discoveries by interacting with patents metadata and text.

 Download paper here.

Click the button below to see how we have leveraged our published research into solutions.

See our solutions