2018 IEEE 4th International Conference on Collaboration and Internet Computing (CIC) (2018)
Philadelphia, PA, USA
Oct 18, 2018 to Oct 20, 2018
In order to help Earth scientists share knowledge and expedite scientific exploration, NASA is developing NASA Earth Science Enterprise (ESE) whose underlying basis is NASA Science Knowledge Graph (SKG). This paper focuses on the information model design of the SKG, where entity-and relationship-encapsulated features are extracted as first-class citizens in SKG. The rationale is that typological structural analysis can thus be exploited for runtime knowledge extraction, discovery, and prediction, in addition to semantic analysis. Based on the information model, a knowledge discovery technique is equipped to answer queries and provide personalized recommendation based upon the SKG. Deep learning techniques, i.e., Stacked Denoising Auto-Encoders (SDAE) and Translating Embeddings (TransE), are applied to detect structural similarity and explore paths at runtime, respectively, supporting higher scalability and performance. Experiments on a science-oriented testbed serves as a proof of concept and demonstrates the feasibility and effectiveness of the techniques.
data analysis, data mining, geophysics computing, learning (artificial intelligence), query processing, recommender systems
J. Zhang et al., "Facilitating Data-Centric Recommendation in Knowledge Graph," 2018 IEEE 4th International Conference on Collaboration and Internet Computing (CIC), Philadelphia, PA, USA, 2018, pp. 207-216.