Discovering Research Ideas Using Semantic Vectors and Machine Learning

Explore how semantic vectors and machine learning can be used to discover research ideas in this 49-minute conference talk from GOTO Copenhagen 2016. Learn about UNSILO's mission to enrich scientific content and improve discoverability across domains. Dive into key challenges in full-text search, keyword usage, and ontologies. Discover UNSILO's innovative approaches, including exhaustive concept extraction, complete semantic mapping, and phrase extraction. Understand the application of word embeddings, Word2Vec, and ontology-augmented vector spaces. Explore how synsets are built using vector cosine similarity and how human-readable fingerprints are created. See demonstrations of UNSILO's discovery widgets for easier content exploration. Gain insights into future directions for research idea discovery and content enrichment in scientific publishing.

Intro
UNSILO Mission
UNSILO Core Technology
Key Challenges
Full Text Search
Using Keywords and Ontologies
UNSILO Exhaustive Concept Extraction
UNSILO Complete Semantic Mapping
Phrase Extraction
Word Embeddings and Word2Vec
Ontology Augmented Vector-space
Synsets built from Vector Cosine Similarity
Human readable Fingerprints
UNSILO Discovery Widgets
Easier Content Exploration
Future Directions