2000 character limit reached
Knowledge Fusion via Embeddings from Text, Knowledge Graphs, and Images
Published 20 Apr 2017 in cs.AI, cs.LG, and stat.ML | (1704.06084v1)
Abstract: We present a baseline approach for cross-modal knowledge fusion. Different basic fusion methods are evaluated on existing embedding approaches to show the potential of joining knowledge about certain concepts across modalities in a fused concept representation.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.