Bisociative Knowledge Discovery for Cross-domain Literature Mining

Bisociative Literature-Based Discovery: Lessons Learned and New Word Embedding Approach

“…"Literature-based discovery" introduces literature-based discovery (LBD), which is the main topic of this research. "Bisociative knowledge discovery" presents the area of computational creativity named bisociative knowledge discovery and the connection between bisociative knowledge discovery and LBD, as published in our past research [21,31]. Finally, "Embeddings" briefly introduces embeddings, the contemporary representation learning technology resulting from recent research in neural networks, which is the enabler for the proposed embedding-based bisociative LBD methodology introduced in "Towards creative embeddings-based bisociative LBD".…”

“…In conclusion, let us summarize the previously published [21,31] relationship between bisociative knowledge discovery and Swanson's ABC model for literaturebased discovery, where the particular focus of interest is the relationship between Koestler's bisociative link discovery framework and Weeber's closed discovery framework, as summarized in Table 1. Similar to a bisociation, which is according to Koestler a result of processes of mind when making new associations between concepts S and T from usually separated contexts (illustrated in Fig.…”

“…The paper is structured as follows. "Background and related work" presents the related work in literature-based discovery (LBD) and bisociative knowledge discovery, including the previously published relationship between the two [21,31]. It presents also the related work in representation learning using the embedding technology.…”