Beyond the Keyword: Creative Search and Query Expansions based on DBpedia
Does the notion of keyword search and relevance that powers search engines really meet the needs of all searchers? What alternative advanced search features can be implemented to meet fuzzier search needs? This talk discusses these questions by demonstrating the potential of a new search tool called "Brainforks" in everyday search, scholarly search and book search. By building an interactive graph of query expansions, the user can take part in a search that is both "exploratory" and "creative". "Exploratory" search removes the assumption that the searcher has a clearly defined search need and recreates the library browsing experience in your browser. "Creative" or "serendipitous" search allows the searcher to find solutions from other fields that are functionally analogous, so ideas can be shared across disciplines. Coded in Python, Brainforks makes use of many semantic tools and natural language processing libraries such as NLTK and data from Dbpedia, Wordnet, Faroo and Google Ngrams.
