In search of solutions, publishers are increasingly considering how Artificial Intelligence technology can streamline processes and answer some of the questions they are challenged with when it comes to managing and marketing an ever-growing repository of content. Specifically, Natural Language Processing (NLP) and Machine Learning have aided content enrichment and taxonomy improvements, leading publishers to improved workflow efficiencies, thus saving significant personnel costs. NLP also provides an unbiased way to get a more granular understanding of publishers’ content corpus. However, these solutions still fall a bit short. Put simply, NLP hunts for words to generate a form or taxonomy for publishers using keywords. However, classical NLP approaches cannot easily disambiguate words having multiple meanings, especially at the single-word level.
The next generation of AI technology is going one step further by mimicking how the human mind works and applying those cognitive functions to get a better handle on vast data sets. The human mind is an ideal model because it is used to consuming large amounts of information from various sources. It automatically creates neural pathways connecting relationships with things it already knows, and creates new pathways as humans learn more things. Nestled in Silicon Valley, Yewno, has been developing AI technology that mimics the way the human mind works.
Their models leverage machine learning, computational linguistics, and graph theory to collectively accomplish two very important objectives:
- Identify and extract concepts from both structured and unstructured information
- Unearth significant knowledge via an inferential chain of connections between identified concepts
The major difference between Yewno’s technology and NLP lies in “semantic spaces.” Semantic spaces in the natural language domain aim to create representations of natural language that are capable of capturing meaning. The original motivation for semantic spaces stems from two core challenges of natural language: vocabulary mismatch and ambiguity of natural language. Rules-Based and Model-Based approaches like NLP which operate on a keyword level cannot take into account polysemy and synonymy. Polysemy means the coexistence of many possible meanings for a word or phrase such as jaguar or java, while synonymy refers to different terms with equivalent meanings, such as the way “valid” can mean authorized, legitimate, or licit. Yewno’s technology enables the application of semantic spaces, overcoming these limitations.
While these nuances seem highly technical, they become particularly important when reading large amounts of text, such as a publishers’ entire collection. To overcome difficulty with ambiguous terms, Yewno embeds the multitude of concepts into a semantic space whereby semantically related concepts are closely grouped together. Yewno hunts for concepts, not keywords, and identifies these as objects that carry a description and a significance.
Yewno and Silverchair partnered together in late 2017 to offer publishers access to this powerful technology. If you are a Silverchair publisher interested in increasing your discoverability and identifying hidden resources in your collection, you can simply opt-in to have your content become discoverable in the Yewno ecosystem.