Constellation Labs publishes first academic paper on breakthrough technology
Constellation AI have published our first academic paper, written by our research and data science team: Constellation Labs. The paper, entitled Fine-Grained Named Entity Recognition using ELMo and Wikidata is the result of the Lab’s collaborative process to enhance the performance of naming and recognising entities within natural language processing.
Named Entity Recognition (NER) is one of the most fundamental elements of natural language processing; alongside sentiment analysis, intent prediction and topic detection. Within that, Fine-Grained Named Entity Recognition is a task used to detect and classify a large number of entity types, spanning diverse domains such as finance, healthcare and politics.
Most existing NER systems are only able to recognise around four things (location, people, organisations etc). We combined our unique natural language processing systems and deep learning techniques to form an entirely novel way to identify over a hundred entities. This means we can more accurately, more effectively and more expansively identify specific types within data. It’s the first time this process has ever been achieved in this way.
The Constellation Labs initiative is a team of data scientists, engineers and researchers within Constellation AI, who are working on breakthrough technology of extreme value to the global AI market and ecosystem. Their focus is on the academic over commercial application of their research and enquiry. It’s a collaborative process on ‘moon-shot’ longer-term projects, with medium-term focus on creating a unique ontological map of a human: processing text and extracting the information that is most useful for an individual.
By looking at which people, places and knowledge resonate the most with a person, our Labs can use that information to build fine-grain insight about them. The more we know, the more we can understand them. This insight, when applied to technology, can be fully utilised for the benefit of all.
In upcoming months, you can expect a Constellation Labs Master’s collaboration with Queen Mary University, conference lectures and further academic papers on subjects such as Ontological Mapping. This academic paper is the first step to Constellation Labs sharing our findings, code and methods for the benefit of wider audiences. Read the full academic paper here.
