AI Research
AI First
We continuously invest in our research and development. It is part of our background and continues to be key to our story. We believe a great product sells itself.
​
We continuously evaluate how to improve our clients' ability to easily curate their data by solving really hard problems in engineering and AI. We are proud to drive new innovations in data curation and data management.
R&D Partner:
Multi-type Entity Resolution: The answer to automated knowledge graph building
Typically, entity resolution platforms concentrate on matching and merging of individuals or companies by utilising attributes from other real-world objects like addresses or phones. This process involves using advanced parsers and cleansers or generating various keys from the same underlying information.
At Knights, we approach each entity as a unique resolution problem, enabling us to customise models based on your data concepts. This allows a new level of accuracy in our framework, but also makes it super resilient to new concepts it's never tried before.
​
Our framework supports matching and merging events in the news, product mentions, addresses, films, or anything else you require. Lending itself to new use cases, and specialising in disambiguating information in unstructured data like images, video, or text.
Focus
Concept Resolution
Match and Merge concepts across data and systems.
Graph Neural Networks
Leverage graph structures for AI resolution.
Program Synthesis
Generate programs with machine learning and AI.
Hugot
Our Open Source Library
Hugot enables us to run machine learning models in GoLang (Go) at scale, particularly transformer models (think LLMs!). It is designed to make it easy to integrate transformers and LLM capabilities within Go applications. Read more on our blog here!
