We believe society’s most difficult problems — in critical fields such as security, healthcare, finance, law and business — will be solved by human creativity and resourcefulness together with smarter, faster machines.

Our challenge today isn’t too little information; it’s too much—too detailed for any team or human or rooms of machines to handle. Kyndi’s technology is based on groundbreaking new approaches to machine intelligence that combine insights from multiple areas of science.

This unique combination of science insights has allowed us to develop methods that map symbolic representations of knowledge into sophisticated numeric domains. When applied to reviewing text, Kyndi’s technology has the world’s richest depth of pattern matching and text review, allowing us to review vast amounts of textual information with a deeper clarity of understanding and reason.

The Secret is Kyndi’s Natural Language Understanding

Artificial intelligence makes a distinction between using techniques that manipulate symbols (called symbolic AI) and using techniques that rely on numeric methods.

Google’s and others’ recent successes with deep learning make use of numeric methods that require very large datasets and provide no explanation of outcomes.

Symbolic AI uses combinations of symbols to capture the semantics or meaning of a piece of text. Beyond being able to fully represent meaning as a logical structure, symbolic AI learns efficiently from small datasets and provides complete explanations for any outcome. Until Kyndi, symbolic methods have proven expensive to configure and expensive to run, even for moderate amounts of data.

Kyndi NLU uses a map between symbolic and numeric versions of knowledge

Kyndi has found a way to unify symbolic and numeric approaches into a single solution. By dovetailing these different expressions of knowledge, Kyndi avoids many of the pitfalls of working with symbolic knowledge and heightens natural language processing to the level of natural language understanding.

Natural language understanding takes machine intelligence out of the realm of processing information into a world where true knowledge can be obtained, allowing people to draw the best possible conclusions at a speed never before possible.

Like to Know More?