AI which can understand, be understood, and understand itself

AI we can trust with what matters most

Learn More
image description
image description

About Us

What is b0hr.ai? b0hr.ai is a division of doc.ai which works on natural language understanding and generation in the field of healthcare and medical research. We’re building an AI center of excellence for language research and development in text, conversation, and voice.

Every valuable human being must be a radical and a rebel, for what they must aim at is to make things better than they are.
image description
image description
image description
The opposite of a profound truth may well be another profound truth. Niels Bohr
image description

The complementary truths about knowledge are that it is both structured and emergent. It is structured so we can share and improve it as the genome of civilization. And it is emergent because it is created, discovered, refined, and transformed.

image description
image description

Current AI systems are difficult to trust because they are:

Opaque: neither domain experts or technologists can explain why a given AI system yields correct or incorrect results; this is why the usual approach to error or bias is to just “add more data”

Context-specific: modern AI systems often yield unexpected or inexplicable results when given inputs outside of their training context, which may not be completely understood; (bohr quote) An expert is a person who has made all the mistakes that can be made but in a very narrow field.

Unconscious: these systems have few if any way of recognizing their errors, qualifying their recommendations, or learning from individual errors or advice.

image description

At b0hr.ai, we build, evaluate, debug, and improve new kinds of AI models for healthcare. Our aim is to take AI beyond the Big Black Box. b0hr.ai’s research is currently focused on:

  • image description
    Dynamic Memory

    Replaces static numeric models with indexed memory structures.

  • image description
    Testing & Evaluation

    Analyzing and characterizing the quality and performance of AI models

  • image description
    Multi-model Architectures

    Creating complex decision and learning systems combining multiple AI models

  • image description
    Explainable Continuous Learning

    Learning from individual examples or counter-examples

  • image description
    Practical Consciousness

    AI systems architectures which include awareness of limitations and context

  • We’re building a center of excellence for language, text, conversation, and voice.

image description

Leadership Team

  • image description
    Co-founder, CEO
    Walter De Brouwer, PhD
  • image description
    CTO
    Akshay Sharma
image description

Careers

Join us

We’re looking for talented experts in NLP, machine learning, and narrative intelligence.

Benefits package includes premium medical insurance, 401(k), vacation time, catered daily lunches, and commuter benefits.

Join Us
  • Contact Us

    Join us or collaborate on our research

    Email Us
  • Follow Us

    See more on our technical blog

    Learn More
image description
image description
image description

Connect with us at
upcoming conferences:

  • NeurIPS
    Vancouver Convention Center December 8-14, 2019
  • Pycon2020
    Pittsburgh, Pennsylvania April 15-23, 2020