Skip to main content

Machine Learning AI

Machine Learning & Neural Nets are being applied in many of Oceanit's breakthrough projects.

Artificial Intelligence (AI) and statistical Machine Learning (ML) are making great advancements worldwide by integrating advanced algorithms, high performance computing, and big data to infer the answers to critical strategic, societal, and scientific problems.

Oceanit is actively applying a range of ‘state of the art’ techniques in ML to solve outstanding problems in industries such as shipping & logistics, energy, infrastructure, agriculture, bioinformatics, optics, and many more. Some examples of our Machine Learning AI work include:

  • Ocular Abnormality classification system – ‘RetinaView’ for ophthalmologists combines data and assists diagnoses from disparate ocular tools/systems.
  • Aerospace Sustainment - Analyzing disparate data to predict and diagnose failures in aircraft and rotorcraft systems and hardware, and quantify our uncertainty in those predictions.
  • Material Properties - Estimation system for material make-up and characteristics using scattered/transmitted light interacting with the material.
  • Rapid Damage Assessment – Identification and classification system for damaged infrastructure such as power transmission lines or cell towers using AI-based computer vision techniques.
  • Directional Drilling – Nicknamed, ‘Deep Thought’, Oceanit’s AI augments a direction drill operators decision making to make the drilling process safer, more efficient, and reduce environmental impact.
  • Pipeline state of health prediction – hydrate adhesion and deposition prediction to predict blockages before the affected pipe reaches critical levels or blockage and failure.

Oceanit tailors its AI/ML approach to suit the diverse projects and products that we are working on. In applications such as atmospheric modeling, Oceanit has leveraged the predictive power of deep learning techniques to infer latent physical parameters from observations. In other settings, more robust ensemble Machine Learning methods were used to identify the minimal number of sensors an optical system needed to make accurate predictions.

Oceanit is also developing uses for artificial intelligence microchip technology that can be trained “in the field,” without the support of data scientists or AI/ML specialists.

In 2018, Oceanit’s ‘Harvest Vision’ team utilized this AI chip technology to train a computer vision device that can identify ripe and un-ripe crops for the agriculture industry. Along with Kamehameha Schools and Kauai Coffee, the team won Hawaii’s for agriculture-tech hackathon, AGathon. Read more about Harvest Vision here and watch our video below.

While statistical machine learning is the state of the art in the current-generation of AI, Oceanit is actively developing the next-generation AI technology known as “anthronoetic artificial intelligence,” that replicates human-style cognition.

The implementation of this natural language processing, linguistic-centric approach is referred to as the Noetic Mathematical Engine (NoME). Unlike existing methods that generate predictions, NoME generates explanations, which will enable a user to understand both what the NoME output was and how NoME arrived at that output. Read more about NoME here.