News

AI Start-Up's 'Continual Learning' Platform Removes Human Element from Algorithm Training

• By David Ramel
• 02/12/2019

CogitAI Inc. wants to improve artificial intelligence (AI) enterprise initiatives by removing the human element of manually labeling data for training machine learning algorithms.

The 2015 Orange County start-up backed by Sony Corp. today (Feb. 12) announced the availability of its Continua Software-as-a-Service (SaaS) self-learning AI platform that takes the "continual learning" approach, which it said builds on AI developments in machine learning, deep learning and reinforcement learning. Concerning the latter, perhaps lesser-known approach, the company says its platform is a new generation of reinforcement learning, which it says in turn "improves on supervised learning by enabling the machine learning algorithms to choose actions and discover which sequences of actions lead to mastering a predefined task."

While still fairly immature -- CogitAI says there's "still more research and development to be done" -- continual learning is advancing on other fronts as well, such as the Continual AI open community of researchers and enthusiasts interested in continual/lifelong learning for AI. "One of the grand goals of AI is building an artificial 'Continual Learning' agent that constructs a sophisticated understanding of the world from its own experience through the autonomous incremental development of ever more complex skills and knowledge," that site says.

CogitAI CEO Mark Ring had similar thoughts in announcing Continua: "The biggest opportunities for AI to expand into practical business applications cannot happen until AI systems can learn on their own," Ring said. "As successful as AI has been so far, its applicability will accelerate dramatically when it no longer depends on data provided by humans. CogitAI's continual learning approach is a watershed moment in the industry, providing a general-purpose solution that can learn from its own decisions, actions, and experience."

The continual learning approach can reportedly be applied to many kinds of resources, with the ability to automate nearly any repetitive decision-making process. The company says its platform can turn any process, system, software bot or physical robot into a self-learning autonomous service that can drive actionable business outcomes.

Example use cases for self-learning across many applications listed by CogitAI include:

• Vehicles
• Video games
• Building management
• Robotic Process Automation (RPA) and Business Process Management (BPM)
• Customer-service bots
• Web marketing
• AI coaches personalizing wellness and fitness
• AI software testing bots that learn to find and fix issues
• On-board calibration and real-time energy management for semiconductors

The platform can be hosted in the cloud or on-premises behind a company firewall. It includes client libraries for Java, JavaScript, Python and other programming languages that work with Continua, along with lower-level APIs for machine learning and customization of reinforcement learning as needed. Using the service requires setting up a corporate account; no pricing details were provided.