Bridging the gap between supervised and unsupervised learning for your AI systems

Abstract: As machine learning and deep learning efforts become commonplace in data science departments, organizations are struggling to wring as many insights as they can from the data they collect, and to utilize the limited artificial intelligence talent they have. Training AI systems using machine learning and deep learning can help organizations create better products, learn more about how users interact with their products, and ultimately serve larger business goals. But with strict time and resources, how should data scientists best balance training their machines with supervised and unsupervised learning methods?

Enter instructors, which serve as the middleman between supervised and unsupervised learning. Sasha Caskey, CTO and co-founder of conversational AI platform Kasisto, first became familiar with instructors during his time at the Stanford Research Institute, where its use was bred from a need to produce high quality machine learning data for new domains at scale. Once just a common tool within the academic community, Kasisto has employed AI instructors and continuous learning in its AI stack since the company was founded five years ago.

In this session, Sasha will outline the many benefits of using AI instructors and continuous learning, including: the ability to train systems on existing data from users/customers more quickly and efficiently; the ability to save time and resources by employing a responsible mix of supervised and unsupervised learning; how the technology can be employed without needing to scale other tools along with it; and the measurable impact such a practice has on product development. Additionally, beginner to immediate-level attendees will leave with a firm understanding of how to employ the use of AI instructors, beginning with an introduction to the use of instructors for ML/DL efforts, all the way to the ability to employ instructors in their own stacks at an enterprise scale.

Bio: Sasha Caskey, CTO and Co-founder of Kasisto, the creators of KAI, the leading conversational AI platform for the finance industry. Sasha has over 15 years of experience inventing and transitioning breakthrough technology into applications and products, and has spent the entirety of his career continually teaching computers to better communicate with humans. At Kasisto, Sasha drives the architectural direction and software development of the company’s KAI platform – integrating its industry-agnostic, full stack AI technologies with prepackaged banking knowledge. Before founding Kasisto, he was a member of the IBM TJ Watson research team, The MITRE Corporation and Speechworks Intl. He has a B.A. from Brandeis University and M.A. from Columbia University.

Open Data Science Conference