Please register to join us November 19 at 6:15pm to hear from our speakers:
- Cyrus Shaoul, Co-founder and CEO, Leela AI
- Amit Seker, Research Engineer, Babel Street
Meetup Program
Talks
Digital Visual Agents Make Manufacturing Safer and More Efficient
Speaker: Cyrus Shaoul, Co-found & CEO, Leela AI
Learn how hybrid causal learning architectures are being used in digital visual agents to provide explainable insights from ‘dark’ (un-instrumented) areas of the manufacturing floor to help operations leaders and safety managers make decisions that optimize operations and keep their people safer. This approach is explainable and much more efficient at parsing visual sensor data than using neural nets alone in real world, industrial applications. Taking Leela AI’s system as an example, we’ll speak about the value that digital visual agents are delivering and show why monolithic systems (such as LLMs) have inherent limitations that will be hard to overcome in this domain.
Leela AI’s technology is deployed at industrial sites globally, where these agents are working together with production supervisors and industrial engineers at some of the world’s largest manufacturers, helping to build critical products that move millions of tons of cargo, cool billions of square feet of built environment and pump trillions of liters of water every year.
Tracking Cognitive Development of Large Language Models
Speaker: Amit Seker, Research Engineer, Babel Street
In this talk, I’ll explore how cognitive development theories can help us understand the evolving capabilities of large language models (LLMs). Drawing on Piaget’s Theory of Cognitive Development, we’ll assess LLMs on cognitive abilities ranging from simple object recognition to complex logical reasoning. The findings reveal that advanced models like GPT-4 display cognitive abilities comparable to a 20-year-old human, and that factors like model size and optimization objectives play a critical role in their cognitive growth. This discussion will shed light on how LLMs’ cognitive abilities are linked to their performance and where future improvements might be focused.
Amit Seker
Amit has been working in natural language processing and AI for over 20 years with a particular focus on morphologically rich languages. Currently he is an engineering manager at Babel Street working on machine learning and text analytics and is active in research at the ONLP Lab of Bar-Ilan University. Previously he was co-founder/VP of R&D at Comprendi, a startup that delivers advanced machine learning and NLP-powered ad targeting automation. Amit holds a BS in computer science from UC Santa Barbara and an MS in computer science from The Open University.
Cyrus Shaoul
Cyrus is an entrepreneur and computational neuroscientist with expertise in computational cognitive modeling. Cyrus is one of the co-founders and the CEO of Leela AI, which uses a novel kind of visual intelligence to improve productivity and safety in manufacturing. The Leela AI platform has been deployed commercially since 2021, and has been is now in production at multiple Fortune 500 companies. Previously, Cyrus was co-founder and CTO of Digital Garage, which operated a search engine called Infoseek Japan, the dominant search engine in Japan at the beginning of the Internet Age. He was at Digital Garage from a Day 1 through to a successful IPO. He holds a BSc. from MIT and a PhD from the University of Alberta.
