In this episode, Anastassia and Dimitry Fisher discuss computer vision technologies and their evolution. Dimitry explains that computer vision is the ability of artificial systems to acquire, process, and act upon visual input. He highlights the three main directions from which computer vision emerged: pre-World War II television, early computers, and the study of animal and human vision. Dimitry also discusses the most critical technologies in computer vision, such as convolutional neural networks (CNNs), generative adversarial networks (GANs), and transformers. He emphasises the importance of labelled data and using pre-trained models in computer vision. The conversation also touches on the ethics and future of computer vision technologies.
About Dimitry Fisher
Dimitry Fisher is a distinguished AI scientist with extensive experience in neuroscience, machine learning, and data science. He serves as the Senior Vice President of Data Science at Aicadium (https://aicadium.ai/), an AI company committed to building AI products across industries and business functions.
Dimitry earned his PhD in Plasma and High-Temperature Physics, atomic physics, and hot-dense matter from the Weizmann Institute of Science. He was a Senior Scientist at Brain Corporation, where he developed large-scale vision models and researched sensory-motor learning algorithms for robots and AI. His postdoctoral work at UC Davis and the Weizmann Institute of Science further solidified his expertise in neuroscience and computational algorithms of the brain cortex.
Key Takeaways
- Computer vision is the ability of artificial systems to acquire, process, and act upon visual input
- Computer vision emerged from three main directions: pre-World War II television, early computers, and the study of animal and human vision
- Groundbreaking technologies in computer vision include convolutional neural networks, GANs, and transformers
- Labelled data is essential in computer vision, and pre-trained models are often used to reduce the need for large amounts of labelled data
- Ethics play a crucial role in developing and deploying computer vision technologies
- The future of computer vision involves advancements in co-bots, autonomous machines, and multimodal AI
Chapters
00:00 Introduction to Computer Vision
02:03 The Three Directions of Computer Vision Emergence
05:16 Groundbreaking Technologies in Computer Vision
07:37 The Importance of Labeled Data and Pre-Trained Models
18:35 Ethics and the Future of Computer Vision
21:14 Advancements in Co-bots, Autonomous Machines, and Multimodal AI