
Building the factories of the future with Covariant CEO Peter Chen
No Priors: Artificial Intelligence | Technology | Startups
Learning Grounded Multimodal Understanding from Image and Text Pairs
Multimodal models can be trained using image and text pairs found on the Internet to develop a grounded understanding of concepts. By analyzing large amounts of data, these models can start to associate visual and symbolic concepts with real-world objects such as apples. However, for robotics to improve, it is essential to collect data that captures missing information such as the physical properties and interactions of the objects.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.