From the course: Implementing a Data Strategy for Responsible AI
Join today to access over 25,600 courses taught by industry experts.
Foundation models of generative AI
From the course: Implementing a Data Strategy for Responsible AI
Foundation models of generative AI
- [Instructor] Generative AI had its public release in late 2022. This meant that people with little to no background could both consume and produce new AI-aided tools. Some reports state that over 2,000 generative AI tools are currently available, and they span industries, sectors, and modalities. Popular ones include Microsoft Copilot, GitHub Copilot, Midjourney, Grammarly, DALLE-3, and Google Gemini. The backbone of generative AI lies in its neural network architecture. This architecture has its foundation in deep learning with a general framework of connected layers. The input layer receives data from various sources, and hidden layers use the Inputs features to make an algorithmic guess of the relationship between the existing inputs. The output layer produces new data, given the statistical guesses computed within the hidden layers. The transformer model combines this neural network architecture with unsupervised learning approaches. The hidden layers consist of encoding and…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.