Google Unveils LaMDA to Solve the Puzzle of Open-Ended Conversation
Google introduces LaMDA, a new language model built specifically for dialogue that aims to make conversational AI feel more natural. Unlike traditional chatbots, this technology freely navigates a seemingly endless number of topics.
Google unveils LaMDA, a breakthrough conversational AI technology designed to tackle the complexities of open-ended dialogue. Unlike traditional chatbots that follow narrow, pre-defined paths, LaMDA engages in free-flowing conversations that can naturally drift across a seemingly endless number of topics, much like a human conversation does.
This new language model builds on the Transformer neural network architecture that Google invented and open-sourced in 2017. While previous models like BERT and GPT-3 utilize this same architecture to predict upcoming words, LaMDA distinguishes itself by undergoing specialized training focused entirely on dialogue.
Through this dialogue-focused training, the system picks up on the subtle nuances that separate open-ended conversation from other forms of written text. Google believes this ability to understand the meandering, adaptable nature of human language unlocks more natural ways of interacting with technology and paves the way for entirely new categories of helpful applications.