The Google Assistant is arguably the smartest voice assistant out there, but it even suffers in certain situations. Fortunately, the software giant has been working to improve natural language recognition capabilities and has Announced LaMDA today at Google I / O 2021.
Google’s new LaMDA language model
The Language Model for Dialogue Applications (LaMDA) is Google’s latest language model built on top of its open source Transformer neural network architecture. What makes LaMDA different from existing language models like BERT and GPT-3 is the fact that it is trained in dialogue..
Google says LaMDA picked up “Several of the nuances that distinguish open conversation from other forms of language” during the training process. This is really promising as we often get irrelevant responses from the Assistant if you lose the context.
At I / O 2021, Google showed a quick LaMDA demo in a conversation involving Pluto and a paper airplane. As you can see in the image below, the sample outputs look decent enough. Google’s Sundar Pichai says LaMDA offers sensible answers while keeping the dialogue open. As a result, it is possible to have a new conversation on a different topic without retraining the model.
However, there is a catch. Like most artificial intelligence systems, LaMDA is not yet perfect. Google even showed a slide where it managed to screw up the conversation and you can see that below.
Google is also working on multimodal models that use images, text, audio, and videos. When this becomes a reality, interacting with Assistant should be a seamless experience. For example, you can ask Google to find a route with beautiful mountain views for your road trip. You can even search for specific parts of a video with voice commands. How cool is that!
With natural conversations as the key focus, LaMDA should, in theory, help significantly improve the Assistant and could also change the way we interact with it. Unfortunately, LaMDA is not immediately rolling out to current Google products yet.
The company plans to use LaMDA in many of its products, including Google Assistant, Search, and Workspace in the future. Google is currently focusing on making LaMDA’s responses compelling and objective before taking advantage of its capabilities.