A Very Beginner Guide to Large Action Models (LAM)
In CES 2024, A company called “Rabbit” introduced a device called “R1” which is an AI personal assistant device. But, the best part is that it uses a new model called the “Large Action Model” which gives lots of ground-breaking opportunities to the existing digital device space. In this article, let’s talk about what is a Large Action Model and how we can use it.
What is a Large Action Model?
A Large Action Model is a model that can take actions by what you prompt to it. According to the Rabbit R1 keynote, LAMs are capable of understanding any sort of user interface and navigating through it just like a human being.
Example #1: Let’s say you want to hire an Uber to your home. Instead of opening the app and tapping a few buttons, you can tell the LAM to book an Uber ride to your home.
Example #2: Imagine a situation where your assistant is on leave and you want to update an important spreadsheet. LAMs can do that too.
LAM vs LLM
Large Language Models can understand what you prompt and give you textual or visual outputs. Of course, you can tweak it to do some actions if you give it a good prompt and pass some APIs 😉 but it takes some hard work.
With Large Action Models, you don’t need to worry about it. They’re designed to understand what you say and take action with it.
Example: If you tell a Large Language Model to book a hotel room from your favorite website, it cannot do that because currently, it does not have the capability to do that. Instead, it will show you the recommendations for a hotel room. If you tell the same thing to a Large Action Model, it will give you recommendations and also book the hotel room for you.
Where can we use Large Action Models?
Large Action Models will be able to be used in various industries such as:
- Healthcare: Patient monitoring and diagnosis support.
- Consumer Electronics: Personalized user experiences in devices such as Rabbit R1
- Robotics: Enhancing automation and human-robot interaction.
- Interactive Learning and Education: Understand student behavior and curate the learning experience
- Retail and Customer Service: Analyze the customer experience in a supermarket and create a personalized customer experience
- Content Creation and Media: Enhance the ability to create adaptive content
The Future of Large Action Models
With the advancement of Large Action Models, we’ll be able to see a huge growth in the consumption of automated devices and the increase in collaboration between humans and machines.
There will be more research into this and we’ll be able to see these types of models everywhere just like what is happening with Large Language Models.
