Productive AI is here

Today, we're going to talk about three mysterious terms you might have heard before:- P, NP, and EXP.

This is also important to understand them as using these you will know what all AI can or can't do, why your digital transactions are safe so far, why AI solutions are associated with GPUs or what makes it difficult to have lighter size AI models or why Amazon is never able to predict with last minute accuracy when will your package arrive, yet. Understanding them is important as helps you appreciate the power and limitations of computers.

Don't worry if they sound complicated—we're going to break them down and explain them in a way that's easy to understand. We are starting this new section, in simple words, to occasionally write concepts in simple words which someone with no background in this scientific concept can understand enough to get started in the domain.

Now back to the topic at hand:

P: Problems with "Quick" Solutions

First, let's start with P. P stands for "Polynomial-time," and it's a class of problems that can be solved "quickly" by computers. By "quickly," we mean that as the size of the problem grows, the time it takes to solve the problem doesn't grow too fast. In other words, these problems have a reasonable solution time.

Example 1:Imagine you have a list of numbers and want to find the largest one. A computer can do this relatively quickly by comparing each number to the others. This is an example of a problem in P.

Example 2: Determining if a number is even or odd is a problem in the P class. We can easily perform this operation by checking if the number is divisible by 2. This operation takes a constant amount of time, regardless of the number's size.

So if we make a table and the difficulty to validate a solution, it will look like this:

Imagine you have a picture book with different images, like animals, toys, and foods. You want to know more about the pictures, like how many animals are there, what toys are in the picture, or which food is the biggest.

Right now, computers use something called end-to-end models to figure out these things. It's like they look at the whole picture and try to guess the answers, but sometimes, they can't explain how they got those answers.

In this paper, authors came up with a new way for computers to understand pictures, just like how you put puzzles together, piece by piece. This new way is called a ViperGPT, and it uses a helper called GPT-3 Codex. This helper is like a super smart friend who knows a lot about different things.

So, when you want to know something about a picture, you can ask GPT-3 Codex a question, like "how many animals are there?" GPT-3 Codex will then create a set of instructions to find the animals and count them. It's like teaching the computer to look at the picture step-by-step, just like you would do.

This new way of looking at pictures is really helpful because it can explain how it found the answers, and it's also easy to teach the computer new things or improve how it understands the pictures. This makes it better at answering questions about the images in your picture book.

People naturally combine individual steps to understand the visual world, employing compositional reasoning. However, the field of computer vision predominantly relies on end-to-end models, which lack compositional reasoning and interpretability. While end-to-end models have advanced object recognition and depth estimation, they struggle to generalize and remain uninterpretable. The author presents a new framework that leverages code-generating large language models, like GPT-3 Codex, to flexibly compose vision models based on textual queries.