What is a tensor in physics terminology and what’s the difference from a tensor in machine learning and AI?

June 24, 2019 / Ask Slater, Machine Learning

There’s no difference. A tensor is just a very, very generic term for: basically anything. A tensor is a generic term for a way of arranging numbers that generally has some geometric interpretation.

A scalar (math speak for a number) is a 0-tensor.

A vector (math speak for a list of numbers) is a 1-tensor

A matrix (math speak for a grid of numbers) is a 2-tensor

So on and so forth. Often people will only use the tensor to refer to 3+ tensors since we’ve got other words for smaller tensors. This is slightly more common in physics, but I don’t have any data that confirms it’s used any differently between physics and ML.

A tensor can represent pretty much anything though. For computational reasons it’s rare to deal with 3+ tensors in ML, and 3 tensors are more common in physics. You see them in ML though for things like color images (3-tensor), video (4-tensor) and more complex imaging (i.e. MRIs).

One important note and point of confusion is that the “dimensions” of the tensor are referred to as modes. A vector is a tensor with a single mode, not a single dimension. A vector can have infinitely many dimensions, but it only has a single mode.

View original question on Quora >

Follow Slater on Quora >>

Don't Miss a Post!

Subscribe to indico's monthly newsletter to receive the latest blog posts and AI industry news.