Tensors Explained - Data Structures of Deep Learning

welcome back to this series on no network programming with pi torch in this video we will kick off section 2 of the series which is all about tensors we'll talk tensors terminology and look at tensor indices this will give us the knowledge we need to look at some fundamental tensor attributes that are used in deep learning without further ado let's get started [Music] tensors are the primary data structures used by neural networks the inputs outputs and transformations within neural networks are all represented using tensors as a result neural network programming utilizes them heavily the concept of a tensor is a mathematical generalization of other more specific concepts each of these examples are specific instances of the more general concept of a tensor let's organize this list of example tensors into two groups the first group of three terms number array in 2d array are all terms that are typically used in computer science while the second group scalar vector and matrix are terms that are typically used in mathematics we often see this kind of thing where different areas of study use different words for this same concept the terms in each group correspond to one another as we move from left to right to show this correspondence we can reshape our list of terms to get three groups of two terms the relationship within each of these pairs has to do with the number of indices required to access a specific element there are 0 indices required for a number and a scalar because you just refer to the actual number or scalar value you don't need an index when we move to an array or a vector we need one index to refer to a specific element and then when we move to a 2d array or a matrix we need two indices to refer to a specific element let's suppose we have an array called a with four elements now suppose we want to access the number three in this data structure we can do it using a single index like so as another example let's suppose we have this 2d array called D D notice that we need two indices to refer to the number three in this 2d array when more than two indices are required to access a specific element we stop giving specific names to the data structures and begin using more general language in mathematics we stop using words like scalar vector and matrix and we start using the word tensor or in the tensor the N tells us the number of indices required to access a specific element within the structure in computer science we stop using words like number array in 2d array and we start using the word multi-dimensional array or the word in D array I very rarely used words like vector and matrix because like they're kind of meaningless specific examples of something more general which is they're all n dimensional tensors so let's make this clear for practical purposes in deep learning and neural network programming tensors are multi-dimensional arrays physicists get crazy when you say that because to a physicist a tensor has quite a specific meaning but in machine learning we generally use it in the same way so tensors are multi-dimensional arrays or in D arrays for short the reason we say a tensor is a generalization is because we use the word tensor for all values of n for example a scalar is a zero dimensional tensor a vector is a one-dimensional tensor a matrix is a two dimensional tensor in an India Ray is an N dimensional tensor tensors allow us to drop these specific terms and just use an end to identify the number of dimensions we are working with one thing to know about the dimension of a tensor is that it differs from what we mean when we refer to the dimension of say a vector in a vector space the dimension of a tensor does not tell us how many components exist within the tensor if we have a three-dimensional vector from three-dimensional Euclidean space we have an ordered triple with three components however a three-dimensional tensor can have many more than three components our two-dimensional tensor dd for example had nine components in the next post we will cover the concepts of rank axes and shape these are the fundamental attributes of tensors that we use in deep learning if readings your thing I highly recommend you check out the blog post for this video on deep lizard com also check out the deep lizard hivemind for exclusive perks and rewards thanks again for contributing to collective intelligence I'll see you in the next one well it used to be that if you wanted to get a computer to do something new you would have to program it now programming for those of you who they haven't done it yourself requires laying out in excruciating detail every single step that you want the computer to achieve to do in order to achieve your goal now if you want to do something that you don't know how to do yourself then this is going to be a great challenge so this was the challenge faced by this man Arthur Samuel in 1956 he wanted to get this computer to be able to beat him at checkers how can you write a program layout in excruciating detail how to be better than you at checkers so he came up with an idea he had to compute a play against itself thousands of times and learn how to play checkers and indeed it worked [Music]

Channel: deeplizard

Views: 70,432

Keywords: deep learning activation function, AI, artificial intelligence, artificial neural network, autoencoders, batch normalization, clustering, CNN, convolutional neural network, data augmentation, deep learning, education, Tensorflow.js fine-tune, image classification, Keras, learning, machine learning, neural net, neural network, Python, relu, Sequential model, SGD, stochastic gradient descent, supervised, Tensorflow, Theano, transfer learning, tutorial, unsupervised learning, TFJS, PyTorch

Id: Csa5R12jYRg

Length: 6min 6sec (366 seconds)

Published: Tue Sep 11 2018