What Tensors Meaning, Applications & Example
Multidimensional arrays used in machine learning and deep learning.
What are Tensors?
Tensors are multi-dimensional arrays or matrices that are a fundamental data structure in machine learning and deep learning . They represent data in the form of arrays with varying dimensions, such as scalars (0D), vectors (1D), matrices (2D), and higher-dimensional arrays (3D or more). Tensors are used to store and manipulate data, such as inputs, outputs, and parameters, in machine learning models.
Types of Tensors
- Scalar (0D Tensor): A single number, such as 5 or -3.14.
- Vector (1D Tensor): An ordered list of numbers, such as [1, 2, 3, 4].
- Matrix (2D Tensor): A 2-dimensional array, such as a table or grid, represented as rows and columns.
- Higher-Dimensional Tensor: Tensors that represent data with three or more dimensions, such as a batch of images with width, height, and color channels (e.g., 3D tensors for images).
Applications of Tensors
- Neural Networks: Tensors are used to represent the inputs, weights, and outputs in neural networks.
- Data Manipulation: Tensors allow for efficient mathematical operations, such as matrix multiplication and element-wise addition, which are essential in training machine learning models.
- Computer Vision and NLP: Tensors represent images, text, and other data types used for tasks like image classification and language modeling.
Example of Tensors
In a neural network , the input data (e.g., an image) is represented as a tensor. For instance, a color image of size 28x28 pixels with RGB color channels would be stored as a 3D tensor of shape (28, 28, 3). Operations like convolution, activation functions, and backpropagation all involve manipulating tensors to learn the model parameters during training.