Basic Tensor Operations

Chief Nazari

Raw data is useless. We need to clean it, combine sensor feeds, and amplify the signal. Every operation needs to be fast — we're processing millions of data points per second.

You've used Array.map() to transform lists, reduce() to aggregate values, and spread operators to combine arrays. Tensor operations are the same ideas — but they run on the GPU and handle millions of elements in parallel.

Element-wise Operations

When you write arr.map((x, i) => x + arr2[i]), you're doing element-wise addition. TensorFlow.js does the same thing, but across any number of dimensions and on the GPU:

import * as tf from '@tensorflow/tfjs';

const sensor1 = tf.tensor([1, 2, 3]);
const sensor2 = tf.tensor([4, 5, 6]);

// Element-wise operations — same as map + zip
const sum = tf.add(sensor1, sensor2);     // [5, 7, 9]
const product = tf.mul(sensor1, sensor2); // [4, 10, 18]
const diff = tf.sub(sensor2, sensor1);    // [3, 3, 3]

// Chaining with method syntax
const result = sensor1.add(sensor2).mul(tf.scalar(2));
// [10, 14, 18]

// Matrix multiplication — the workhorse of neural networks
const weights = tf.tensor2d([[1, 2], [3, 4], [5, 6]]);  // [3, 2]
const input = tf.tensor2d([[1], [1], [1]]);               // [3, 1]
const output = weights.transpose().matMul(input);          // [2, 1]

Matrix multiplication (matMul) is the operation that makes neural networks work. Every layer in a neural network is essentially: output = matMul(input, weights) + bias. You'll see this pattern everywhere from here on.

Challenge

Add two tensors together to combine sensor readings.

const a = tf.tensor([1, 2, 3]);
const b = tf.tensor([4, 5, 6]);
const result = null; // add a and b