Machine Learning in TypeScript: Libraries

Machine Learning in TypeScript: Overview of Key Libraries

TypeScript — a statically typed superset of JavaScript — is increasingly used in projects related to machine learning (ML). Thanks to strict typing, improved code readability, and excellent IDE support, TypeScript allows you to create reliable ML applications that run both in the browser and on the server (Node.js). In this article, we will review the main libraries for machine learning in TypeScript, their capabilities, and provide practical examples.

1. TensorFlow.js — The Flagship of Machine Learning in TypeScript

TensorFlow.js is a port of the popular TensorFlow framework for JavaScript and TypeScript. It allows you to run machine learning models directly in the browser or on Node.js without server infrastructure. The library supports all major ML stages: loading pre-trained models, training new models, inference (prediction), and even fine-tuning.

Installation:

npm install @tensorflow/tfjsnpm install @tensorflow/tfjs-node  # for Node.js (with acceleration)

Example: Image classification with a pre-trained MobileNet model

import * as tf from '@tensorflow/tfjs';import { loadGraphModel } from '@tensorflow/tfjs-converter';

async function classifyImage(imageElement: HTMLImageElement) {  // Load the pre-trained MobileNet model  const model = await loadGraphModel(    'https://tfhub.dev/google/tfjs-model/imagenet/mobilenet_v2_100_224/classification/3/default/1'  );

  // Convert the image to a tensor  const tensor = tf.browser.fromPixels(imageElement)    .resizeNearestNeighbor([224, 224])    .toFloat()    .expandDims();

  // Get the prediction  const prediction = model.predict(tensor) as tf.Tensor;  const topK = tf.topk(prediction, 3);  const indices = await topK.indices.data();  const values = await topK.values.data();

  console.log('Top-3 predictions:', indices, values);}

// Usageconst img = document.getElementById('myImage') as HTMLImageElement;classifyImage(img);

TensorFlow.js is ideal for tasks in computer vision, text processing (NLP), and time series. In TypeScript, you get autocompletion for all methods, which significantly speeds up development.

2. ML5.js — A Friendly Interface for Creative ML

ML5.js is built on top of TensorFlow.js but offers a simpler and higher-level API. This library is aimed at artists, designers, and developers who want to quickly add ML functionality without diving deep into the mathematics. ML5.js supports tasks such as image classification, object detection, text generation, sound analysis, and much more.

Installation:

npm install ml5

Example: Object detection using COCO-SSD

import ml5 from 'ml5';

async function detectObjects(videoElement: HTMLVideoElement) {  // Load the object detection model  const objectDetector = await ml5.objectDetector('cocossd');

  // Perform detection on the video  const results = await objectDetector.detect(videoElement);

  results.forEach((obj: any) => {    console.log(`Object: ${obj.label}, confidence: ${obj.confidence.toFixed(2)}`);    console.log(`Coordinates: x=${obj.x}, y=${obj.y}, width=${obj.width}, height=${obj.height}`);  });}

// Usage with webcamconst video = document.getElementById('webcam') as HTMLVideoElement;navigator.mediaDevices.getUserMedia({ video: true })  .then(stream => { video.srcObject = stream; })  .then(() => detectObjects(video));

ML5.js is great for prototyping, educational projects, and interactive installations. In TypeScript, you can define types for detection results, which makes