IBM adds machine learning to quantum computer SDK

Get a free Techzine subscription!

IBM has released Qiskit Machine Learning. The new module for the devkit for quantum computers allows developers to outsource machine learning tasks to quantum computers.

Quantum computers offer a new way to apply machine learning models, Medium writes. Quantum machine learning (QML), as the technique is called, should capitalise on the unique capabilities offered by quantum computers and offer better performance than traditional machine learning. After all, in theory, quantum computers can handle vast amounts of data extremely well.

Fast calculations, slow input

With QML, however, it is difficult to feed the large data sets required for machine learning into the quantum computer. If the calculations are blazingly fast, but the process is slower in other respects, there is no bottom-line benefit. Nevertheless, researchers have already found that quantum computers can, in theory, handle certain tasks that a normal computer cannot solve efficiently.

However, we are still a long way from this being the case in practice. In order to give machine learning developers some training, IBM has added a new module to Qiskit. This allows developers who have no expertise in quantum computers to try out how to work with such computers.

Features

An example of the functions of Qiskit Machine Learning is QuantumKernel, a tool that calculates kernel matrices for a data set in a quantum framework. Furthermore, the module contains several implementations of neural networks for quantum computers and the algorithms to train and use them. This allows developers to build and maintain their own networks.

Finally, developers can use Qiskit Machine Learning to integrate their neural networks directly with the open source PyToch library for machine learning. This is a Facebook platform used primarily for computer vision and natural language processing.

SDK for quantum computers

Qiskit is a framework from IBM with which developers can build applications for quantum computers. The system is intended to be able to work on different quantum computers, making development more general. IBM recently introduced a new certification that allows developers to demonstrate proficiency with Qiskit.