Introduction

Artificial Intelligence programme recreates entire periodic table of elements

A new AI programme developed by Stanford scientists recreated the periodic table in just a few hours
  • Client

    Stanford University

  • Services

    To distinguish between different atoms after analysing a list of chemical compound names

  • Technologies

    Artificial Intelligance (AI)

  • Dates

    28/06/2018

Description

The tabular arrangement of the chemical elements, ordered by atomic numbers, electron configuration and chemical properties — commonly called the periodic table — not only came to be the favourite poster in school classrooms around the world, but also represents a century of fine tuning by the brightest scientists around the globe.

 

The first version of the modern periodic table, which has been instrumental in predicting new elements in the universe and their behaviour, was first proposed by Russian chemist Dmitri Mendeleev in 1869. But the journey to the periodic table in its current form began in 1789 when Antoine Lavoisier published a list of 33 chemical elements, grouping them into gases, metals, nonmetals and earths, eventually leading to a century of chemists dedicated to classification of elements.

 

This monumental exercise in science was replicated by an artificial intelligence (AI) programme developed by Stanford physicists within hours.

 

The AI programme, called Atom2Vec, first learned to distinguish between different atoms after analysing a list of chemical compound names from an online database, Phys reported. Then, it used concepts borrowed from the field of natural language processing to cluster the elements according to their chemical properties. From such sparse data, the programme figured out, for example, that potassium and sodium must have similar properties because both elements can bind with chlorine.

 

"We wanted to know whether an AI can be smart enough to discover the periodic table on its own, and our team showed that it can," the report quoted study leader Shou-Cheng Zhang as saying. Zhang said the research was published in the 25 June issue of Proceedings of the National Academy of Sciences.

 

He hopes that in the future, scientists can harness Atom2Vec's knowledge to design new materials. "For this project, the AI program was unsupervised, but you could imagine giving it a goal and directing it to find, for example, a material that is highly efficient at converting sunlight to energy," Zhang was quoted as saying.

Related