Researchers have developed a brand new AI algorithm, called Torque Clustering, that is way closer to natural intelligence than current methods. It significantly improves how AI systems learn and uncover patterns in data independently, without human guidance.
Torque Clustering can efficiently and autonomously analyse vast amounts of information in fields akin to biology, chemistry, astronomy, psychology, finance and medicine, revealing recent insights akin to detecting disease patterns, uncovering fraud, or understanding behaviour.
“In nature, animals learn by observing, exploring, and interacting with their environment, without explicit instructions. The following wave of AI, ‘unsupervised learning’ goals to mimic this approach,” said Distinguished Professor CT Lin from the University of Technology Sydney (UTS).
“Nearly all current AI technologies depend on ‘supervised learning’, an AI training method that requires large amounts of information to be labelled by a human using predefined categories or values, in order that the AI could make predictions and see relationships.
“Supervised learning has plenty of limitations. Labelling data is dear, time-consuming and sometimes impractical for complex or large-scale tasks. Unsupervised learning, in contrast, works without labelled data, uncovering the inherent structures and patterns inside datasets.”
A paper detailing the Torque Clustering method, Autonomous clustering by fast find of mass and distance peaks, has just been published in IEEE Transactions on Pattern Evaluation and Machine Intelligence, a number one journal in the sphere of artificial intelligence.
The Torque Clustering algorithm outperforms traditional unsupervised learning methods, offering a possible paradigm shift. It’s fully autonomous, parameter-free, and may process large datasets with exceptional computational efficiency.
It has been rigorously tested on 1,000 diverse datasets, achieving a mean adjusted mutual information (AMI) rating — a measure of clustering results — of 97.7%. As compared, other state-of-the-art methods only achieve scores within the 80% range.
“What sets Torque Clustering apart is its foundation within the physical concept of torque, enabling it to discover clusters autonomously and adapt seamlessly to diverse data types, with various shapes, densities, and noise degrees,” said first creator Dr Jie Yang.
“It was inspired by the torque balance in gravitational interactions when galaxies merge. It is predicated on two natural properties of the universe: mass and distance. This connection to physics adds a fundamental layer of scientific significance to the tactic.
“Last 12 months’s Nobel Prize in physics was awarded for foundational discoveries that enable supervised machine learning with artificial neural networks. Unsupervised machine learning — inspired by the principle of torque — has the potential to make the same impact,” said Dr Yang.
Torque Clustering could support the event of general artificial intelligence, particularly in robotics and autonomous systems, by helping to optimise movement, control and decision-making. It is ready to redefine the landscape of unsupervised learning, paving the best way for truly autonomous AI. The open-source code has been made available to researchers.
