MIT Researchers Resort To Quantum Computation For Handling Big Data

MIT, the University of Waterloo and the University of Southern California have collaborated to design a new quantum computing approach to handle big data. Complex topological systems are the key required to unlock thousands of compound networks such as neural networks, global interconnections of the Internet, country based power grid etc.

However, such systems require large time to process and compute using classical algorithms. The team devised a quantum machine learning technique that can solve the problems in real time, without any extra lag. The report was published in “Nature Communication” open access journal under the title “Quantum algorithms for topological and geometric analysis of data”.




Every complex study requires huge amount of data to be monitored, which takes help of machine learning. Some of these analytical methods are based on a branch of mathematics known as topology. It is the study of properties that are preserved through deformations aka twisting, stretching but not tearing. Seth Lloyd, the paper's lead author and the Nam P. Suh Professor of Mechanical Engineering explained that algebraic topology forms the skeleton of their method. This theory helps to reduce the distortions in the raw data, thus making them error free.

Topologically speaking, a circle and an ellipse are equivalent. Inspired by such conception, the algorithm only treats objects categorically, interpreting their topology. However, by using classical computation, the system would take enormous time to process, making it computationally expensive. This is where the team applied the principles of quantum mechanics to simplify the time based solution. Lloyd explained that if a sample had 300 data points, then classical algorithm would take 2^300(two to the power 300) processing units-approximately equal to number of particles in universe whereas the new algorithm would take only 300 quantum bits. This clearly indicates that Quantum computation is the only way forward.

The paper is based primarily on strong theoretical background and practical implementation would take considerable amount of time. Lloyd however claimed that experimentalists have already contacted them regarding trying prototypes for experimental realization, and that the system could easily be built in next few years.

Watch a relevant explanation of Topology:


Source: A new quantum approach to big data | MIT News | Massachusetts Institute of Technology | #-Link-Snipped-#