The quantum tunneling (QT) effect discovered in the 1920s was a major achievement in the field of quantum mechanics. Since there is a major fundamental difference between artificial and biological neurons in every aspect, Artificial intelligence struggles to replicate human perception, particularly with complex visuals like the Necker cube and Rubin’s vase. Artificial Intelligence cannot switch between different interpretations of these illusions due to their neurological and psychological processes, just as humans have.
Current approaches to machine vision, highly dependent on conventional deep neural networks (DNNs) with standard activation functions like ReLU, face limitations in duplicating human-like perception of optical illusions. Deep Neural Networks often struggle to capture the physical, physiological, and psychological factors that guide human vision, thus making it difficult to behave in the same manner as humans do. A deep neural network is a collection of machine learning algorithms inspired by the overall structure and functioning of the brain, with multiple layers of nodes between the input and outputs and many hidden layers. The human brain, which can effortlessly switch between multiple interpretations of a single visual stimulus, exhibits quantum-like behavior that traditional DNNs fail to replicate. Researchers from Charles Sturt University, Australia, have therefore proposed a novel neuromorphic DNN model that incorporates quantum tunneling (QT) as its activation function, called QT-DNN.
The QT-DNN model is tested on Necker Cube and Rubin’s vase illusions. The research team’s major innovation revolves around the idea of using quantum tunneling probability equations as the activation function for the hidden layer nodes, replacing activation functions like sigmoid or ReLU. QT-DNN uniquely uses a physical quantum random number generator to ensure unbiased visual information processing.
Its architecture consists of an input layer with 100 nodes, three hidden layers with 20 nodes each, and an output layer with two nodes for classification. It can initiate not just switching between different interpretations of the illusions but also intermediate states that represent a superposition of multiple perceptions, which is easily observed in human subjects but is difficult to replicate with classical DNNs. Compared to traditional DNNs, QT-DNN showed better alignment with theoretical predictions from quantum models and experimental observations of human perception.
In conclusion, QT-DNN, designed on quantum mechanical principles, is a unique method to reduce the gap between machine and human perception with its helpful applications in fields requiring human-like visual processing, such as aviation safety, augmented reality systems, and medical diagnostics. The research opens a new gateway for developing more sophisticated AI systems that better interpret visual information in ways similar to human perception.
Check out the Paper and Details. All credit for this research goes to the researchers of this project. Also, don’t forget to follow us on Twitter and join our Telegram Channel and LinkedIn Group. If you like our work, you will love our newsletter.. Don’t Forget to join our 55k+ ML SubReddit.
[Sponsorship Opportunity with us] Promote Your Research/Product/Webinar with 1Million+ Monthly Readers and 500k+ Community Members
Nazmi Syed is a consulting intern at MarktechPost and is pursuing a Bachelor of Science degree at the Indian Institute of Technology (IIT) Kharagpur. She has a deep passion for Data Science and actively explores the wide-ranging applications of artificial intelligence across various industries. Fascinated by technological advancements, Nazmi is committed to understanding and implementing cutting-edge innovations in real-world contexts.
Leave a comment