Quantum Extreme Learning Machine Noise Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current quantum machine learning algorithms face challenges in adapting to complex non-temporal classification tasks due to decoherence and hardware noise in NISQ devices, particularly in leveraging noisy dynamics for improved learning effectiveness.
Innovation Solution
A quantum-based extreme learning machine system utilizes noise in a quantum substrate to enhance classical machine learning methods by generating complex output states, employing a gate-based implementation with noisy quantum gates and encoding techniques like basis encoding, amplitude encoding, and Hamiltonian encoding to compute optimal weights for prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If quantum computing is applied to machine learning algorithms, then training speedup and performance enhancement are achieved, but decoherence and hardware noise in NISQ devices prevent effective adaptation to complex non-temporal classification tasks
Solution Approach 1:
The patent converts the harmful effect of quantum noise and decoherence into a beneficial feature by designing a quantum extreme learning machine that leverages noisy quantum dynamics to enhance learning effectiveness. The noise in the quantum substrate is used to generate complex output states that improve classification performance on non-temporal tasks, transforming the previously detrimental noise into a useful resource for achieving both speedup and reliability.
2Adaptability or versatility
If noise in quantum substrate is utilized to enhance learning, then complex output states are generated for improved classification, but quantum noise remains a strong limiting factor in gate-based quantum computing
Solution Approach 1:
The patent applies the blessing in disguise principle by accepting and utilizing quantum noise as a beneficial feature rather than attempting to eliminate it. The noisy quantum substrate generates complex dynamics and output states that enhance the model's adaptability to non-temporal classification tasks, turning the harmful noise into a useful resource for improving learning effectiveness.
3Ease of manufacture
If gate-based implementation is used on superconducting quantum processors, then quantum substrate can be constructed with noisy quantum gates, but the implementation complexity increases due to encoding requirements
Solution Approach 1:
The patent employs parameter changes by utilizing the noisy dynamics of quantum gates as a key feature of the model. Instead of attempting to correct or eliminate noise, the system changes the operational parameters to leverage the noisy quantum evolution, allowing the quantum substrate to generate complex output states that improve learning while maintaining feasibility on NISQ devices.
Data Source
AI summary
A quantum-based extreme learning machine and a method for training a quantum-based extreme learning machine using a quantum processor (50) implementing a quantum substrate (220) and a set of training data (260) is disclosed. The training data comprises input features vectors (400) with a plurality of N parameters (262) and true labels vector (265). The method comprises uploading (S610) the training data (260) to the quantum processor (50), encoding (S620) the uploaded training data (260), passing (S410, S630) a plurality of subsets of the input features vector (400) from the training data (260) through the quantum substrate (220) to obtain a plurality of output vectors of expectation values (420), concatenation (S420) of the plurality of output vectors of expectation values (420) to construct (S430) a matrix (430), computation (S440) of an inverse matrix (H) from the matrix (430) and multiplication (S450) of the inverse matrix (H) by the true labels vector to obtain a vector (470) of optimal weights β.