Graham D. Bruce

Kwang Jun Lee, Alexander C. Trowbridge, Graham D. Bruce et al.

Reliable identification of high-value products such as whisky is an increasingly important area, as issues such as brand substitution (i.e. fraudulent products) and quality control are critical to the industry. We have examined a range of machine learning algorithms and interfaced them directly with a portable Raman spectroscopy device to both identify and characterize the ethanol/methanol concentrations of commercial whisky samples. We demonstrate that machine learning models can achieve over 99% accuracy in brand identification across twenty-eight commercial samples. To demonstrate the flexibility of this approach we utilised the same samples and algorithms to quantify ethanol concentrations, as well as measuring methanol levels in spiked whisky samples. Our machine learning techniques are then combined with a through-the-bottle method to perform spectral analysis and identification without requiring the sample to be decanted from the original container, showing the practical potential of this approach to the detection of counterfeit or adulterated spirits and other high value liquid samples.

Roopam K. Gupta, Graham D. Bruce, Simon J. Powis et al.

The speckle pattern produced when a laser is scattered by a disordered medium has recently been shown to give a surprisingly accurate or broadband measurement of wavelength. Here it is shown that deep learning is an ideal approach to analyse wavelength variations using a speckle wavemeter due to its ability to identify trends and overcome low signal to noise ratio in complex datasets. This combination enables wavelength measurement at high precision over a broad operating range in a single step, with a remarkable capability to reject instrumental and environmental noise, which has not been possible with previous approaches. It is demonstrated that the noise rejection capabilities of deep learning provide attometre-scale wavelength precision over an operating range from 488 nm to 976 nm. This dynamic range is six orders of magnitude beyond the state of the art.