Data Science

/ Software Development / Data Science/

What is Data Science?

Data science combines algorithms, tools, computer programming and machine learning to discover information within datasets. Machine learning is a type of artificial intelligence that trains a machine to learn from past experience. A subset of machine learning that Cerberus Nuclear regularly uses is called deep learning, where the machine makes more intelligent decisions based on a neural network (which is based off the workings of the human brain).

Our Capability

At Cerberus Nuclear, we’re working hard to utilise our data science and computer vision expertise, combined with our in-house machine learning capabilities and optimisation experience to provide nuclear data science solutions and training for bespoke applications. We mainly use Python with Libraries such as Tensorflow and OpenCV to develop our data science products.

CASe study 1
Criticality Training Aid, CARTA

CARTA is an immersive and realistic criticality safety training aid, which focusses on counter-intuitive interactions. It is a highly adaptable system that can be tailored to create bespoke training scenarios.

CARTA uses a novel machine learning approach comprised of an Artificial Neural Network (ANN) that takes into account the multiple variables to produce an accurate value for k-effective.

CASe study 2
Automated Nuclear Damage Inspection, ANDI

We have built an innovative program for Sellafield Ltd that uses deep learning (a Mask R-CNN) and computer vision to automatically detect damage and defects in inspection footage. The neural network is trained off previous example images of damage, such as scratches, dents and abrasion.

ANDI is a high-quality use-friendly software program, and has been featured as a Sellafield Ltd success story, read more about it here.