Technology

We are passionate about assisting clinicians and doctors to transform the diagnosis and treatment of diseases through machine learning systems.

WSK Medical is an innovative software company focused exclusively on developing leading edge machine learning applications for the medical community. Our aim is to bring you the most advanced AI technologies to assist clinicians and doctors in the early detection and classification of cancerous tumours. We do this by working with you in partnership, to develop sophisticated image recognition systems.

AI - More than a movie!

Artificial intelligence is a branch of computer science that aims to create intelligent software and machines. As such, it has become an essential part of the Information Technology industry. Research associated with artificial intelligence is highly specialised. The core problems of artificial intelligence have traditionally included programming computers for certain tasks such as: * Knowledge acquisition * Reasoning * Problem solving * Perception * Learning * Planning * Robotics.

zeno pathology

From Engineering to Learning

Traditional AI focused on Knowledge Engineering. This is still an important part of AI. Machines can often act and react like humans only if they have abundant information relating to the world. Artificial intelligence systems must have access to objects, categories, properties and relations between all of them to implement knowledge engineering. However, programming knowledge, common sense and reasoning in machines has proven to be a difficult and tedious task. And this is where Machine Learning comes in.

Machine Learning

Machine learning has become a core part of AI. Learning without any kind of supervision requires an ability to identify patterns in streams of data inputs, whereas learning with adequate supervision involves classification. Classification determines the category an object belongs to. Mathematical analysis of machine learning algorithms and their performance is a well-defined branch of theoretical computer science often referred to as computational learning theory.

zeno pathology

The Challenge

To develop practical autonomous learning systems has been a challenge for AI developers. A challenge that’s been made a good deal easier now that computing power has increased to a level where complex AI solutions are now cost effective to deliver within realistic time limits. WSK Medical leverages these technologies to deliver tailored AI based solutions to its clients.

zeno pathology

Explainable AI (xAI)

At WSK Medical we believe in transparent, explainable and traceable AI software solutions, from design to decision-making. As AI solutions are becoming an important part of the digitalisation of health care, we apply modern methods of Explainable AI (xAI) to create transparent and explainable AI solutions. With these methods, we can provide our users with the required guidance and therefore create the perfect human-machine collaboration.

We build commercial bespoke machine learning systems. From prototype to production.

Our approach is to use two methodologies that allow us to follow an end-to-end process – Agile and CRISP-DM. They both help us to achieve data mining and software development. Agile allows for rapid software prototyping, while staying customer focused. CRISP-DM (CRoss-Industry Standard Process for Data Mining) ensures we deliver a deep learning model.

1. Problem Understanding

The key phase for success. Understanding the problem from a user perspective. Assessing the success criteria and the clinical environment from a perspective of requirements, risks, benefits, resources, and costs.

2. Data Understanding

This is all about collecting the data and quickly confirming if we have data quality issues. Occasionally initial data does not always fit the problem that is being solved. This will mean assessing the structure and quality of the data. Typical questions, “Is the data complete, is there any missing values, have we explored the data enough”.

3. Data Preparation

Data cleansing and preparing the dataset. A key step before the modelling phase which may involve the following: Selecting the data that is most relevant for the data mining objective; Cleansing the data to ensure the quality of the data is correct; Reconstructing the data to ensure it meets the requirements of the model.

4. Modelling

Run the data mining tools. Once the problem is sufficiently understood, then we select a suitable model for supplying training data to the learning algorithm.

U

5. Evaluation

Determine if the results meet the project objectives; Identify project issues that should have been addressed earlier before we go into the deployment phase.

6. Deployment

Deploying the algorithm into a production environment. If an algorithm is being deployed into a larger system environment, then additional software engineering aspects have to be considered. Such as reliability, security, hardware & operating system requirements. Which is all outside the data mining tasks outlined.