Byowave Ltd.
Nominated Award: Best Application of AI in a Startup
Website of Company: https://byowave.io/
ByoWave is a Galway based startup company founded in June 2020 by cofounder s Brandon Blacoe and Eibhlin O’Riordan. Both share a passion for gaming and when they were children used to play ‘Shrek 2’ together on the PS2. They went their separate ways, Brandon went to study Electronic Engineering at UC D and Eibhlin studied Game Design at DIT. In May of 2020 they reconnected and Brandon found out about a problem which Eibhlin had.
Eibhlin has a connective tissue disorder called Ehlers Danlos Syndrome (hEDS). Like all impairments, Ehlers Danlos Syndrome affects everyone differently, for her hEDS causes her a lot of pain, fatigue and a multitude of other issues. When she was 14 she injured a nerve in my right hand and due to her hEDS it did not heal properly and she developed a painful nerve condition called Complex Regional Pain Syndrome (CRPS).
While her hEDS was challenging it was not until she developed CRPS that she was unable to use my hand for any extended period of time without experiencing extreme pain. She was suddenly unable to do many of the activities she loved to do and she wasn’t able to use generic video game controllers. Gaming was and is a huge part of her life, it is an escape from her chronic pain. She spent man y years and a lot of money trying to find an accessible alternative to solve this problem, but she could not find a controller that worked for her.
She just wanted to be able to build her dream controller and believed that her inability to play games was for no other reason than that gaming was inaccessible and had nothing
to do with her disability. While in search of a solution, she also found many others just like her, who we re struggling to play games due to their disability who felt the same.
Brandon took the idea of allowing disabled gamers to build their own controller and made it a reality. They combined Brandon’s skills and expertise in biomedical and electronic engineering as well as his knowledge and passion for start-ups with Eibhlin’s personal experiences with the problem and her knowledge of game design, and together they started ByoWave.
Since then, ByoWave has come a long way. Through the research and development stage they talked with hundreds of potential customers from around the world and came up with the idea or a customisable gaming controller kit. This kit is a modular selection of small cubed blocks and peripheral buttons and allow for a completely customisable controller that the user can design for themselves. In March of 2021 I (Ronan Murphy) joined the team as CTO after completing a Masters in Artificial Intelligence to develop the controller further and an ecosystem for our community through a number of software services. We are preparing for our soft launch where the Proteus Controller will be available to selected users at the end of October 2021.
Reason for Nomination:
The reason we believe Byowave should deserve the award for the AI in a startup is because of the software solutions we are providing to our community. We are using two solutions – a Controller Matchmaker and an Analytics Dashboard.
The Controller Matchmaker is a recommender system which predicts the optimal controller configuration for a user based on a survey they complete, when they enter our controller builder. The features of the survey include their disability type, age, console, and game they intend to use the controller for. Our controller builder saves the configurations that each user makes and stores their survey details to our NoSQL database. Based on the details a user enters, our Matchmaker outputs a controller configuration prediction. Our AI model outputs this prediction based on the most similar users to them. Our recommender system is a model based collaborative based filtering method. The model finds the most similar interaction profiles or other users from matrix factorisation. As more controllers get created by users our model will be better able to find similar users based on their categorisation. The Matchmaker outputs the top 3 most similar controllers for the user which they can interact with and edit for further customisations. The overall goal of this system is to reduce the time it takes a user to build their own controller by learning from users similar go them which greatly enhances our user experiences.
Our second AI service is our Analytics Dashboard. We gather primary data directly from our users controllers. This data includes statistics such as button presses, latency, and in
game performance. This information is stored to our backend also and then using a data pipeline cleaned, filtered and sorted for individual user preferences. We created an ensemble model to compare the highest performing users for each game and their configurations to predict the most important features of the optimally performing controllers for their specific categorisation (i.e. console, game, and disability if applicable). This allows users to enhance their performance by comparing themselves to the highest performing users in their categorisation.
The method of feature importance inspection uses shapely values, and is a game-theoretic approach to assessing how a model explains it’s predictions. When the model predicts the efficiency score for the controller configuration, this prediction can be explained based on the contributions of the different features which contribute to this prediction. For example, using block A in combination with the other blocks yields a prediction of a lower efficiency score, where not using block A in combination with others yields a prediction of a higher efficiency score. By explaining what combinations of what features (in our case features = module blocks) used in what amounts yield to a high er or lower gamer efficiency score, a player can adjust their controller configuration based on what modules are contributing to high efficiency the most, and remove modules than contribute to lower efficiency. This allows the user to improve their gaming performance by adding blocks that in their case improve the prediction of higher efficiency and remove blocks that in their specific case reduce the predicted score for efficiency.
Additional Information:
Mission statement of Byowave:
Empowering our community to build their own solutions through our precision engineered assistive technology ecosystem. Initially our products and services will deliver for the global gaming market.
Vision statement of ByoWave:
Empowering communities to bridge the gap across the digital world through our magical user experience and world class engineering without exceptions.