Silicon Studio has created a platform known as Yokozuna Data: a state of the art machine learning engine that can utilize next generation artificial intelligence algorithms to predict individual player behavior – scale to big data – and provide an intuitive, unique simulation for each user.
Welcome to the digital age of gaming!
If you are an avid gamer, chances are you’ve had the pleasure of experiencing life-like cinematography enhanced by the cutting edge technology of Silicon Studio.
Based out of Tokyo and sporting a very diverse and experienced team, Silicon Studio was established with the primary goal of delivering state of the art innovation to the digital entertainment technology industry.
They have achieved this by specializing in rendering technology, research and development methods, game content development, online gaming solutions, and most importantly they are behind one of the first post-processing middleware technologies known as Yebis – one of many tools they have in their arsenal.
Currently on its third version, Yebis 3 allows developers to create high-quality lens-simulation optical effects – basically make games, such as Final Fantasy XV look like they are from the future.
Silicon Studio’s Newest Project: Yokozuna Data
Under the lead of Dr. África Periáñez: a senior data scientist (Head of Game Data Science of Silicon Studios) and researcher with 12+ years of experience, gaming is about to evolve in ways we can only begin to imagine.
Dr. Periáñez holds a PhD in Mathematics from the University of Reading, a MSc in Theoretical Physics from the Autonomous University of Madrid, a Master of Advanced Studies in String Theory Physics carried out at CERN (under a Marie Curie Excellence grant). She also has significant industry experience at SPSS Inc., IBM and now Silicon Studio.
For those of you who are unfamiliar, CERN has been responsible for much of the technological advancement our species as seen over the last 3 decades, including the world wide web.
I had the rare opportunity to interview both Vitor Santos and Dr. África Periáñez, in which they explained how their technology is going to revolutionize the gaming industry.
Her passion as a data scientist to further understand social phenomenon combined with her background in satellite data, mathematics, and string theory has created the platform Yokozuna Data (named after the highest achievable rank in sumo wrestling), and with it a state of the art algorithm that is beyond anything her competition is developing.
Dr. Periáñez’s algorithms are unique, in that they combine state of the art deep learning with something called a survival ensemble.
The ensemble method so far has only been utilized in the field of biology, particularly in attempting to understand behaviors in for example breast cancer – at the individual cell level.
What ensemble means is that instead of there being just one algorithm, there is a cloud of learning algorithms that break down the information presented into separate parts, allowing it to take a multi-dimensional approach, scaling with big data, examining more correlations and alternative models, thus drastically improving prediction rates.
Imagine a tree with roots.
None of her competition is utilizing this technique.
She is the first to incorporate the ensemble design into the field of gaming. Yokozuna Data’s algorithm is so accurate, it won both the Churn Prediction Track and Survival Analysis Track at the NCSoft’s Blade & Soul Game Data Science competition at the IEEE 2017 Conference on Computational Intelligence in Games.
Simply put, her algorithm knew precisely how long people were going to play, exactly when they would leave, where (which level of the game) and why.
Her algorithm will calculate exactly what is needed in the game to tempt you the most into following the path of becoming a whale.
The more you play, the more it adapts to you.
This actually works to create a balance between what are known as free to play players vs. the wide ranging demographic of those who are willing to spend money to gain more advantage.
Both the gamers and the companies that produce the games win. Gamers don’t have to spend money in the game if they don’t want to and still progress. Corporations maximize profit.
Scalable to Big Data:
Until this interview, I was under the impression social media data was the sandbox of data scientists. It makes perfect sense hearing her describe how within video games, every click, every action we take – there is so much more of your personality being digitally transferred.
Gaming companies have amassed petabytes of player data, as every action, in-app purchase, guild conversation or in-game social interaction performed by the player is recorded.
Surprisingly and against what I had always believed, most of them do nothing with this information.
It is no easy task obtaining and storing this kind of data, therefore they horde it.
Silicon Studio wants this kind of knowledge accessible to anyone. They are leading the revolution in democratizing game data science between these companies so that not only can all games benefit, so too can research into human behavior.
They are not creating their own database, they are merely breaking down borders between already existing databases as they continue to forge new relationships.
This is exactly how A.I. and gaming should progress with each other. Companies should not be withholding data, but rather sharing it so all of mankind can learn from it and benefit.
Simply put, Silicon Studio is exactly the kind of company we want to taking gaming into the future – and thanks to them, that future is here.