Researchers at Aalto University have developed a new predictive typing model that can simulate different types of users to optimize the way we use our phones. Typing on a phone involves manual dexterity, visual perception, working memory, and proofreading, and automatic text correction can either help or hinder the typing process. The machine-learning model created by the researchers acts like a human, making mistakes and correcting them as it types out sentences.
The simulated user created by the researchers has a human-like visual and motor system and was trained millions of times in a keyboard simulator to learn typing skills that can be applied in various real-world situations. The model, developed in collaboration with Google, aims to complement the costly and time-consuming real user testing typically used to evaluate new designs for phone keyboards. By using this model, keyboards can be evaluated and optimized more quickly and easily, saving time and resources.
This new predictive typing model is part of a larger effort by Professor Antti Oulasvirta and his research group at Aalto University to improve user interfaces and understand human behavior in task-oriented situations. By using computational models of human behavior, the group aims to create a more functional society and smoother everyday life by training computer models to make predictions without requiring observation of many people. Ultimately, the goal is to improve user interfaces in all aspects of modern life.
The researchers will present their findings on the new predictive typing model at the CHI Conference in May, showcasing how this technology can revolutionize the way phone keyboards are designed and optimized. By simulating different types of users and capturing the differences between typing with one or two hands or between younger and older users, the model can help identify ways to enhance the typing experience for all users.
By training computer models to act like humans and make predictions based on user interactions, the research team at Aalto University is paving the way for more efficient and effective user interface design. By understanding how different users interact with technology and how to optimize their experience, the researchers are contributing to the creation of a more functional society where everyday tasks are smoother and more enjoyable. The predictive typing model developed by the researchers has the potential to revolutionize the way phone keyboards are evaluated and designed, saving time and resources in the process.
