DIGITAL LIFE

Scientists develop virtual simulator capable of typing like a human

A group of researchers from Aalto University, in partnership with Google, developed a virtual simulator capable of typing like a human. The predictive typing model can distinguish between different ways of using the cell phone keyboard, whether with one hand or two, and even replicates the way young and old people type.

The project's proposal is to support the development of interfaces that best adapt to human movements, making, for example, cell phone use more practical and optimized.

When you click on the letters on your cell phone keyboard, you are using manual dexterity, visual perception and working memory at the same time.

Simply put, you need to move your fingers precisely while viewing and reviewing what you write.

The simulator has virtual “eyes and fingers” and memory to have the same level of skills as you. This means that you also make the same mistakes.

He has been trained thousands of times to absorb the human typing style, from agility, movements and even their specific difficulties.

For example, some people hate using autocorrect, others can't live without it. Technology also analyzes and applies this.

The model is especially useful for interface development tests, which typically only use real users.

Interface Improvements...Traditionally, the process of testing new keyboard designs involves a significant investment of time and resources, as it requires the participation of real users. Therefore, the project seeks to complement these tests with a digital simulator. This way, it would be possible to evaluate keyboards more quickly and efficiently.

Antti Oulasvirta, leader of a research group at Aalto University, believes this initiative represents a crucial step towards improving user interfaces and understanding human behavior in task-oriented situations. Using computational models of human behavior, his team is committed to investigating fundamental questions related to human-machine interaction.

''We can train computer models so we don't need to observe many people to make predictions. User interfaces are everywhere today – fundamentally, this work aims to create a more functional society and a smoother everyday life''...Said Antti Oulasvirta

Next month, the research team will present the project at the CHI Conference, a forum focused on the area of human-computer interaction.

mundophone