Note
Technology that understands human change, for people.
Heartbeats, sleep, and voice. From biosignal research to technology that fits individual lives.
Our bodies change from day to day: while we sleep, while we walk, and while we talk with someone. Some of that change appears in heart rate and sleep records, or in a voice. We pay attention to those small clues.
At Lumo, we research and develop ways to interpret biosignals and understand how people change. We want to expand what we can understand, step by step, while being honest about what remains unknown. That steady work is how we hope to make technology useful in everyday life.
Read the signals with care.
The intervals between heartbeats vary. We work on extracting measures of that heart rate variability, or HRV, and studying how they change over time, including during sleep. Our work with voice also includes detecting speech segments and extracting features from sound.
Sensor records contain noise and gaps. First we check the quality of the signal and determine which parts can support analysis. This careful preparation, before a model is built, is an essential part of the research.
Start with that person’s own usual.
The same number can mean different things depending on the person and how it was measured. We focus on understanding someone’s usual range through records gathered over time, then considering changes in relation to that personal reference.
The time of measurement, sleep, and activity offer context for interpretation. We consider signals alongside what came before and after, as well as the person’s own experience. We investigate which combinations help us account for individual differences.
Make each hypothesis testable.
We use signal processing to extract features, then statistical and machine learning models to examine relationships. We check whether the same procedure reproduces a result and whether it holds for other people or conditions. How data is split and what methods are compared are part of that evaluation.
When results do not match our expectations, we record the conditions and limitations and use them to shape the next question. We want the work to explain what has been established and what remains open, beyond any single promising result.
Be open about what remains unknown.
A number reflects one aspect of a person. Heart rate alone cannot tell us what someone feels or why they feel physically unwell. Sparse records and poor signal quality also change how much we can reasonably interpret.
We aim to distinguish observations from model estimates and communicate how certain an interpretation is. We do not promise benefits beyond what the research has established. We see that honesty as part of the quality of technology that concerns people.
Keep the person’s choices at the centre.
Our design starts with what information is needed and why, and with keeping analysis on a person’s device where it can be done there. We want people’s choices to shape how information is collected and retained. Those principles matter from the moment we choose a technical approach.
How people focus, experience things, and use their strengths varies. Neurodiversity is one of the perspectives we bring to designing explanations and interactions around those differences. We want technology to leave room for people to choose at their own pace.
We want research to offer a starting point for understanding yourself and choosing how to spend your time. Clear language and a form you can turn to when needed are part of the technical work, too.