inCognita Data Science components
Volv has additional capabilities that make up the inCognita methodology, which enrich the insights and learnings that the core predictive modelling produces.
For example, we have created a new method for leveraging passively-collected smartphone data, and performing machine learning on this to detect and monitor brain disorders such as depression and Parkinson’s disease.
inCognita AI & Data Science
Volv’s approach is unique and provides state of the art performance, defining disease state more accurately than any other current known method.
The ubiquity of smartphones in modern life suggests the possibility to use them to continuously monitor patients, for instance to detect undiagnosed diseases or track treatment progress.
Such data collection and analysis may be especially beneficial to patients with
Volv has developed a new method of leveraging passively-collected smartphone data and using machine learning to detect and monitor brain disorders such as depression and Parkinson’s disease. Crucially, the algorithm is able learn accurate, interpretable models from small numbers of labelled examples, for example from smartphone users for whom sensor data has been gathered and disease status has been determined. Predictive modelling is achieved by learning from augmented training sets comprising both real and ‘synthetic’ patients. This approach is shown to outperform state-of-the-art techniques in experiments involving disparate brain disorders and multiple patient datasets.
The Volv approach is shown to outperform state-of-the-art techniques in experiments involving disparate brain disorders and multiple patient datasets.
Volv is exploring the applicability of the approach to other disorders.
Given consent to access passively-collected smartphone data for a patient and with a wish to predict whether the individual has a brain disorder of interest. A supervised learning approach to this task is challenging. We can see that the objective entails:
Unfortunately, in medical passive-sensing applications there are rarely enough labeled examples to support induction of good, generalizable models.
Volv has constructed accurate, interpretable models for detecting target brain disorders, even in settings with few labeled training examples, by using machine learning guided by domain knowledge. The predictive modeling process integrates three steps:
Given passively-collected smartphone sensor data and associated TD diagnoses for a (small) set of patients, the goal is to learn a model capable of accurately predicting the probability that a new patient is suffering from the TD. The focus on depression and Parkinson's Disease (PD) as brain disorders of interest leads to the following specific prediction targets, determine whether an individual:
Volv has successfully built robust models to determine these outcomes.
Learning from Synthetic Patients
In some applications, the labeled training data is extremely limited and/or class-imbalanced, and this is an obstacle to effective learning. This situation arises, for example, because it is often difficult to collect data traces for individuals with confirmed diagnoses for the Target Disease (TD) of interest, and because typical datasets have far fewer ‘cases’ (patients with the TD) than ‘controls’ (unaffected patients). We therefore use synthetic patients to address the issue of class imbalance.
inCognita - continuous monitoring of mental disorders and progressive neurological diseases
inCognita helps detect signals related to mental disorders and progressive neurological diseases. inCognita brings real-time monitoring that can facilitate earlier diagnosis and more effective treatments. inCognita is a unique data science framework that Volv have developed.
