Project 1: In this project we aim to develop deep learning algorithms to identify and classify blood diseases directly from data gathered by photonic sensors. To do so, we will make use of the raw data provided from our sensors when light interacts with the substances in place and we will explore what information is encoded in these signals via algorithms such as variational autoencoders, multiple instance learning, U-NETs, etc. We will use these algorithms to make a decision whether a particular disease is present in the biological sample.
Project 2: In this project we will implement advanced signal processing solutions in order to improve our current algorithms for classification of blood diseases directly from data gathered by photonic sensors. Examples of the approaches we will explore are wavelet decomposition, compressive sensing, random matrix theory, PCA, UMAP, and more. Our objective is to clean the signals and extract meaningful information that will allow us to identify features in the samples that are linked to the presence of different diseases.
Project 3: At iLoF we use photonic sensors to gather information from the light-matter interaction occurring when a laser beam is sent to a biological fluid (e.g. serum, plasma, blood). This light-matter interaction is dominated by scattering of photons within the nano and micro elements present in the sample. In this project we will develop numerical statistical models that are based on this interaction and that will allow us to generate high volumes of synthetic data to potentially train our machine learning algorithms and therefore accelerate our platform.
Contact: Alex Turpin, Head of Biosignal and Analytics at iLoF
EMAIL: aturpin@ilof.tech