Sensing specific molecules is essential for life and forms the basis of many diagnostic technologies. Often, this is achieved by one sensor-one analyte lock-and-key mechanisms, e.g., by antibody-antigen binding. This project develops materials for an alternative approach to sensing based on an artificial perception system. This system comprises a set of nanosensor elements chosen from the family of DNA/Single-Walled Carbon Nanotubes (SWCNT) and will work on bodily fluids. Each nanosensor element (a particular DNA/SWCNT combination) will have only a weakly specific response to an analyte or physiological state. However, acting in concert with machine learning (ML) techniques and high throughput experimental interrogation, a suitably designed nanosensor array can accurately detect or measure multiple analytes or physiological states in biofluids. A compelling feature of this approach will be the ability for a nanosensor array to be designed first and primarily by the choice of nanosensor elements and secondarily by the use of machine learning algorithms. Because the nanosensor array will be built as initially analyte-agnostic, the optimally designed sensor array has the potential to be a universal biofluid sensing system capable of diagnosing many diseases.