Applying Machine Learning Techniques for Sensitive Spectral Identification and Detection of Hazardous Target Molecules

Natural, extreme climate change, and human-induced disasters bring together teams of scientists and practitioners supported by several agencies, such as NOAA, NASA, EPA, and FEMA. These scientists and practitioners work at monitoring for conditions leading to, working through, and recovering from disastrous events. They currently use state-of-the-art sensors capable of detecting hazardous gases and chemical hazardous materials, which these sensors are individually calibrated for.Such sensor devices are in effect mobile lab stations, miniaturized, and target specific chemical toxin detectors.For the phase 1 study, Caelum has assembled a team of physicists and data scientists with experience in spectrometry and full-stack AI solutions development. Caelum Research Corporation has selected Spectromatics, LLC for its specific expertise as a subcontractor. We will develop an Artificial Intelligence (AI) platform to augment and integrate into currently existing commercial chemical detectors produced by such firms as Honeywell, Ventis and ALTAIR. Our goal is to significantly increase the capacity and proficiency of these devices by adding AI capabilities through computational spectral analysis. We have aligned our vision with SBIR’s three phases. Phase 1: feasibility of spectral data and AI algorithms, Phase 2: functional prototype, and Phase 3: operational commercialization. Such computational analysis using AI increases the flexibility, capability and adaptability of toxin detection, as new toxins are discovered throughout the platform’s lifecycle. Our comprehensive and versatile AI algorithm, once implemented, will help multiple agencies and many of their commercial partners in the field of detection and mitigation of chemical hazards.