** AWARDS ISSUED PRIOR TO JANUARY 20, 2025, WERE FUNDED UNDER PREVIOUS ADMINISTRATIONS AND MAY NOT REFLECT THE PRIORITIES AND POLICIES OF THE CURRENT ADMINISTRATION.** BETTER DATA ANALYSIS MEANS BETTER FOOD SAFETY. IN RECENT YEARS, ARTIFICIAL INTELLIGENCE (AI)-POWERED SOLUTIONS ARE EMERGING IN AGRICULTURE AND FOOD SCIENCE DUE TO THEIR SUPERIOR PATTERN RECOGNITION AND PREDICTION CAPABILITY. INTEGRATION OF FOOD SCIENCE AND AI CAN HELP ADDRESS NEW CHALLENGES SUCH ASINCREASING VOLUME OF DATA ACQUIRED FROM MIXED SAMPLES AND COMPLEX FOOD MATRICES, WHICH POSES CHALLENGES TO DATA ANALYSIS USING TRADITIONAL METHODS.THE OVERALL GOAL IS TOINTEGRATE NOVEL MACHINE LEARNING (ML) METHODS, SUCH ASATTENTION-BASED DEEP NETWORKS,WITHSURFACE-ENHANCED RAMAN SPECTROSCOPY (SERS) PLATFORM FORMULTIPLEXDETECTION AND QUANTIFICATIONOF FOOD CONTAMINANTS WITH HIGH ACCURACY.SPECIFIC OBJECTIVES ARE TOSYNTHESIZE NANOSUBSTRATES AND ACQUIRE SERS SPECTRAL DATA OF DIFFERENT TYPES AND QUANTITIES OF PESTICIDES BY SERS MEASUREMENT;DEVELOPATTENTION-BASEDDEEP LEARNING PREDICTION METHODS FOR QUALITATIVE AND QUANTITATIVE ANALYSIS OF SINGLE FOOD CONTAMINANTS AND MULTIPLE/MIXED FOOD CONTAMINANTS; VALIDATE AND ASSESS SERS-ML TECHNIQUES FOR MULTIPLEX DETECTION OF CHEMICAL CONTAMINANTS IN FRESH PRODUCE; AND ESTABLISH PROTOCOLS/DATABASES FOR MEASURING FOOD CONTAMINANTS.THIS STUDY WILL BE THE FIRST SYSTEMATICINVESTIGATION OF PESTICIDES IN FRESH PRODUCE BY SERS COUPLED WITH MACHINE LEARNING ALGORITHMS, WHICH WILL BE MORE ACCURATE, SENSITIVE, AND RELIABLE THAN CURRENT METHODS. THE PROJECT WILL BROADEN THE APPLICATIONS OF DATA SCIENCE IN MAINTAINING THE SUSTAINABILITY OF U.S. AGRICULTURE AND FOOD SYSTEMS.