Identifying and Enabling Emerging Technology Leaders

1. Collaborate with Army to summarize a corpus data set of representative selection criteria for Army Emerging Technology Leaders (ETL):WSI team develops the “fingerprint” for Army ETL requirements including the Army Modernization Plan. Use the Elsevier Fingerprint Engine (FPE) with machine learning (ML) and Artificial Intelligence (AI)-based techniques to analyze the unstructured text of ETL requirements. Output consists of a “word map” ontology and “fingerprint” of ETL requirements with key terms and associated weighting. 2. Develop representative metadata: Identify a sample pool of WSI’s retired military officers with technical degrees and experience. Gather novel personnel data that mirrors Army metadata for candidate ETLs. Identify and define unstructured textual information such as resumes, ASVAB and other test scores, evaluation reports, career selection information (e.g., AIM), and unit assignments. Strip out all Personally Identifiable Information (PII) and obtain individual release approval for novel metadata. Develop a draft ontology to fingerprint the metadata.  3. Develop a construct to quantify cognitive and values-based attributes of ETLs using behavioral analytics tools (VIZEON): Identify subjective descriptors of ideal ETL candidates. Use those descriptors, WSI’s online VIZEON assessments, and associated WSI cognitive behavior algorithms to create a template to guide evaluation of the employee sample pool. Use Elsevier FPE to fingerprint cognitive attributes of the sample pool and develop a VIZEON attribute ontology. 4. Develop a construct for text extraction, followed by Natural Language Processing (NLP) and semantic profiling of unstructured metadata text to mine the novel metadata: Use NLP, data mining, text mining, and information retrieval processes to develop an AI-based construct of semantic search fingerprints for the novel (sample pool) metadata. Develop initial FPE models through analysis of novel metadata to assess common data elements across the representative (sample) group of personnel. Tune the FPE model and associated ML to accommodate all types of unstructured text including Army HR modules, resumes, transcripts, evaluation reports, and publications.  5. Develop a construct to map fingerprints of personnel metadata to Army ETL requirements: Using the ontologies, cognitive assessments, and fingerprints created in Tasks 1-4, develop an AI-driven ML model to map personnel data to ETL requirements. Identify and catalog key phrases that emerge from the novel metadata to weight terms and evaluate their relative strength. Create a graduated ranking of personnel data fingerprints against the weighted terms. Use a bracketing approach to achieve an 80th – 90th percentile probability of best fit for ETL candidate selection using key terms for the requirements ontology and key terms from the personnel fingerprints. Develop an AI-based construct to automate this cognitive computing process. 6. Develop/present Phase I deliverable.