Norman “John” Mapes Jr. Ph. D. is currently a research and development scientist in the nuclear science and technology directorate at INL. He has over 20 years of industrial and academic experience in computational modeling and reporting. In a 15-year period he worked on semi-automated asset integrity. He examined and reported on infrastructure, aviation and plant equipment utilizing electronic, chemical, ultrasonic, RF, microwave, visual, magnetic, and radiographic devices. He has been involved in programming artificial intelligence and advanced analytics into reporting, for example the visualization of the settlement of 80,000-barrel chemical tanks and GPR for locating underground and embedded objects. Since the mid 2010’s, he has focused on discovering previously unknown and potentially useful information through cyber engineering and data mining, especially relating to unstructured data such as the natural language found in text field reports. He has researched natural language modeling and pattern recognition themes including large-language models, summarization, text classification, topic models, semantic searching, question answering, seq2seq, masked LM, transfer learning, fine tuning, pretraining, non-negative matrix factorization, biclustering, document embeddings, regression, gradient boosting, DNA, and protein sequence analysis.

Large language models, automation, computational modeling, data mining, pattern recognition, bioinformatics and asset integrity.