AI Assistant for Process Plant Operations



AI-based Decision Assistant for Process Plant Operations (ADAPPO or Plant Assistant in short) is an applied research project conducted as a collaboration between GippLab and eschbach GmbH funded by ZIM (Zentrales Innovationsprogramm Mittelstand) run by the German Ministry of Economic Affairs and Climate Action.

NLP-driven Plant Assistant aims at creating a tool that supports plant operators at their daily operations on a industrial processing plant. The Plant Assistant aggregates knowledge and experience from logged plant operations and provides fast, efficient, and interactive feedback when users need solutions to encountered problems. A goal of the project is to leverage the recent advances in NLP to tailor language models towards domain-specific text data of multiple languages with uneven data quality, data scarcity, and lack of annotated sources.

Plant Assistant involves the following NLP research and directions:

  • Domain adaption of language models
  • Coreference resolution
  • Extractive Question & Answering
  • Text summarization
  • Automated annotation of data corpora
  • Semantic Information Retrieval
  • Domain-specific sentiment analysis
  • Model improvement with a human-in-loop

Check out our ongoing research topics available here.