logseq/journals/2022_09_12.md

• PAPERS/Software-ecosystem-book-chapter
  • Lavorare su sezione presentazione processo
• #Highlights
  • Machine Learning is a method to find a relationship between different sets of values by utilizing ==statistical models== and ==algorithms==.
  • ((631f19cb-c451-4e9b-bb59-1d22d7bac8b2))
  • ((631f1a34-3f81-4a1b-9fcb-29b8d39c020e))
  • ((631f1aa1-6e38-4cb4-a283-150b6fac9e58))
    • ML courses, videos, and websites id:: 631f2025-9cad-413a-8fcf-c1c78bd68635
• type:: meeting external-links:: people:: #people/alfonso tags:: PAPERS/Software-ecosystem-book-chapter date:: 12-09-2022 duration-min:: 73
  • MDE for Digital Twin (Schloss Dagstuhl : Seminar Homepage) Ideas
    • Lavoro di Andreas che ha fatto qualche survey
      • ci sono diverse componenti tra cui quella prescrittiva
      • Anomaly detection
      • Predictive maintenance
      • Dataset disponibili?
        • Petrobras ha messo disponibile dei #dataset
          • UCI Machine Learning Repository: 3W dataset Data Set
            • ricardovvargas/3w_dataset: The first realistic and public dataset with rare undesirable real events in oil wells. (github.com)
            • @A realistic and public dataset with rare undesirable real events in oil wells
            • Da vedere
              • Machine learning and transport simulations for groundwater anomaly detection - ScienceDirect (cineca.it)
              • Imminence Monitoring of Critical Events: A Representation Learning Approach | Proceedings of the 2021 International Conference on Management of Data (cineca.it)
        • Disponibilità di simulatori
          • Ventilatore meccanico
    • Domini consolidati
      • Manufatturiero dove hai una controparte fisica / meccanica
      • Sistemi umani (es. medicina di precisione, agricoltura di precisione)
  • Aggiornamento book chapter
    • Condiviso il paper @AI-driven streamlined modeling: experiences and lessons learned from multiple domains con Alfonso. Ha detto che se lo vedrà per Venerdi 16 Settembre
• type:: meeting external-links::
  tags::
  people:: #people/Vittorio_Cortellessa #people/traini #people/riccardo #people/imran tags:: date:: 12-09-2022 duration-min:: 60
  • Benchmarking process
    • 4 different steps
      • Identify code segment
        • Recommender systems for recommending parts of software systems that should be benchmarked
          • Mixing static and dynamic analysis
          • Related work
            • AutoJMH (2016)
            • J23jmh (2022)
      • Wrap code into payload
        • remove code smells (of the code we want to benchmark)
          • Smells that affect the performance
            • We need to be cautious on this because we might introduce some bias or in general, affect the behaviour of the original system
          • Dead code identification
      • [...]
      • Execution large # of times
        • Reduce time to execute tests
  • Ideas
    • Identify features (static or dynamic) of existing projects that can underpin the recommender system
      • The goal is to recommend code locations that need to be benchmarked via JMH
    • projects.csv