Software Engineering for Machine Learning: Characterizing and Detecting Mismatch and Predicting Inference Degradation in ML Systems

Abstract

An area of work within the SEI is developing practices, methods and tools for reliable end to-end development, deployment, and evolution of AI-enabled systems. Our goal is to develop empirically validated practices to guide AI engineering and support software engineering for machine learning (SE4ML) systems. This webinar reports on two focus areas: Characterizing and Detecting Mismatch in ML-Enabled Systems Predicting Inference Degradation in Production ML Systems

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2021
Accession Number
AD1146933

Entities

People

  • Grace Lewis

Organizations

  • Carnegie Mellon University

Tags

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Artificial Intelligence Software
  • Computer Programming
  • Data Science
  • Data Sets
  • Degradation
  • Department Of Defense
  • Detection
  • Engineering
  • Environment
  • Governments
  • Literature Surveys
  • Machine Learning
  • Programming Languages
  • Retraining
  • Software Development
  • Software Testing
  • Training

Fields of Study

  • Computer science
  • Engineering

Readers

  • Computational Modeling and Simulation
  • Neural Network Machine Learning.
  • Software Engineering.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • AI & ML - Neural Networks