Course content

This course provides an overview of practical applications of machine learning (ML) in industrial contexts, focusing on goal setting, use of real data, domain knowledge, and problem-solving strategies. Students will gain practical experience with large, complex datasets, with examples of classification, regression, and clustering, and learn to handle uncertainties inherent in ML issues.

The course emphasizes the importance of subject and domain knowledge to understand data quality, measurement uncertainty, noise sources - and how these affect model accuracy and predictive performance. Participants will learn to define clear objectives for machine learning projects, aligned with the needs of stakeholders and users. The curriculum covers best practices in analysis evaluation, interpretation of results, and parameter adjustment to optimize model performance.

Upon completion of the course, students will possess the skills to effectively apply machine learning techniques in industrial settings and use data-driven insights to achieve organizational goals.

Learning outcome

KnowledgeAfter completing the course, students will:• Understand practical applications of machine learning (ML) in industrial contexts, including typical challenges and limitations.• Gain foundational knowledge of both supervised and unsupervised learning techniques and how they are applied to real datasets.• Learn about the concept of uncertainty in ML models and methods to handle and quantify this.• Understand the importance of domain knowledge to improve the accuracy of ML models and data processing, especially in relation to identifying and handling noise, errors, and unreliable data.• Be able to explain the role of machine learning in supporting production-related issues.

SkillsUpon course completion, students will be able to:• Apply machine learning techniques to real datasets, developing both supervised and unsupervised models to solve relevant industrial problems.• Use domain knowledge to preprocess and clean data, improving model reliability by identifying and correcting errors, noise, and inconsistencies.• Formulate clear, targeted objectives for ML projects based on stakeholder and user needs and expectations.• Assess the performance of ML models by analyzing results, interpreting analyses, and making data-driven parameter adjustments to optimize prediction accuracy.• Communicate complex ML concepts, processes, and findings effectively to both technical and non-technical stakeholders.

General CompetenceAfter successful completion of the course, students will:• Be able to place ML projects in a strategic context, considering both technical and business goals to create value in industrial applications.• Recognize the ethical implications of data handling and decision-making in machine learning, ensuring responsible and transparent model implementation.• Demonstrate a critical understanding of the procedural and iterative nature of ML work, with the ability to adapt and improve models based on changes in data or objectives.• Exhibit independent learning and problem-solving skills, and be able to tackle new and unfamiliar datasets and ML challenges in professional settings.• Collaborate in interdisciplinary teams, combining domain knowledge with machine learning methods to achieve common goals and maximize the impact of ML solutions.

Learning methods and activities

Learning Methods and Activities•

• Project-based teaching
• Student-active learning methods - self-study
• Group work
• Lectures or guest lectures
• Student presentations
• Field-specific laboratory work
• Company visits
• Report writing

Compulsory assignments

• Assignments

Specific conditions

Recommended previous knowledge

An introductory course on industrial AI is recommended

Required previous knowledge

Some knowledge of industrial production systems is required.

Course materials

Jay Lee: Industrial AI: Applications with Sustainable Performance, Springer, 2020

Escobar CA, Morales-Menendez R. Machine Learning in Manufacturing: Quality 4.0 and the Zero Defects Vision. Elsevier; 2024 Mar 17.

Papers and other resources will be available through the NTNU learning management system