Back to TUM-PREP Projects overview

Project Overview

Project Code: MGT 02

Project name:

Learning to Collaborate: A Meta-Learning Approach to Forecast Integration

TUM Department:

MGT – School of Management

TUM Chair / Institute:

Chair of Logistics and Supply Chain Management

Research area:

Human-AI Interaction; Supply Chain Management

Student background:

Computer ScienceManagement / Economics

Further disciplines:

Participation also possible online only:

Planned project location:

Chair of Logistics and Supply Chain Management
TUM School of Management
Arcisstraße 21
80333 Munich, Germany

Project Supervisor - Contact Details


Title:

Prof. Dr.

Given name:

Stefan

Family name:

Minner

E-mail:

stefan.minner@tum.de

Phone:

+49 (0)89 289 28201

Additional Project Supervisor - Contact Details


Title:

M.Sc.

Given name:

Yuxuan

Family name:

Zhu

E-mail:

yuxuan_thu.zhu@tum.de

Phone:

Additional Project Supervisor - Contact Details


Title:

Given name:

Family name:

E-mail:

Phone:

Project Description


Project description:

Background
The Chair of Logistics and Supply Chain Management launched a research project with a company in the tire manufacturing industry.
The TUM PREP project will focus on the human-AI collaborative forecasting scenario. The objective is to use Big Data and Machine Learning to improve overall forecasting accuracy and promote efficiency in operation managements.

Project
Collaborative forecasting between humans and AI systems increasingly relies on dynamic adaptation to varying tasks, environments, and user behaviors. This project explores the use of meta-learning techniques to enhance human–AI integration in demand forecasting.
Meta-learning, or "learning to learn," aims to develop models that can rapidly adapt to new tasks with limited data by leveraging experience from previous tasks. In the context of human–AI forecasting, meta-learning can enable the system to tailor its collaboration strategies—such as weighting human versus AI inputs—based on contextual signals like data availability, product characteristics, or forecaster expertise. This approach allows for more flexible and personalized integration, improving both prediction accuracy and user acceptance. By continuously updating its own learning strategy, a meta-learning framework supports long-term human–AI co-evolution and facilitates scalable, context-sensitive forecasting solutions.

Requirements
Proficiency in Python programming and basic knowledge of machine learning is required. A keen interest in data analytics is crucial. The interest in working in an industry-related project is recommended.

This project offers an exciting opportunity to explore the intersection of demand forecasting, Big Data, and AI, with a focus on improving inventory decisions.

Working hours per week planned:

40

Prerequisites


Required study level minimum (at time of TUM PREP project start):

2 years of bachelor studies completed

Subject related:

Other:

  • Keine Stichwörter