Lecturer: Prof. Dr. Christoph Ann (TUM)

The course will cover the theoretical aspects of the module in a discussion with the lecturer. It will also provide the opportunity to work individually or in groups on case scenarios, covering issues of technology protection and patent law. The purpose is to repeat and to intensify the content discussed in course meetings and to review and evaluate legal issues from different
areas of law in everyday situations.

The course is designed to provide doctoral candidates with an understanding of how technology protection works and the stakeholders involved. Topics covered are:

• Technology protection?
• Why and how best?
• Patent quality & patent trolls
• Standard essential patents (SEPs) & FRAND licensing
• Patent courts & patent offices
• From Patents to profits

Syllabus: Technology protection for Doctoral Candidates

Lecturer: Prof. Dr. Robert Graf (International University of Applied Sciences)

Knowledge Objectives
Doctoral candidates will understand key concepts related to conditionality of probability. They will
comprehend key theorems and be able to follow core ideas of proofs. They will be able to connect different results.
Skills Objectives
Doctoral candidates learn how to derive results formally. They will also learn to transfer abstract and fundamental theorems to their research questions. They will further be able to build on fundamental theorems in their research.
Learning Objectives
Doctoral candidates will learn how to read mathematical texts and proofs. They will learn to state technical results in English, that is, in their own words. In doing so, they will learn how to rephrase technical expressions without losing rigor.

Syllabus: Conditionality in probability

Lecturer: Prof. Dr. G.P. Kiesmüller (TUM) 

Many real life systems are subject to uncertainty and should therefore be modelled with stochastic models. In this course, we focus on the theory and the application of Markov Decision Processes and Semi Markov Decision Processes. The students should gain knowledge about these models such that they are able to construct these models and apply them to solve real life problems. For illustration, we use among others, models of inventory systems, manufacturing systems and maintenance systems. We practice to derive the Bellmann equation for these systems and show how an optimal solution can be computed numerically. Besides the traditional solution approaches, we also discuss approaches based on reinforcement learning.

Syllabus: Markov Decision Processes

Lecturer: Prof. Dr. Johannes Resin (TUM)

We will discuss

• concepts such as forecast calibration (reliability) and sharpness (discrimination);
• various types of forecasts:
  • probabilistic (density) forecasts for uncertainty quantification,
  • point predictions and how to make sense of them despite forecast uncertainties,
  • other types of forecasts such as prediction intervals conveying uncertainty;
• tools to assess calibration such as PIT histograms and reliability diagrams;
• summary measures of forecast performance provided by
  • proper scoring rules,
  • consistent scoring or loss functions,
  • score decompositions;
• regression, as well as model diagnostics, from a predictive perspective;
• combination of predictive distributions.

The course enables doctoral candidates to apply these tools in their own research to identify good forecasts, as well as potential short-comings in poor forecasts. The practical relevance of the course materials will be illustrated throughout using multiple case studies.

Syllabus: Probabilistic Forecasting: What Makes a Good Forecast?