Who Builds Reinforcement Learning? A Patent Analysis

Forschungsthema/Bereich: Reinforcement Learning · Patent Analysis
Typ der Abschlussarbeit: Master
Startzeitpunkt: -
Bewerbungsschluss: 31.08.2026
Dauer der Arbeit: 6 Monate, ab sofort

Beschreibung

Who Builds Reinforcement Learning? A Patent Analysis

Reinforcement learning is one of the most powerful ideas in machine learning. An agent acts in an environment, takes sequential decisions, and learns by trial and error to maximise a cumulative reward over multiple periods. It is a prominent topic in research. Where it is actually used outside academia is not that easy to answer.

Part of the reason could be that RL is almost never a product in itself. It sits somewhere inside one, as a controller, a scheduler, a pricing policy, and companies have little reason to advertise it. Patents are one of the few places where the trace becomes visible. Filing one means a firm decided the idea was worth real money, often years before anything ships. Whether the firm keeps paying the renewal fees afterwards tells you if companies still believe in the idea.

In this thesis you build a patent corpus on reinforcement learning and turn it into an empirical picture of industrial RL: which industries, which kinds of decisions, internal operations or customer-facing systems, which RL methods, and which of it a company keeps paying for.

Specifically:

Building a de-duplicated corpus from patent databases
Operationalising a strict definition of sequential RL, and measuring how much of the official "reinforcement learning" classification actually meets it
Statistical analysis of the corpus metadata: how RL patenting develops over time and across methods, which companies and countries drive it, and whether patents are maintained or allowed to lapse
Manual deep coding of a stratified sample: e.g. application domain, internal or customer-facing use, learning setup, whether empirical results are reported at all

What you bring:

Python, since you will be pulling and cleaning the data yourself
patience with messy text data, and an interest in empirical work on a dataset nobody has cleaned for you
An advantage, but not required: familiarity with RL concepts, or a background in innovation and technology management

What you get:

close supervision on research design and analysis
a topic with a clear empirical contribution plus a dataset that will be reused beyond your thesis

Disclaimer
The thesis is meant to teach you the scientific approach. Please keep that in mind. We expect you to learn and use scientific methods and to submit a scientifically sound result. Our role is to provide you with everything you need to succeed at that. We expect that you create a piece of work, in which you demonstrate a sound data collection method, an extensive dataset and your analysis of that dataset. A motivated student who works consistently on the thesis throughout the designated period can expect a very good grade. Details on the process and the format that we use to support you in succeeding with your thesis can be found here:

Master: https://bwsyncandshare.kit.edu/s/MiN8yTkfgeCrnSW

If you are interested, please send a cover letter and transcript of records by email to julius.grams@kit.edu.

References

Dulac-Arnold, G., Levine, N., Mankowitz, D. J., Li, J., Paduraru, C., Gowal, S., & Hester, T. (2021). Challenges of real-world reinforcement learning: definitions, benchmarks and analyses. Machine Learning, 110, 2419-2468.
Lin, Y. & Maruping, L. M. (2025).Organizing for AI Innovation: Insights From an Empirical Exploration of U.S. Patents1. MIS Quarterly 1 September 2025; 49 (3): 1095–1122. https://doi.org/10.25300/MISQ/2025/18765
Nickerson, R. C., Varshney, U., & Muntermann, J. (2013). A method for taxonomy development and its application in information systems. European Journal of Information Systems, 22(3), 336-359.

Voraussetzung

Voraussetzungen an Studierende

Python, since you will be pulling and cleaning the data yourself
Patience with messy text data, and an interest in empirical work on a dataset nobody has cleaned for you
An advantage, but not required: familiarity with RL concepts, or a background in innovation and technology management

Studiengangsbereiche

Wirtschafts- und Rechtswissenschaften
Wirtschaftsinformatik

Betreuung

Titel, Vorname, Name: Julius Vincent Grams
Organisationseinheit: KIT WIN - ESIS
E-Mail Adresse: julius.grams@kit.edu
Link zur eigenen Homepage/Personenseite: Website

Bewerbung per E-Mail

Bewerbungsunterlagen

Anschreiben
Notenauszug

E-Mail Adresse für die Bewerbung
Senden Sie die oben genannten Bewerbungsunterlagen bitte per Mail an julius.grams@kit.edu

Zurück