Login
Register
Help
Documentation
Studentische Hilfskräfte, Praktikantenstellen, Studienarbeiten
Hier finden Sie Stellen für Studentische Hilfskräfte, Praktikantenstellen an der TU München sowie Studienarbeiten
Eine Anleitung zum Erstellen finden Sie in der Kurzanleitung für Stellenanzeigen und (ausführlicher) im Best Practice Manual Stellenanzeigen (pdf)
23.08.2022
Studentische/r Mitarbeiter/in (m/w/d) für die Professur Recht, Wissenschaft und Technologie
Wir, die Professur Recht, Wissenschaft und Technologie (Professor Dr. Christian Djeffal), suchen zum nächstmöglichen Zeitpunkt eine/n Studentische/n Mitarbeiter/in (m/w/d) 8-16 Std./Woche.
read more
Kontakt: professur.djeffal@tum.de
22.08.2022
Studentische Hilfskraft (w/m/d) an der Professur für Entrepreneurial Finance
Wir, der Chair of Entrepreneurial Finance (Prof. Momtaz) der TUM School of Management, suchen ab sofort eine/n
Studentische/n Mitarbeiter/in (w/m/d)
im Bereich Entrepreneurial Finance, insb. Crypto und Blockchain
(6 bis 9 Wochenstunden)
read more
Kontakt: office.entfin@mgt.tum.de
19.08.2022
Working student - Full stack web development
We are looking for a student to expand our development team for an online self-service platform to upload, review, and find research data.
read more
Kontakt: michael.eichelbeck@tum.de
19.08.2022
Studentische Hilfskraft (m/w/d) für Munich Dual Career Office gesucht
Das Munich Dual Career Office (MDCO) ist Teil des Präsidialstabs Berufungen, Karriereaufstieg und Dual Career und richtet sich an Spitzenwissenschaftler*innen und ihre Dual Career Partner*innen. Den Schwerpunkt des MDCO bildet die Beratung zur Arbeitsmarktintegration und Jobsuche der Dual Career Partner*innen in München. Wir bieten individuelle Karriereplanung und -beratung, entwickeln neue Karriereperspektiven und vermitteln Kontakte zu unserem MDCO-Netzwerk sowie berufliche Anschlussmöglichkeiten am neuen Wohnort. Dazu unterstützen wir die gesamte Familie beim Umzug und Ankommen an allen Standorten der TUM.
read more
Kontakt: dualcareer@zv.tum.de
18.08.2022
MA: How to Achieve Optimality in Safe Reinforcement Learning?
We are looking for a highly motivated student for a master thesis project in the area of safe Reinforcement Learning (RL)! Within this project, you will modify various RL algorithms to account for a safety-preserving correction of the action proposed by the algorithm. You will evaluate the effect of these modifications using a simulation of an inverted pendulum.
read more
Kontakt: hannah.markgraf@tum.de
18.08.2022
MA: How to Achieve Optimality in Safe Reinforcement Learning?
Many real-world applications such as autonomous driving, power system operation, or robot navigation require powerful decision-making tools. Reinforcement Learning (RL) has proven its potential to control complex systems from various applications by learning through interaction with the environment. However, most state-of-the-art RL algorithms have a significant disadvantage which prevents their deployment beyond simulated environments: They cannot guarantee fulfilling safety specifications, e.g., obstacle avoidance in autonomous driving. A simple yet effective approach for safeguarding RL algorithms works as follows: If the action proposed by the RL algorithm steers the agent into an unsafe state, a safety controller is employed to modify that action so that the agent does not leave the safe set. The safety-preserving action can be obtained by projecting the (potentially unsafe) action onto a set of admissible actions. However, adjusting the action proposed by the RL algorithm may disrupt the learning process and result in suboptimal policies. Gros et al. therefore derived corrections that ensure optimality in both Q-Learning and policy gradient methods despite safe action projections as described above. However, they did not provide practical examples showcasing the effects of their theoretical results.
The goal of this thesis is to evaluate the influence of applying the corrections proposed by Gros et.al. to RL control of a dynamic system, e.g., an inverted pendulum. This includes extending an existing framework for safe RL of an inverted pendulum with the corrections of the learning process. Three different RL algorithms will be examined: Q-learning, deterministic policy gradient
optimization and stochastic policy gradient optimization. For Q-learning, an additional goal is to analyze two different possibilities for integrating exploration into the learning process. The effect of the corrections should be evaluated empirically first. Optionally, it is then possible to derive theoretic bounds on the error introduced without the correction. Another extension would be to test the framework for more complex dynamic systems.
read more
Kontakt: hannah.markgraf@tum.de
18.08.2022
Studentische Hilfskraft (m/w/d) für Virtual Reality Probandenerhebung gesucht
Der Lehrstuhl für Ergonomie sucht für die Durchführung einer Probandenstudie in virtueller Realität zum Thema Mensch-Roboter Interaktion eine studentische Hilfskraft (m/w/d).
Voraussetzung sind sehr gute Deutschkenntnisse , Gewissenhaftigkeit und ein sorgsamer und professioneller Umgang mit Equipment und Proband:innen.
read more
Kontakt: olivia.herzog@tum.de
17.08.2022
Student assistants "Digital Logistics"
The chair of Logistics and Supply Chain Management looks for student assistants in the project "Digital Logistics".
read more
Kontakt: logistics.log@mgt.tum.de
12.08.2022
Student research assistant position in DFG-funded project
The Chair of Science and Technology Policy at the Department of Science, Technology and Society (STS) is searching for a student research assistant to support the DFG-funded project led by Dr. Susanne Koch. “In-Forest: a multi-method empirical study of inequality and its epistemic effects in forest research” seeks to enhance the understanding of how dimensions of inequality in academia intersect and impact on knowledge production, using forest science as empirical case. It is funded by the German Research Foundation (DFG) and involves international collaboration with partners at the Centre for Research on Evaluation, Science and Technology (CREST, Stellenbosch University, South Africa), the Centre for Science and Technology Studies (CWTS, Leiden University, The Netherlands) and the Tanzania Forestry Research Institute (TAFORI).
read more
Kontakt: camilla.tetley@tum.de
11.08.2022
Practical course / HiWi: Development of a PyTorch-based framework for deep learning research on autonomous driving
Talented students wanted for practical course project with downstream HiWi positions.
read more
Kontakt: eivind.meyer@tum.de
Print
RSS 
