Seminar HWS 2019/20

Course Information:

  • Course ID: CS 704
  • Credit Points:
    • B.Sc. Wirtschaftsinformatik: 5 ECTS
    • M.Sc: Wirtschaftsinformatik: 4 ECTS
    • MMDS: 4 ECTS
    • Supervision: Dr. Christian Bartelt, Prof. Dr. Heiner Stuckenschmidt


  • Until Thursday, September 05, 2019 (23:59 CEST): Please register for the kick-off meeting by sending two preferred topics and a list of your completed courses (Transcript of Records, CV optional) via mail to Nils Wilken (
  • Friday, September 06, 2019: As we can only offer a limited amount of places, you will be informed whether you can participate in this seminar.
  • Monday, September 09, 2019: Latest possible drop-out date without a penalty (A drop-out after this date will be graded with 5.0)
  • Monday, September 09, 2019: Milestone 1 – Kick-Off Meeting
  • Sunday, November 03, 2019 (23:59 CET): Milestone 2 – Submission of final seminar paper
  • Sunday, November 10, 2019 (23:59 CET): Milestone 3 – Submission of reviews
  • Monday, November 18, 2019: Milestone 4 – The presentation session will take place in room 422/423 at the InES on Monday, November 18, 2019 from 10:15 to 12:15

Important notes

  • Missing a mile-stone will result in a final grade of 5.0.
  • The four parts final paper version, camera-ready paper version, feedback (reviews + presentation feedback), and presentation will be graded separately, where each part counts 25% of the final grade. Receiving a grade of 5.0 in one of the parts will result in a final grade of 5.0.
  • This seminar is open for Bachelor and Master Students focusing on “Business Informatics” and “Data Science”. Master students enrolled in the “Mannheim Master in Data Science” are also highly welcome to apply for this seminar.
  • Only Master Students enrolled in the program “Business Informatics”: This seminar will be held as Module “CS 704” and is thus only applicable for the Specialization Tracks “Information Technology”, “System Design and Development” and “Data and Web Science”.

Suggested Topics:

TOPIC 1: Current Trends in Deep Reinforcement Learning (DRL)

Introduction: Merging methods from Reinforcement Learning and Deep Learning has proven in recent years to result in a powerful approach for solving previously unsolvable problems, for example in computer vision, robotics, or multi-agent coordination.

As with many advancements in AI, there is a vast variety of research articles exploring the applications of DRL, as well as potential refinements and connections with other fields. Therefore, selecting the right approach for a given problem requires not only an overview of the ongoing work, but also an understanding of the connections and discrepancies between the different trends and methods.

Goal and Objective: In this seminar, you will gather an overview of the field of Deep Reinforcement Learning, present current approaches together with their respective advantages, drawbacks and main application areas, and particularly explore the potential usage of DRL for cooperative Multi-Agent Systems.

Starting Papers:

  • V. François-Lavet et al. (2018). An Introduction to Deep Reinforcement Learning, Foundations and Trends in Machine Learning: Vol. 11, No. 3-4. DOI: 10.1561/2200000071.
  • A. Tampuu et al. (2017). Multiagent cooperation and competition with deep reinforcement learning. PLOS ONE 12(4): e0172395.

TOPIC 2: Application of NLP technologies in Requirements Engineering

Introduction: Requirements Engineering consists of multiple activities and techniques in order to define, document and maintain software artefacts. Advances in Natural Language Processing (NLP) in the last years have also impacted Requirements Engineering. The goal of this seminar work is to provide a clear overview of approaches that apply various NLP methods to support Requirements Engineering (especially Requirements elicitation) with a focus on current approaches.

Goal and Objective: Overview of different state-of-the-art approaches that apply NLP technologies in the field of Requirements Engineering and comparison between them

Starting Papers:

  • Automated Identification of Component State Transition Model Elements from Requirements, Kaushik Madala et al., 2017

TOPIC 3: Microservice architectures - Integration/Composition and Security - Challenges and Solutions

Introduction: In the recent years microservice architectures have gained popularity to build more manageable applications compared to monolithic systems in certain situations. But microservices have their own problems and limitations in terms of complexity and security.

The greater number of independent components in the system makes it harder to compose or integrate the functionality of these components or services to supply an API for the different needs of various specific clients or use cases. Furthermore the greater number of components makes it more difficult to achieve fine-grained authorization.

The goal of this seminar is to clearly identify the problems (and the respective solutions discussed in the literature) associated with microservice architectures in terms of integration/composition and security.

Goal and Objective: Gain an understanding of the core concepts and patterns of microservice architectures in comparison to monolithic architectures. Examine the challenges of microservice architectures especially in terms of composition/integration and security and provide an overview of the available solutions or patterns for these challenges alongside with their advantages and disadvantages.

Starting Papers:

  • N. Dragoni u. a., „Microservices: Yesterday, Today, and Tomorrow“, in Present and Ulterior Software Engineering, M. Mazzara und B. Meyer, Hrsg. Cham: Springer International Publishing, 2017, S. 195–216.
  • D. Yu, Y. Jin, Y. Zhang, and X. Zheng, “A survey on security issues in services communication of Microservices-enabled fog applications,” Concurrency and Computation: Practice and Experience, vol. 0, no. 0, p. e4436, 2018.
  • D. Taibi, V. Lenarduzzi, and C. Pahl, “Architectural Patterns for Microservices: A Systematic Mapping Study,” in CLOSER, 2018.

TOPIC 4: Policy Enforcement Approaches

Introduction: In terms of software development requirements are common, regarding if the behavior of ready to use and running code the conforms to user demands is not as widespread as, but a very important aspect to prevent misuse of data. There are different known approaches of policy enforcement, including static and dynamic approaches and also approaches putting the software component or the runtime environment in charge of enforcement.

Goal and Objective: The objective of this seminar is to give an overview of different approaches to express and enforce policies. This overview should arrange and compare approaches, work out advantages and drawbacks found in the literature.

Starting Papers:

  • Khaitzin, E., Stephen, J. J., Anderson, M., Jamjoom, H., Kat, R., Natarajan, A., ... & Solomon, T. (2019). Deep enforcement: policy-based data transformations for data in the cloud. In 11th {USENIX} Workshop on Hot Topics in Cloud Computing (HotCloud 19).
  • Qazi, Z. A., Tu, C. C., Chiang, L., Miao, R., Sekar, V., & Yu, M. (2013, August). SIMPLE-fying middlebox policy enforcement using SDN. In ACM SIGCOMM computer communication review (Vol. 43, No. 4, pp. 27-38). ACM.
  • Leng, X., Hou, K., Chen, Y., Bu, K., Song, L., & Li, Y. (2019). A Lightweight Policy Enforcement System for Resource Protection and Management in the SDN-based Cloud. Computer Networks.

TOPIC 5: Machine learning techniques for preference learning in pervasive and autonomous systems

Introduction: One of the major goals of pervasive systems is to adapt to user preferences with minimal intervention of the user. Over the past decades several approaches have utilized different machine learning approaches for preference learning. The kind of machine learning approaches used, depend on the focus of the preference learning approach that utilize them. Some approaches focus on the observation of user behavior, whereas other approaches learn preferences on the basis of observed context states.

Goal and Objective: In this seminar you will research and compare several approaches for preference learning in the context of pervasive and autonomous systems. Another important aspect of the seminar is to name the advantages and disadvantages of the approaches as well as highlighting which machine learning techniques are commonly used.

Starting Papers:

  • Gallacher, S., Papadopoulou, E., Taylor, N. K., & Williams, M. H. (2013). Learning user preferences for adaptive pervasive environments: An incremental and temporal approach. ACM Transactions on Autonomous and Adaptive Systems (TAAS)8(1), 5.
  • Lim, J., Son, H., Lee, D., & Lee, D. (2017, June). An MARL-based distributed learning scheme for capturing user preferences in a smart environment. In 2017 IEEE International Conference on Services Computing (SCC) (pp. 132-139). IEEE.

Nach oben

PostanschriftBesucheradresse  EmailTelefonnummer
Institute for Enterprise Systems  L15,  Office Management
Schloss68131 Mannheim +49 621 181-3560
68131 MannheimGermany