Navigation

Skiplinks

Contact

  • Director
    Prof. Dr. Claus Möbus

    Room: Uni, A02 2-226
    Tel:+49-441-798-2900
                                                                                                     claus.moebus(at)uol.de
    claus.moebus(at)uni-oldenburg.de 
  • Secretary
    Manuela Wüstefeld

    Room: OFFIS, O42
    Tel:+49-441-9722-117

    manuela.wuestefeld(at)uni-oldenburg.de 

Topics for Bachelor and Master Theses

Bachelor or master theses in probabilistic modeling, machine learning, or applied artificial intelligence are directed by me and other researchers.

The production of the thesis starts and ends with a presentation (Vortrag). In the start presentation you present the topic and the milestone blueprint. This takes place in my research seminar "Probabilistic Modelling" (inf 533 and/or inf533). You should enroll in this seminar via Stud.IP and also qualify for success ("Bescheinigung der erfolgreichen Teilnahme"). This can be achieved by the (successful) presentation and the written milestone description.

In the end presentation you summarize and demonstrate the results of the thesis. This takes place in the relevant seminar ("Oberseminar", etc) of the co-examiner of the thesis. Depending on the special topic of the thesis (e.g. probabilistic robotics, computational intelligence, machine learning, business intelligence) the following colleagues (Prof. Fränzle, Prof. Kramer, Prof. Sauer) are recommended as coexaminers.

Interested students should consult this schedule ("Leitfaden") and are invited to contact me via email:

Prof. Dr. Claus Möbus claus.moebus(at)uol.de

________________________________________________________________________________________________________________________________

Deep Probabilistic Programming with Edward and WebPPL

Edward and WebPPL are a new Turing-complete probabilistic programming languages PPLs). Edward (named after the famous statistician Edward Box) is a TensorFlow library which is used for deep learning (DL) artificial neural networks (ANNs). WebPPL is a domain specific language (DSL) embedded in the functional part of JavaScript (JS). PPLs are used for building generative probabilistic models (GPMs). These models represent causal background knowledge which is characteristic for experts. In contrast, deep learning models only represent shallow knowledge which is useful for pattern matching. In that respect the latter have become rather successful. This paper describes with many examples the fundamental difference between DL and GPMs (Lake, et al., 2016).

The thesis should survey models programmed in Edward and WebPPL. This set is called the paradigmatic example set. This set should be partitioned in the joint set Edward /\ WebPPL and the two difference sets Edward\WebPPL and WebPPL\Edward. The last two sets are of special interest. Are these sets existent by accident or are there fundamental difficulties in compiling examples from one PPL to the other ?

Interest students are invited to contact me via email:

Prof. Dr. Claus Möbus claus.moebus(at)uol.de

Thema: Portfoliooptimierung durch Diversifikation von Risiken

Spätestens bei Kurseinbrüchen würde sich mancher Wertpapierbesitzer wünschen, er hätte etwas für die Risikostreuung und -minimierung getan. Nun kann nach Martin Weber (Prof. an der Uni Mannheim) ein Laie zwar nicht besser performen als der Markt, aber er/sie kann etwas für das Risikomanagement im Portfolio tun. Das Vorgehen wurde vom Nobelpreistträger Markowitz 1952 im Aufsatz Portfolio Selection theoretisch beschrieben. Markowitz erhielt dafür 1990 den Nobelpreis. Weber gibt in seinem Buch Genial einfach investieren; Mehr müssen Sie nicht wissen - das aber unbedingt! in Kapitel 6 praktische Handlungsanweisungen. Etwas mathematischer - aber immer noch gut lesbar - ist die Behandlung des Themas Portfoliooptimierung im Kap. 4.2 Diversifikation von Risiken im überaus empfehlenswerten Buch von Cottlin & Döhler, Risikoanalyse, 2013 2/e.

Dazu müssen einige statistische Berechnungen (Standardabweichungen, Korrelationen) durchgeführt werden. Diese stellen für Laien ohne Assistenz eine unüberwindbare Hürde dar. Einschlägige Berechnungen lassen sich auch mit dem Opensourcepaket OLPS (= On-Line Portfolio Selection Toolbox) oder besser noch mit den npm libraries finance und portfolio-allocation durchführen. Dazu werden aber ebenfalls für Endbenutzer technisch und inhaltlich unüberwindbare Hürden aufgebaut.

Hier soll nun die Bachelor- oder Masterarbeit, unter Beachtung juristischer Rahmenbedingungen (Lizenzgebühren, Ausschluss von Haftung etc), Abhilfe schaffen. Der Ratsuchende soll browseruntzerstützt ein Portfolio zusammenstellen, bewerten und mit der Performance des von Prof. Weber gemanagten ARERO-Fonds vergleichen. ARERO (Acronym für Aktien, REnten und ROhstoffe) ist ein nicht aktiv gemanagter Fonds, der u.a. den MSCI-Worldindex nachführt) und als Maß der Dinge beim Vermeiden von Anlegerfehlern angesehen wird. 

Zusätzlich soll in der Masterarbeit untersucht werden, ob die dynamische Webseite als PWA (Progressive Web Application) realisiert werden kann.

Vorkenntnisse: Der/die Kandidat/In sollte WI mit Erfolg studieren bzw studiert haben, Grundkenntnisse in deskriptiver Statistik (Mittelwert, Standardabweichung, Varianz, Korrelation, Regression, etc) vorweisen können, Kapitel 6 in Weber's Buch verstanden haben und dynamische Webseiten realisieren können.

Kontakt: Prof. Dr. Claus Möbus

Kloning a Human Telecontroller with a Probabilistic or Deep Learning Model in the Realm of Sumo-Robotics

2017 an explorative MSc-thesis with a similar topic was finished. In the thesis Sumo-bots in the Lego league were controlled by agents using simple probabilistic models of the "naive" Bayesian type. These results were encouraging so that a more refined methodology should provide more perfect results. Especially the lessons learnt are important for further research.

The new research should improve following aspects:

1. Human telecontrolers shouldn't use solely a bird's eyes view but the view of the controlled bot or a combination of both.

2. Model evaluation should be more systematic.

3. The behavior of the human telecontroller and the sumo bot should be organized in a behavior hierarchy along maneuvers, tactics, and strategies.

4. Bot-actions should be inferred in real-time by alternative Bayesian or Deep Learning models.

5. The sumo-bots should fall into a standard sumo-bot competition category.

6. The sumo-bots should be controlled by standard game controller.

Prerequisites: The candidate should have basic knowledge in (educational) robotics and probabilistic and/or deep learning modeling. The first knowledge can be obtain when playing with e.g. Lego bots, the second when attending the seminars "Probabilistic Modelling I & II" (Inf533, Inf534), and the third, when reading the book "Deep Learning".

Contact: Mark Eilers (MSc), Claus Möbus (Prof. Dr.)

Probabilistic Modeling with Model Fragments, Patterns, or Templates

WebPPL is a web-based probabilistic programming language (PPL) embedded in JavaScript. PPLs are used for implementing probabilistic models in domains with uncertain knowledge (cognition, medicine, traffic, finance, etc). Dependent on the situation-specific problem questions in the form of unknown (conditional) probabilities are formalized. Models and programs generate answers by numerical inference processes.

The interactive tutorial "Probabilistic Models of Cognition" provides a variety of WebPPL-models. There is nearly always a fixed sequence modelling steps: 1) modelling the causal process of interest (root causes, expositions -> syndroms -> symptoms), 2) observation of evidence (data), 3) (diagnostic) inference (most often) contrary to the causal direction (symptoms -> syndroms -> expositions).

Challenge: The research question is, whether the modelling process can be improved substantially by a library of model fragments, patterns, or templates.

Prerequisites: The topic is suited for a master thesis. Preknowledge can be acquired by successful participation in the seminar "Probabilististic Modeling I & II" (Inf533, Inf534) and studying the above mentioned tutorial.

Contact: Prof. Dr. Claus Möbus

Webmaster (Changed: 2018-09-10)