Paul Maier
Researcher/PhD Student
Department of Informatics of TU München
Boltzmannstraße 3, 85748 Garching
Office: 02.013.035
Phone: +49 (89) 289-19595
Fax: +49 (89) 289-17757
maierpa@in.tum.de
Relocated to new office: 02.013.035
For students: Check out my currently open projects (BA/M/DA/SEP) here. Other topics within our group can be found here.
Intro
I started my PhD in September 2007 here at Informatics/TUM in the group of Dr. Martin Sachenbacher. My research interests are
- model-based development and operation of technical systems
- constraint-based diagnosis and planning
- model-based reasoning, specifically model-based diagnosis
- constraint optimization
- probabilistic reasoning
On a more general scope I'm interested in the elements that constitute intelligent behavior and problem solving. I find it deeply fascinating to view and analyze tasks such as bringing a product design to live as a real product as computational problems.
Prior to my PhD studies I made my diploma in Computer Science at the University of Passau. In my diploma thesis, "Content-Based Image Retrieval: Review and Benchmarking", I analyzed and benchmarked a number of content-based image retrieval systems.
Research
In my current research I focus on the model-based operation of manufacturing plants. Future factories are envisioned as highly automated facilities that can help human workers not only with mundane, repetitive tasks, but also support them by, e.g., automatically generating manufacturing schedules for customized products. Typically, automated scheduling is computationally intensive and therefore is done "off-line", e.g. during the night. The next day schedules can be executed to manufacture the products. However, while executing, current sensor data might indicate faults in this complex environment. The question arises, how this affects the single products currently being processed. In my thesis, I address this question as the problem of plan assessment: computing the probability that a product's manufacturing plan will successfully create the product, given sparse on-line sensor information and a probabilistic model of factory stations and products.
Selected Publications
-
P. Maier and D. Jain and M. Sachenbacher:
Compiling AI Engineering Models for Probabilistic Inference,
Proc. KI-2011,7006,191--203,Springer,(2011)
(pdf)
- P. Maier and D. Jain and S. Waldherr and M. Sachenbacher: Plan Assessment for Autonomous Manufacturing as Bayesian Inference, Proc. KI-2010,6359,263--271,Springer,(2010)
(pdf)
- P. Maier and M. Sachenbacher and T. R\"uhr and L. Kuhn: Automated Plan Assessment in Cognitive Manufacturing, Adv. Eng. Informat.,24,241--376,(2010)
(pdf)
- P. Maier and M. Sachenbacher: Diagnosis and Fault-Adaptive Control for Mechatronic Systems using Hybrid Constraint Automata, Proc. {PHM}-2009,(2009)
(pdf)
- P. Maier and M. Sachenbacher and T. Rühr and L. Kuhn: Integrating Model-based Diagnosis and Prognosis in Autonomous Production, Proc. {PHM}-2009,(2009)
(pdf)
- P. Maier and M. Sachenbacher: Adaptive Domain Abstraction in a Soft-Constraint Message-Passing Algorithm, Workshop {P}roc. {SOFT}-2008,(2008)
(pdf)
- P. Maier: Adaptive Abstraction of Constraint-Based Models for Self-Diagnosis and Planning, Proc. {AAAI}/sigart doctoral consortium,1859--1860,The AAAI Press,(2008)
(pdf)
- P. Maier and D. Jain and S. Waldherr and M. Sachenbacher: Plan Assessment for Autonomous Manufacturing as Bayesian Inference, Proc. KI-2010,6359,263--271,Springer,(2010)
A list of all publications can be found here.
Teaching
Current:
Past:
- SS10: Seminar - Intelligente Produktionsanlagen mit modellbasierter Selbstwahrnehmung
- WS09/10: Seminar - Intelligente Produktionsanlagen mit modellbasierter Selbstwahrnehmung
- WS08/09: Einführung in die Informatik 1 - Praktikum Grundlagen der Programmierung
- SS08: Grundlangen Algorithmen und Datenstrukturen - Übung
Misc
Talks at conferences and workshops:
KI'10, KI'09, ETFA'09, PHM'09, CPAIOR'08, AAAI'08, WARTS'08 (webpage)
Visits of and talks at institutions:
- October 2010: Stanford Research Institute (SRI)
- October 2009: Palo Alto Research Center (PARC)
Software I use:
- Toolbar is a open source suite of constraint optimisation algorithms (based on weighted CSP). I integerated my algorithm BEDA into this suite. It can be downloaded at: toolbar
- Toulbar2, the advancement of toolbar, is a very efficient state-of-the-art constraint optimization solver
- Bayesian Logic Networks, a framework for compact modeling of statistical relational problems
- Tools of the trade: eclipse (pydev, texlipse), lyx (for the thesis), bibdesk, NoteTaker HD, iAnnotate
