Seminar

Seminar

Seminar "Reasoning, Planning, and Scheduling with Uncertainty"

Submitted by Till Hofmann on 13. January 2018 - 13:46

In this seminar, we will study uncertainty in the context of reasoning, planning, and scheduling. For reasoning about actions, we will look into stochastic extensions of the Situation Calculus, a well-known formalism for reasoning about dynamic domains. In the classical Situation Calculus, all actions are deterministic. In this seminar, we will learn about extensions that allow non-deterministic and probabilistic actions. For planning, we will investigate probabilistic extensions to classical planning frameworks such as the Planning Domain Definition Language (PDDL) and compare them to Markov Decision Processes (MDPs). For scheduling, we will learn about mechanisms for solving scheduling problems with probabilistic task durations.

Seminar "Selected Topics on Planning and Plan Execution for Robotic Systems" SS 2017

Submitted by Till Hofmann on 9. January 2017 - 18:41

In this seminar we will study plan generation and plan execution in robotic domains. While planning deals with the problem of finding suitable actions to accomplish a certain goal, plan execution and monitoring on robots come with additional problems: The agent's knowledge is no longer complete, the environment may change during plan execution, humans and other agents may interfere with the agent's actions, and resource and temporal constraints need to be honored. The plan executive needs to execute actions physically on a robot, adapt to changes during execution, and repair the current plan or replan if necessary. In this seminar, we will look at various methods to deal with those problems and we will investigate how these methods are applied on real-world robotic systems.

Seminar Forgetting and Relevance SS 2016

Submitted by Christoph Schwering on 30. December 2015 - 20:59

In this seminar we will study concepts of forgetting and relevance developed in the field of Knowledge Representation. Humans not only forget unintentionally but also intentionally, for example, when they obtain new information and delete previous contradicting and apparently false information. In fact, humans usually do not forget irrevocably; we can bring back memories and reconstruct forgotten knowledge. In formal languages as studied in KR, however, even supposedly simple forgetting often turns out very difficult. In this seminar, we will study techniques for forgetting in KR languages such as predicate logic, description logics, or answer set programs. Related to forgetting is the concept of relevance among different facts. For example, when some fact is to be forgotten, this may also affect other facts relevant to the forgotten fact. Relevance is useful to develop tractable reasoning and handling inconsistencies in a knowledge base.

Seminar Beliefs and Causality WS 2015/2016

Submitted by Christoph Schwering on 2. August 2015 - 6:40

In this seminar we will study papers from various subfields of Knowledge Representation. This includes belief revision; argumentation theory; planning; causality in general and the situation calculus in particular. All of these subfields of AI are related to each other. For example, one usually acts depending on what one believes, and actions in turn affect one's beliefs. In this seminar, we will study recent papers on these topics (particularly from the IJCAI-15 conference).

Seminar "Selected Topics in Multi-Agent Reasoning Systems" WS 2014/2015

Submitted by stf on 1. August 2014 - 14:22

In this seminar we will study several systems used for reasoning in multi-agent scenarios.

We plan to implement a peer review process for this seminar. That is, every student will read some other students' term paper and provide feedback in form of a written review. This shall not only deepen your understanding of the other topics, but it also introduces you to the academic review process.

Seminar Dynamics of Knowledge and Belief SS 2014

Submitted by Christoph Schwering on 6. January 2014 - 0:08

In this seminar we will study several modeling and reasoning techniques for knowledge and belief in dynamic systems. Knowledge is an important aspect of intelligent programs: while most of today's systems assume a closed world, i.e., everything they don't know to be true is assumed to be false, an intelligent system needs to consider possible that there are truths not known to the system. In a dynamic environment, i.e., an environment where one or multiple agents (inter)act, the system will usually have to acquire new knowledge through sensing. Potentially it may even revise its beliefs when it realizes some beliefs were wrong. In this seminar we will study various aspects of action, knowledge, and belief.

We plan to implement a peer review process for this seminar. That is, every student will read some other students' term paper and provide feedback in form of a written review. This shall not only deepen your understanding of the other topics, but it also introduces you to the academic review process.

Seminar Plan and Activity Recognition SS 2013

Submitted by Christoph Schwering on 11. January 2013 - 15:14

In this seminar we will study several different approaches to and aspects of plan and activity recognition. Recognizing what other agents are doing is an important aspect of intelligent systems. For example, a domestic service robot needs to understand what the human is doing in order to interact with him in a reasonable way. And a self-driving car should know what the other traffic participants are doing right now and infer what they are going to do in the next moments. Different domains bring along different problems and needs for levels of expressiveness like partial observability, incomplete knowledge, non-deterministic actions, adversarial agents, potentially hazardous situations. We will study some of the latest research on these problems and work out the particular strengths and weaknesses.

The topics include recent papers on plan and activity recognition.

Seminar Robust Reliable Robotics SS 2012

Submitted by tim on 9. January 2012 - 13:10

In this seminar we will study several aspects of robust and reliable robotics. Robots are machines created to fulfill particular tasks instead of or in cooperation with humans. In virtually all scenarios a failure is annoying or even catastrophic. Planetary rovers cannot be repaired easily or at all, broken factory robots can become vastly expensive not only due to the cost to repair the robot itself, but the problems they cause for the overall supply chain; and domestic service robots operate in close proximity to humans in their habitats and must take special precautions as not to harm a human or damage the interior. These considerations make it necessary to develop techniques and systems that enable a robot system to detect failures or unexpected behavior and at least stop, better even work around the problem.

The topics include recent papers on execution monitoring, robot system debugging, and fault detection.

Seminar Knowledge Representation and Computational Tractability SS 2011

Submitted by vaishakbelle on 4. January 2011 - 18:11
Introduction

Knowledge Representation (KR) is a vibrant and exciting field in artificial intelligence. The endeavor rests on two fundamental ideas. First, to reason about the problem domain one must formalize it, perhaps in some logical formalism such as propositional logic or first-order logic. Second, for the representation to be useful one must be able to obtain reasonable and intuitive inferences in a timely fashion.
Unfortunately, propositional reasoning is intractable (Boolean reasoning is NP-COMPLETE) and first-order logic is undecidable. Thus, an important goal in the KR enterprise is to find a tradeoff be- tween the expressiveness of the representational language and the computational behavior of associated reasoning tasks. A main objective of this seminar is to discuss approaches bordering this tradeoff.