Content: Twelfth Workshop

Twelfth Workshop


The twelfth Hybris Workshop will take place on November 12th-13th 2018 at the technical faculty of the University of Freiburg, Georges-Köhler-Allee in building 101.

Venue

Click here for directions.

Program

Monday, November 12th

12:45-13:00Registration and Welcome
13:00-14:00Invited talk by Hector Geffner, ICREA and Universitat Pompeu Fabra.
Representation and Learning for Generalized Planning
14:00-14:30David Speck, Uni Freiburg
Symbolic Planning with Axioms: Representation and Reasoning
14:30-15:00Bernhard Nebel, Uni Freiburg
Implicitly Coordinated Multi-Agent Path Finding under Destination Uncertainty
15:00-15:30Coffee Break
15:30-17:30Invited talk by Alessandro Artale, University of Bolzano.
Tutorial: Ontology-Mediated Query Answering over Temporal Ontologies
17:45-19:00Hybris PI Meeing
20:00Dinner at Paradies,
Paradies, Mathildenstraße 28, 79106 Freiburg im Breisgau

Tuesday, November 13th

9:00-10:00Invited talk by Craig Boutilier, Google.
Towards User-centric Recommender Systems
10:00-10:30Coffee Break
10:30-11:00Tim Welschehold, Uni Freiburg
Combined Task and Action Learning from Human Demonstrations
11:00-11:30Torsten Schaub, Uni Potsdam
Introducing Temporal Stable Models for Linear Dynamic Logic
11:30-12:00Jens Claßen, RWTH Aachen
Reasoning about Conditional Beliefs for the Winograd Schema Challenge
12:00-13:00Lunch Break
13:00-13:30Andreas Ecke, TU Dresden
Consistency in Probabilistic Description Logic using Types
13:30-14:00Benjamin Zarrieß, TU Dresden
On Relatively Complete Verification of Golog Programs
14:00-14:30Marco Wilhelm, TU Dortmund
ALC^ME - A Probabilistic Description Logic Under the Aggregating Semantics and the Principle of Maximum Entropy
14:30-15:00Gerd Brewka, Uni Leipzig
Strong Inconsistency and Strong Explanations in Nonmonotonic Logics
15:00Coffee and Farewell

Abstracts

Representation and Learning for Generalized Planning

Invited talk by Hector Geffner, ICREA and Universitat Pompeu Fabra

Abstract: Generalized planning is concerned with the computation of plans that solve multiple planning instances at once. For example, the plan that iteratively picks up the clear block that is above x achieves the goal clear(x) in any instance of the blocks world. Versions of this problem have attracted attention in both planning and learning. In the talk, I'll go over three ideas that appear to provide a joint computational solution to the problem. The first is a reduction of the generalized planning problem to a qualitative numerical problem and then to a fully observable but non-deterministic (boolean) planning problem. The second is an abstraction that enables the use of such reductions even when the different planning instances involve different actions. The third is a way for learning this abstract representation automatically. Examples and empirical results will also presented for illustrating these ideas.

Ontology-Mediated Query Answering over Temporal Ontologies

Invited talk by Alessandro Artale, University of Bolzano

Abstract: In this talk we introduce the notion of ontology-based data access and extend it to the case where both the data and the Ontology have a temporal dimension. We present a language of ontology-mediated queries by extending OWL 2 QL and SPARQL with temporal operators, and investigate rewritability of these queries into two-sorted first-order logic with < and PLUS over the temporal dimension.

Toward User-centric Recommender Systems

Invited talk by Craig Boutilier, Google

Abstract: Artificial intelligence and machine learning technologies continue to broaden and influence our access to information, entertainment, products and services---and each other---through data-driven recommendations. While the increased access afforded by AI has undoubtedly improved certain aspects of social welfare, the ability of recommenders to generate genuinely personalized recommendations and engage users in meaningful ways remains limited. Furthermore, our understanding of how AI recommenders shape long-term user behavior is poorly understood. In this talk, I will discuss the role that various AI techniques have to play in next-generation, user-centric recommender systems. Among these are preference modeling and preference elicitation; reinforcement learning and latent state models; behavioral decision theory and economics; and modeling of user behavioral preferences. I will also highlight challenges that emerge when putting these methods into practice. I’ll conclude with some discussion of recent work applying reinforcement learning in a practical recommender system.