Hybrid Learning and Reasoning

Seminar an der Universitšt des Saarlandes, Fachrichtung Informatik,  LSF 140201

The topics for seminar time slots (see schedule) are as follows. Currently assigned topics are marked in red. Selected background literature is indicated below; topic reference papers are available in the web or on request from seminar organizers.

Topic #

Topic

1

Informed Learning with Prior Knowledge (I)

Phan, N., et al. (2017): Ontology-based Deep Learning for Human Behavior Prediction with Explanations in Health Social Networks. Journal of Information Sciences, 384, Elsevier.  [paper]

2

Informed Learning with Prior Knowledge (II)

Marra, G., et al. (2020): Inference in Relational Neural Machines. Proceedings of 1st International Workshop on New Foundations for Human-Centered AI co-located with 24th European Conference on Artificial Intelligence (ECAI).  [paper]

3

Informed Learning with Prior Knowledge (III)

De Giacomo, G., et al. (2019): Foundations for Restraining Bolts: Reinforcement Learning with LTLf/LDLf Restraining Specifications. Proceedings of 30th International Conference on Automated Planning and Scheduling (ICAPS).  [paper]

4

Learning Logical Representations for Statistical Inferencing (I)

Cohen, W.W., et al. (2020):  Scalable Neural Methods for Reasoning with a Symbolic Knowledge Base. Proceedings of 8th International Conference of Learning Representations (ICLR). [paper]

5

Learning Logical Representations for Statistical Inferencing (II)

Almasan, P., et al. (2019): Deep Reinforcement Learning Meets Graph Neural Networks: Exploring a Routing Optimization Use Case. Journal of Computer Communications.  [paper]

6

Learning for Planning (I)

Ferber, P. et al. (2022): Neural Network Heuristic Functions for Classical Planning: Bootstrapping and Comparison to Other Methods. Proceedings of 32nd International Conference on Automated Planning and Scheduling (ICAPS).   [paper] [sup]

7

Learning for Planning (II)

Gehring, C. et al. (2022): Reinforcement Learning for Classical Planning: Viewing Heuristics as Dense Reward Generators. Proceedings of 32nd International Conference on Automated Planning and Scheduling (ICAPS).  [paper]

8

Learning for Planning (III)

Pusse, F. & Klusch, M. (2019): Hybrid Online POMDP Planning and Deep Reinforcement Learning for Safer Self-Driving Cars. Proc. 30th IEEE International Intelligent Vehicles Symposium (IV), I.  [paper] [sup]

9

Explainable Learning Through Rational Reconstruction

Yang, Y., et al. (2022): LOGICDEF: An Interpretable Defense Framework Against Adversarial Examples via Inductive Scene Graph. Proceedings of AAAI Conference on Artificial Intelligence (AAAI).  [paper] 

Selected background references: