TY - CONF AB - Graph neural networks (GNNs) have been successfully applied in many structured data domains, with applications ranging from molecular property prediction to the analysis of social networks. Motivated by the broad applicability of GNNs, we propose the family of so-called RankGNNs, a combination of neural Learning to Rank (LtR) methods and GNNs. RankGNNs are trained with a set of pair-wise preferences between graphs, suggesting that one of them is preferred over the other. One practical application of this problem is drug screening, where an expert wants to find the most promising molecules in a large collection of drug candidates. We empirically demonstrate that our proposed pair-wise RankGNN approach either significantly outperforms or at least matches the ranking performance of the naive point-wise baseline approach, in which the LtR problem is solved via GNN-based graph regression. AU - Damke, Clemens AU - Hüllermeier, Eyke ED - Soares, Carlos ED - Torgo, Luis ID - 27381 KW - Graph-structured data KW - Graph neural networks KW - Preference learning KW - Learning to rank SN - 0302-9743 T2 - Proceedings of The 24th International Conference on Discovery Science (DS 2021) TI - Ranking Structured Objects with Graph Neural Networks VL - 12986 ER - TY - THES AU - Wever, Marcel Dominik ID - 27284 TI - Automated Machine Learning for Multi-Label Classification ER - TY - CONF AU - Hanselle, Jonas Manuel AU - Tornede, Alexander AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 21198 TI - Algorithm Selection as Superset Learning: Constructing Algorithm Selectors from Imprecise Performance Data ER - TY - CHAP AU - Pfannschmidt, Karlson AU - Hüllermeier, Eyke ID - 19521 SN - 0302-9743 T2 - Lecture Notes in Computer Science TI - Learning Choice Functions via Pareto-Embeddings ER - TY - CONF AB - Current GNN architectures use a vertex neighborhood aggregation scheme, which limits their discriminative power to that of the 1-dimensional Weisfeiler-Lehman (WL) graph isomorphism test. Here, we propose a novel graph convolution operator that is based on the 2-dimensional WL test. We formally show that the resulting 2-WL-GNN architecture is more discriminative than existing GNN approaches. This theoretical result is complemented by experimental studies using synthetic and real data. On multiple common graph classification benchmarks, we demonstrate that the proposed model is competitive with state-of-the-art graph kernels and GNNs. AU - Damke, Clemens AU - Melnikov, Vitaly AU - Hüllermeier, Eyke ED - Jialin Pan, Sinno ED - Sugiyama, Masashi ID - 19953 KW - graph neural networks KW - Weisfeiler-Lehman test KW - cycle detection T2 - Proceedings of the 12th Asian Conference on Machine Learning (ACML 2020) TI - A Novel Higher-order Weisfeiler-Lehman Graph Convolution VL - 129 ER - TY - CONF AU - Bengs, Viktor AU - Hüllermeier, Eyke ID - 21534 T2 - International Conference on Machine Learning TI - Preselection Bandits ER - TY - GEN AB - We consider a resource-aware variant of the classical multi-armed bandit problem: In each round, the learner selects an arm and determines a resource limit. It then observes a corresponding (random) reward, provided the (random) amount of consumed resources remains below the limit. Otherwise, the observation is censored, i.e., no reward is obtained. For this problem setting, we introduce a measure of regret, which incorporates the actual amount of allocated resources of each learning round as well as the optimality of realizable rewards. Thus, to minimize regret, the learner needs to set a resource limit and choose an arm in such a way that the chance to realize a high reward within the predefined resource limit is high, while the resource limit itself should be kept as low as possible. We derive the theoretical lower bound on the cumulative regret and propose a learning algorithm having a regret upper bound that matches the lower bound. In a simulation study, we show that our learning algorithm outperforms straightforward extensions of standard multi-armed bandit algorithms. AU - Bengs, Viktor AU - Hüllermeier, Eyke ID - 21536 T2 - arXiv:2011.00813 TI - Multi-Armed Bandits with Censored Consumption of Resources ER - TY - CONF AU - Tornede, Alexander AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 17407 T2 - Discovery Science TI - Extreme Algorithm Selection with Dyadic Feature Representation ER - TY - CONF AU - Hanselle, Jonas Manuel AU - Tornede, Alexander AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 17408 T2 - KI 2020: Advances in Artificial Intelligence TI - Hybrid Ranking and Regression for Algorithm Selection ER - TY - CONF AU - Tornede, Tanja AU - Tornede, Alexander AU - Wever, Marcel Dominik AU - Mohr, Felix AU - Hüllermeier, Eyke ID - 17424 T2 - Proceedings of the ECMLPKDD 2020 TI - AutoML for Predictive Maintenance: One Tool to RUL Them All ER -