TY - GEN AB - Syntactic annotation of corpora in the form of part-of-speech (POS) tags is a key requirement for both linguistic research and subsequent automated natural language processing (NLP) tasks. This problem is commonly tackled using machine learning methods, i.e., by training a POS tagger on a sufficiently large corpus of labeled data. While the problem of POS tagging can essentially be considered as solved for modern languages, historical corpora turn out to be much more difficult, especially due to the lack of native speakers and sparsity of training data. Moreover, most texts have no sentences as we know them today, nor a common orthography. These irregularities render the task of automated POS tagging more difficult and error-prone. Under these circumstances, instead of forcing the POS tagger to predict and commit to a single tag, it should be enabled to express its uncertainty. In this paper, we consider POS tagging within the framework of set-valued prediction, which allows the POS tagger to express its uncertainty via predicting a set of candidate POS tags instead of guessing a single one. The goal is to guarantee a high confidence that the correct POS tag is included while keeping the number of candidates small. In our experimental study, we find that extending state-of-the-art POS taggers to set-valued prediction yields more precise and robust taggings, especially for unknown words, i.e., words not occurring in the training data. AU - Heid, Stefan Helmut AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 17605 T2 - Journal of Data Mining and Digital Humanities TI - Reliable Part-of-Speech Tagging of Historical Corpora through Set-Valued Prediction ER - TY - CONF AU - Tornede, Alexander AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 20306 T2 - Workshop MetaLearn 2020 @ NeurIPS 2020 TI - Towards Meta-Algorithm Selection ER - TY - CHAP AU - El Mesaoudi-Paul, Adil AU - Weiß, Dimitri AU - Bengs, Viktor AU - Hüllermeier, Eyke AU - Tierney, Kevin ID - 18014 SN - 0302-9743 T2 - Learning and Intelligent Optimization. LION 2020. TI - Pool-Based Realtime Algorithm Configuration: A Preselection Bandit Approach VL - 12096 ER - TY - GEN AB - We consider an extension of the contextual multi-armed bandit problem, in which, instead of selecting a single alternative (arm), a learner is supposed to make a preselection in the form of a subset of alternatives. More specifically, in each iteration, the learner is presented a set of arms and a context, both described in terms of feature vectors. The task of the learner is to preselect $k$ of these arms, among which a final choice is made in a second step. In our setup, we assume that each arm has a latent (context-dependent) utility, and that feedback on a preselection is produced according to a Plackett-Luce model. We propose the CPPL algorithm, which is inspired by the well-known UCB algorithm, and evaluate this algorithm on synthetic and real data. In particular, we consider an online algorithm selection scenario, which served as a main motivation of our problem setting. Here, an instance (which defines the context) from a certain problem class (such as SAT) can be solved by different algorithms (the arms), but only $k$ of these algorithms can actually be run. AU - El Mesaoudi-Paul, Adil AU - Bengs, Viktor AU - Hüllermeier, Eyke ID - 18017 T2 - arXiv:2002.04275 TI - Online Preselection with Context Information under the Plackett-Luce Model ER - TY - CONF AB - Algorithm selection (AS) deals with the automatic selection of an algorithm from a fixed set of candidate algorithms most suitable for a specific instance of an algorithmic problem class, where "suitability" often refers to an algorithm's runtime. Due to possibly extremely long runtimes of candidate algorithms, training data for algorithm selection models is usually generated under time constraints in the sense that not all algorithms are run to completion on all instances. Thus, training data usually comprises censored information, as the true runtime of algorithms timed out remains unknown. However, many standard AS approaches are not able to handle such information in a proper way. On the other side, survival analysis (SA) naturally supports censored data and offers appropriate ways to use such data for learning distributional models of algorithm runtime, as we demonstrate in this work. We leverage such models as a basis of a sophisticated decision-theoretic approach to algorithm selection, which we dub Run2Survive. Moreover, taking advantage of a framework of this kind, we advocate a risk-averse approach to algorithm selection, in which the avoidance of a timeout is given high priority. In an extensive experimental study with the standard benchmark ASlib, our approach is shown to be highly competitive and in many cases even superior to state-of-the-art AS approaches. AU - Tornede, Alexander AU - Wever, Marcel Dominik AU - Werner, Stefan AU - Mohr, Felix AU - Hüllermeier, Eyke ID - 18276 T2 - ACML 2020 TI - Run2Survive: A Decision-theoretic Approach to Algorithm Selection based on Survival Analysis ER - TY - JOUR AU - Richter, Cedric AU - Hüllermeier, Eyke AU - Jakobs, Marie-Christine AU - Wehrheim, Heike ID - 16725 JF - Journal of Automated Software Engineering TI - Algorithm Selection for Software Validation Based on Graph Kernels ER - TY - CONF AB - In multi-label classification (MLC), each instance is associated with a set of class labels, in contrast to standard classification where an instance is assigned a single label. Binary relevance (BR) learning, which reduces a multi-label to a set of binary classification problems, one per label, is arguably the most straight-forward approach to MLC. In spite of its simplicity, BR proved to be competitive to more sophisticated MLC methods, and still achieves state-of-the-art performance for many loss functions. Somewhat surprisingly, the optimal choice of the base learner for tackling the binary classification problems has received very little attention so far. Taking advantage of the label independence assumption inherent to BR, we propose a label-wise base learner selection method optimizing label-wise macro averaged performance measures. In an extensive experimental evaluation, we find that or approach, called LiBRe, can significantly improve generalization performance. AU - Wever, Marcel Dominik AU - Tornede, Alexander AU - Mohr, Felix AU - Hüllermeier, Eyke ID - 15629 TI - LiBRe: Label-Wise Selection of Base Learners in Binary Relevance for Multi-Label Classification ER - TY - JOUR AB - In software engineering, the imprecise requirements of a user are transformed to a formal requirements specification during the requirements elicitation process. This process is usually guided by requirements engineers interviewing the user. We want to partially automate this first step of the software engineering process in order to enable users to specify a desired software system on their own. With our approach, users are only asked to provide exemplary behavioral descriptions. The problem of synthesizing a requirements specification from examples can partially be reduced to the problem of grammatical inference, to which we apply an active coevolutionary learning approach. However, this approach would usually require many feedback queries to be sent to the user. In this work, we extend and generalize our active learning approach to receive knowledge from multiple oracles, also known as proactive learning. The ‘user oracle’ represents input received from the user and the ‘knowledge oracle’ represents available, formalized domain knowledge. We call our two-oracle approach the ‘first apply knowledge then query’ (FAKT/Q) algorithm. We compare FAKT/Q to the active learning approach and provide an extensive benchmark evaluation. As result we find that the number of required user queries is reduced and the inference process is sped up significantly. Finally, with so-called On-The-Fly Markets, we present a motivation and an application of our approach where such knowledge is available. AU - Wever, Marcel Dominik AU - van Rooijen, Lorijn AU - Hamann, Heiko ID - 15025 IS - 2 JF - Evolutionary Computation TI - Multi-Oracle Coevolutionary Learning of Requirements Specifications from Examples in On-The-Fly Markets VL - 28 ER - TY - GEN AB - We study the problem of learning choice functions, which play an important role in various domains of application, most notably in the field of economics. Formally, a choice function is a mapping from sets to sets: Given a set of choice alternatives as input, a choice function identifies a subset of most preferred elements. Learning choice functions from suitable training data comes with a number of challenges. For example, the sets provided as input and the subsets produced as output can be of any size. Moreover, since the order in which alternatives are presented is irrelevant, a choice function should be symmetric. Perhaps most importantly, choice functions are naturally context-dependent, in the sense that the preference in favor of an alternative may depend on what other options are available. We formalize the problem of learning choice functions and present two general approaches based on two representations of context-dependent utility functions. Both approaches are instantiated by means of appropriate neural network architectures, and their performance is demonstrated on suitable benchmark tasks. AU - Pfannschmidt, Karlson AU - Gupta, Pritha AU - Hüllermeier, Eyke ID - 19523 T2 - arXiv:1901.10860 TI - Learning Choice Functions: Concepts and Architectures ER - TY - JOUR AU - Merten, Marie-Luis AU - Seemann, Nina AU - Wever, Marcel Dominik ID - 17565 IS - 142 JF - Niederdeutsches Jahrbuch TI - Grammatikwandel digital-kulturwissenschaftlich erforscht. Mittelniederdeutscher Sprachausbau im interdisziplinären Zugriff ER - TY - GEN AB - A common statistical task lies in showing asymptotic normality of certain statistics. In many of these situations, classical textbook results on weak convergence theory suffice for the problem at hand. However, there are quite some scenarios where stronger results are needed in order to establish an asymptotic normal approximation uniformly over a family of probability measures. In this note we collect some results in this direction. We restrict ourselves to weak convergence in $\mathbb R^d$ with continuous limit measures. AU - Bengs, Viktor AU - Holzmann, Hajo ID - 18018 T2 - arXiv:1903.09864 TI - Uniform approximation in classical weak convergence theory ER - TY - GEN AU - Wever, Marcel Dominik AU - Mohr, Felix AU - Hüllermeier, Eyke AU - Hetzer, Alexander ID - 8868 TI - Towards Automated Machine Learning for Multi-Label Classification ER - TY - JOUR AU - Tagne, V. K. AU - Fotso, S. AU - Fono, L. A. AU - Hüllermeier, Eyke ID - 10578 IS - 2 JF - New Mathematics and Natural Computation TI - Choice Functions Generated by Mallows and Plackett–Luce Relations VL - 15 ER - TY - JOUR AU - Couso, Ines AU - Borgelt, Christian AU - Hüllermeier, Eyke AU - Kruse, Rudolf ID - 15001 JF - IEEE Computational Intelligence Magazine SN - 1556-603X TI - Fuzzy Sets in Data Analysis: From Statistical Foundations to Machine Learning ER - TY - JOUR AB - Many problem settings in machine learning are concerned with the simultaneous prediction of multiple target variables of diverse type. Amongst others, such problem settings arise in multivariate regression, multi-label classification, multi-task learning, dyadic prediction, zero-shot learning, network inference, and matrix completion. These subfields of machine learning are typically studied in isolation, without highlighting or exploring important relationships. In this paper, we present a unifying view on what we call multi-target prediction (MTP) problems and methods. First, we formally discuss commonalities and differences between existing MTP problems. To this end, we introduce a general framework that covers the above subfields as special cases. As a second contribution, we provide a structured overview of MTP methods. This is accomplished by identifying a number of key properties, which distinguish such methods and determine their suitability for different types of problems. Finally, we also discuss a few challenges for future research. AU - Waegeman, Willem AU - Dembczynski, Krzysztof AU - Hüllermeier, Eyke ID - 15002 IS - 2 JF - Data Mining and Knowledge Discovery SN - 1573-756X TI - Multi-target prediction: a unifying view on problems and methods VL - 33 ER - TY - CONF AU - Mortier, Thomas AU - Wydmuch, Marek AU - Dembczynski, Krzysztof AU - Hüllermeier, Eyke AU - Waegeman, Willem ID - 15003 T2 - Proceedings of the 31st Benelux Conference on Artificial Intelligence {(BNAIC} 2019) and the 28th Belgian Dutch Conference on Machine Learning (Benelearn 2019), Brussels, Belgium, November 6-8, 2019 TI - Set-Valued Prediction in Multi-Class Classification ER - TY - CHAP AU - Ahmadi Fahandar, Mohsen AU - Hüllermeier, Eyke ID - 15004 SN - 0302-9743 T2 - Discovery Science TI - Feature Selection for Analogy-Based Learning to Rank ER - TY - CHAP AU - Ahmadi Fahandar, Mohsen AU - Hüllermeier, Eyke ID - 15005 SN - 0302-9743 T2 - KI 2019: Advances in Artificial Intelligence TI - Analogy-Based Preference Learning with Kernels ER - TY - CHAP AU - Nguyen, Vu-Linh AU - Destercke, Sébastien AU - Hüllermeier, Eyke ID - 15006 SN - 0302-9743 T2 - Discovery Science TI - Epistemic Uncertainty Sampling ER - TY - CONF AU - Melnikov, Vitaly AU - Hüllermeier, Eyke ID - 15007 T2 - Proceedings ACML, Asian Conference on Machine Learning (Proceedings of Machine Learning Research, 101) TI - Learning to Aggregate: Tackling the Aggregation/Disaggregation Problem for OWA ER - TY - CONF AU - Epple, Nico AU - Dari, Simone AU - Drees, Ludwig AU - Protschky, Valentin AU - Riener, Andreas ID - 15009 SN - 9781728105604 T2 - 2019 IEEE Intelligent Vehicles Symposium (IV) TI - Influence of Cruise Control on Driver Guidance - a Comparison between System Generations and Countries ER - TY - CONF AU - Tornede, Alexander AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ED - Hoffmann, Frank ED - Hüllermeier, Eyke ED - Mikut, Ralf ID - 15011 SN - 978-3-7315-0979-0 T2 - Proceedings - 29. Workshop Computational Intelligence, Dortmund, 28. - 29. November 2019 TI - Algorithm Selection as Recommendation: From Collaborative Filtering to Dyad Ranking ER - TY - CONF AU - Brinker, Klaus AU - Hüllermeier, Eyke ID - 15013 T2 - Proceedings ECML/PKDD, European Conference on Machine Learning and Knowledge Discovery in Databases TI - A Reduction of Label Ranking to Multiclass Classification ER - TY - CONF AU - Hüllermeier, Eyke AU - Couso, Ines AU - Diestercke, Sebastian ID - 15014 T2 - Proceedings SUM 2019, International Conference on Scalable Uncertainty Management TI - Learning from Imprecise Data: Adjustments of Optimistic and Pessimistic Variants ER - TY - JOUR AU - Henzgen, Sascha AU - Hüllermeier, Eyke ID - 15015 JF - ACM Transactions on Knowledge Discovery from Data SN - 1556-4681 TI - Mining Rank Data ER - TY - JOUR AU - Bengs, Viktor AU - Eulert, Matthias AU - Holzmann, Hajo ID - 14027 JF - Journal of Multivariate Analysis SN - 0047-259X TI - Asymptotic confidence sets for the jump curve in bivariate regression problems ER - TY - JOUR AU - Bengs, Viktor AU - Holzmann, Hajo ID - 14028 JF - Electronic Journal of Statistics SN - 1935-7524 TI - Adaptive confidence sets for kink estimation ER - TY - GEN AU - Mohr, Felix AU - Wever, Marcel Dominik AU - Tornede, Alexander AU - Hüllermeier, Eyke ID - 13132 T2 - INFORMATIK 2019: 50 Jahre Gesellschaft für Informatik – Informatik für Gesellschaft TI - From Automated to On-The-Fly Machine Learning ER - TY - CONF AB - Existing tools for automated machine learning, such as Auto-WEKA, TPOT, auto-sklearn, and more recently ML-Plan, have shown impressive results for the tasks of single-label classification and regression. Yet, there is only little work on other types of machine learning problems so far. In particular, there is almost no work on automating the engineering of machine learning solutions for multi-label classification (MLC). We show how the scope of ML-Plan, an AutoML-tool for multi-class classification, can be extended towards MLC using MEKA, which is a multi-label extension of the well-known Java library WEKA. The resulting approach recursively refines MEKA's multi-label classifiers, nesting other multi-label classifiers for meta algorithms and single-label classifiers provided by WEKA as base learners. In our evaluation, we find that the proposed approach yields strong results and performs significantly better than a set of baselines we compare with. AU - Wever, Marcel Dominik AU - Mohr, Felix AU - Tornede, Alexander AU - Hüllermeier, Eyke ID - 10232 TI - Automating Multi-Label Classification Extending ML-Plan ER - TY - JOUR AU - Rohlfing, Katharina AU - Leonardi, Giuseppe AU - Nomikou, Iris AU - Rączaszek-Leonardi, Joanna AU - Hüllermeier, Eyke ID - 20243 JF - IEEE Transactions on Cognitive and Developmental Systems TI - Multimodal Turn-Taking: Motivations, Methodological Challenges, and Novel Approaches ER - TY - CONF AU - Mohr, Felix AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke AU - Faez, Amin ID - 2479 T2 - SCC TI - (WIP) Towards the Automated Composition of Machine Learning Services ER - TY - GEN AB - Object ranking is an important problem in the realm of preference learning. On the basis of training data in the form of a set of rankings of objects, which are typically represented as feature vectors, the goal is to learn a ranking function that predicts a linear order of any new set of objects. Current approaches commonly focus on ranking by scoring, i.e., on learning an underlying latent utility function that seeks to capture the inherent utility of each object. These approaches, however, are not able to take possible effects of context-dependence into account, where context-dependence means that the utility or usefulness of an object may also depend on what other objects are available as alternatives. In this paper, we formalize the problem of context-dependent ranking and present two general approaches based on two natural representations of context-dependent ranking functions. Both approaches are instantiated by means of appropriate neural network architectures, which are evaluated on suitable benchmark task. AU - Pfannschmidt, Karlson AU - Gupta, Pritha AU - Hüllermeier, Eyke ID - 19524 T2 - arXiv:1803.05796 TI - Deep Architectures for Learning Context-dependent Ranking Functions ER - TY - CONF AU - Mohr, Felix AU - Lettmann, Theodor AU - Hüllermeier, Eyke AU - Wever, Marcel Dominik ID - 2857 T2 - Proceedings of the 1st ICAPS Workshop on Hierarchical Planning TI - Programmatic Task Network Planning ER - TY - JOUR AU - Ramaswamy, Arunselvan AU - Bhatnagar, Shalabh ID - 24150 IS - 6 JF - IEEE Transactions on Automatic Control TI - Stability of stochastic approximations with “controlled markov” noise and temporal difference learning VL - 64 ER - TY - JOUR AU - Demirel, Burak AU - Ramaswamy, Arunselvan AU - Quevedo, Daniel E AU - Karl, Holger ID - 24151 IS - 4 JF - IEEE Control Systems Letters TI - Deepcas: A deep reinforcement learning algorithm for control-aware scheduling VL - 2 ER - TY - CONF AU - Mohr, Felix AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 2471 T2 - SCC TI - On-The-Fly Service Construction with Prototypes ER - TY - JOUR AB - In machine learning, so-called nested dichotomies are utilized as a reduction technique, i.e., to decompose a multi-class classification problem into a set of binary problems, which are solved using a simple binary classifier as a base learner. The performance of the (multi-class) classifier thus produced strongly depends on the structure of the decomposition. In this paper, we conduct an empirical study, in which we compare existing heuristics for selecting a suitable structure in the form of a nested dichotomy. Moreover, we propose two additional heuristics as natural completions. One of them is the Best-of-K heuristic, which picks the (presumably) best among K randomly generated nested dichotomies. Surprisingly, and in spite of its simplicity, it turns out to outperform the state of the art. AU - Melnikov, Vitalik AU - Hüllermeier, Eyke ID - 3402 JF - Machine Learning SN - 1573-0565 TI - On the effectiveness of heuristics for learning nested dichotomies: an empirical analysis ER - TY - JOUR AB - Automated machine learning (AutoML) seeks to automatically select, compose, and parametrize machine learning algorithms, so as to achieve optimal performance on a given task (dataset). Although current approaches to AutoML have already produced impressive results, the field is still far from mature, and new techniques are still being developed. In this paper, we present ML-Plan, a new approach to AutoML based on hierarchical planning. To highlight the potential of this approach, we compare ML-Plan to the state-of-the-art frameworks Auto-WEKA, auto-sklearn, and TPOT. In an extensive series of experiments, we show that ML-Plan is highly competitive and often outperforms existing approaches. AU - Mohr, Felix AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 3510 JF - Machine Learning KW - AutoML KW - Hierarchical Planning KW - HTN planning KW - ML-Plan SN - 0885-6125 TI - ML-Plan: Automated Machine Learning via Hierarchical Planning ER - TY - CONF AU - Mohr, Felix AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 3552 T2 - Proceedings of the Symposium on Intelligent Data Analysis TI - Reduction Stumps for Multi-Class Classification ER - TY - CONF AB - In automated machine learning (AutoML), the process of engineering machine learning applications with respect to a specific problem is (partially) automated. Various AutoML tools have already been introduced to provide out-of-the-box machine learning functionality. More specifically, by selecting machine learning algorithms and optimizing their hyperparameters, these tools produce a machine learning pipeline tailored to the problem at hand. Except for TPOT, all of these tools restrict the maximum number of processing steps of such a pipeline. However, as TPOT follows an evolutionary approach, it suffers from performance issues when dealing with larger datasets. In this paper, we present an alternative approach leveraging a hierarchical planning to configure machine learning pipelines that are unlimited in length. We evaluate our approach and find its performance to be competitive with other AutoML tools, including TPOT. AU - Wever, Marcel Dominik AU - Mohr, Felix AU - Hüllermeier, Eyke ID - 3852 KW - automated machine learning KW - complex pipelines KW - hierarchical planning T2 - ICML 2018 AutoML Workshop TI - ML-Plan for Unlimited-Length Machine Learning Pipelines ER - TY - CONF AB - In multinomial classification, reduction techniques are commonly used to decompose the original learning problem into several simpler problems. For example, by recursively bisecting the original set of classes, so-called nested dichotomies define a set of binary classification problems that are organized in the structure of a binary tree. In contrast to the existing one-shot heuristics for constructing nested dichotomies and motivated by recent work on algorithm configuration, we propose a genetic algorithm for optimizing the structure of such dichotomies. A key component of this approach is the proposed genetic representation that facilitates the application of standard genetic operators, while still supporting the exchange of partial solutions under recombination. We evaluate the approach in an extensive experimental study, showing that it yields classifiers with superior generalization performance. AU - Wever, Marcel Dominik AU - Mohr, Felix AU - Hüllermeier, Eyke ID - 2109 KW - Classification KW - Hierarchical Decomposition KW - Indirect Encoding T2 - Proceedings of the Genetic and Evolutionary Computation Conference, GECCO 2018, Kyoto, Japan, July 15-19, 2018 TI - Ensembles of Evolved Nested Dichotomies for Classification ER - TY - GEN AU - Wever, Marcel Dominik AU - Mohr, Felix AU - Hüllermeier, Eyke ID - 17713 TI - Automated Multi-Label Classification based on ML-Plan ER - TY - GEN AU - Mohr, Felix AU - Wever, Marcel Dominik AU - Hüllermeier, Eyke ID - 17714 TI - Automated machine learning service composition ER - TY - GEN AU - Graf, Helena ID - 5693 TI - Ranking of Classification Algorithms in AutoML ER - TY - GEN AU - Scheibl, Manuel ID - 5936 TI - Learning about learning curves from dataset properties ER - TY - CHAP AU - Schäfer, Dirk AU - Hüllermeier, Eyke ID - 6423 SN - 0302-9743 T2 - Discovery Science TI - Preference-Based Reinforcement Learning Using Dyad Ranking ER - TY - GEN ED - Abiteboul, S. ED - Arenas, M. ED - Barceló, P. ED - Bienvenu, M. ED - Calvanese, D. ED - David, C. ED - Hull, R. ED - Hüllermeier, Eyke ED - Kimelfeld, B. ED - Libkin, L. ED - Martens, W. ED - Milo, T. ED - Murlak, F. ED - Neven, F. ED - Ortiz, M. ED - Schwentick, T. ED - Stoyanovich, J. ED - Su, J. ED - Suciu, D. ED - Vianu, V. ED - Yi, K. ID - 10591 IS - 1 TI - Research Directions for Principles of Data Management VL - 7 ER - TY - CHAP AU - Couso, Ines AU - Hüllermeier, Eyke ED - Mostaghim, Sanaz ED - Nürnberger, Andreas ED - Borgelt, Christian ID - 10783 T2 - Frontiers in Computational Intelligence TI - Statistical Inference for Incomplete Ranking Data: A Comparison of two likelihood-based estimators ER - TY - JOUR AU - Schäfer, D. AU - Hüllermeier, Eyke ID - 16038 IS - 5 JF - Machine Learning TI - Dyad ranking using Plackett-Luce models based on joint feature representations VL - 107 ER - TY - CONF AU - Ahmadi Fahandar, Mohsen AU - Hüllermeier, Eyke ID - 10145 T2 - Proc. 32 nd AAAI Conference on Artificial Intelligence (AAAI) TI - Learning to Rank Based on Analogical Reasoning ER -