Paweł Liskowski
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Paweł Liskowski has a total of 18 co-authors. Here are the top co-authors by number of shared publications:
- Krzysztof Krawiec
Poznan University of Technology
7 shared publications - Marcin Szubert
Poznan University of Technology
2 shared publications - Wojciech Jaśkowski
Institute of Computing Science Poznań University of Technology Poznań Poland
2 shared publications - Aaron Lin
Snowflake Inc., Menlo Park, CA, USA
1 shared publication - Anupam Datta
Snowflake Inc., Menlo Park, CA, USA
1 shared publication - Benjamin Han
Snowflake Inc., Menlo Park, CA, USA
1 shared publication
Latest Publications
- Open Access
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 400-412
Cortex AISQL: A Production SQL Engine for Unstructured Data
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 400-412
Open AccessCortex AISQL: A Production SQL Engine for Unstructured Data
- 2 September 2020
- Published by Springer Nature
- p. 515-527
ClipUp: A Simple and Powerful Optimizer for Distribution-Based Policy Evolution
- 2 September 2020
- Published by Springer Nature
- p. 515-527
ClipUp: A Simple and Powerful Optimizer for Distribution-Based Policy Evolution
- Open Access
- 9 December 2017
- Published by Walter de Gruyter GmbH in Foundations of Computing and Decision Sciences
- Vol. 42 (4) , 339-358
Adaptive Test Selection for Factorization-based Surrogate Fitness in Genetic Programming
Genetic programming (GP) is a variant of evolutionary algorithm where the entities undergoing simulated evolution are computer programs. A fitness function in GP is usually based on a set of tests, each of which defines the desired output a correct program should return for an exemplary input. The outcomes of interactions between programs and tests in GP can be represented as an interaction matrix, with rows corresponding to programs in the current population and columns corresponding to tests. In previous work, we proposed SFIMX, a method that performs only a fraction of interactions and employs non-negative matrix factorization to estimate the outcomes of remaining ones, shortening GP’s runtime. In this paper, we build upon that work and propose three extensions of SFIMX, in which the subset of tests drawn to perform interactions is selected with respect to test difficulty. The conducted experiment indicates that the proposed extensions surpass the original SFIMX on a suite of discrete GP benchmarks.
- 9 December 2017
- Published by Walter de Gruyter GmbH in Foundations of Computing and Decision Sciences
- Vol. 42 (4) , 339-358
Open AccessAdaptive Test Selection for Factorization-based Surrogate Fitness in Genetic Programming
Genetic programming (GP) is a variant of evolutionary algorithm where the entities undergoing simulated evolution are computer programs. A fitness function in GP is usually based on a set of tests, each of which defines the desired output a correct program should return for an exemplary input. The outcomes of interactions between programs and tests in GP can be represented as an interaction matrix, with rows corresponding to programs in the current population and columns corresponding to tests. In previous work, we proposed SFIMX, a method that performs only a fraction of interactions and employs non-negative matrix factorization to estimate the outcomes of remaining ones, shortening GP’s runtime. In this paper, we build upon that work and propose three extensions of SFIMX, in which the subset of tests drawn to perform interactions is selected with respect to test difficulty. The conducted experiment indicates that the proposed extensions surpass the original SFIMX on a suite of discrete GP benchmarks.
- 1 September 2017
- Published by MIT Press in Evolutionary Computation
- Vol. 25 (3) , 375-406
Online Discovery of Search Objectives for Test-Based Problems
In test-based problems, commonly approached with competitive coevolutionary algorithms, the fitness of a candidate solution is determined by the outcomes of its interactions with multiple tests. Usually, fitness is a scalar aggregate of interaction outcomes, and as such imposes a complete order on the candidate solutions. However, passing different tests may require unrelated “skills,” and candidate solutions may vary with respect to such capabilities. In this study, we provide theoretical evidence that scalar fitness, inherently incapable of capturing such differences, is likely to lead to premature convergence. To mitigate this problem, we propose disco, a method that automatically identifies the groups of tests for which the candidate solutions behave similarly and define the above skills. Each such group gives rise to a derived objective, and these objectives together guide the search algorithm in multi-objective fashion. When applied to several well-known test-based problems, the proposed approach significantly outperforms the conventional two-population coevolution. This opens the door to efficient and generic countermeasures to premature convergence for both coevolutionary and evolutionary algorithms applied to problems featuring aggregating fitness functions.
- 1 September 2017
- Published by MIT Press in Evolutionary Computation
- Vol. 25 (3) , 375-406
Online Discovery of Search Objectives for Test-Based Problems
In test-based problems, commonly approached with competitive coevolutionary algorithms, the fitness of a candidate solution is determined by the outcomes of its interactions with multiple tests. Usually, fitness is a scalar aggregate of interaction outcomes, and as such imposes a complete order on the candidate solutions. However, passing different tests may require unrelated “skills,” and candidate solutions may vary with respect to such capabilities. In this study, we provide theoretical evidence that scalar fitness, inherently incapable of capturing such differences, is likely to lead to premature convergence. To mitigate this problem, we propose disco, a method that automatically identifies the groups of tests for which the candidate solutions behave similarly and define the above skills. Each such group gives rise to a derived objective, and these objectives together guide the search algorithm in multi-objective fashion. When applied to several well-known test-based problems, the proposed approach significantly outperforms the conventional two-population coevolution. This opens the door to efficient and generic countermeasures to premature convergence for both coevolutionary and evolutionary algorithms applied to problems featuring aggregating fitness functions.
- Top Cited Papers
- 24 March 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 35 (11) , 2369-2380
Segmenting Retinal Blood Vessels WithDeep Neural Networks
The condition of the vascular network of human eye is an important diagnostic factor in ophthalmology. Its segmentation in fundus imaging is a nontrivial task due to variable size of vessels, relatively low contrast, and potential presence of pathologies like microaneurysms and hemorrhages. Many algorithms, both unsupervised and supervised, have been proposed for this purpose in the past. We propose a supervised segmentation technique that uses a deep neural network trained on a large (up to 400 \thinspace000) sample of examples preprocessed with global contrast normalization, zero-phase whitening, and augmented using geometric transformations and gamma corrections. Several variants of the method are considered, including structured prediction, where a network classifies multiple pixels simultaneously. When applied to standard benchmarks of fundus imaging, the DRIVE, STARE, and CHASE databases, the networks significantly outperform the previous algorithms on the area under ROC curve measure (up to > 0.99) and accuracy of classification (up to > 0.97). The method is also resistant to the phenomenon of central vessel reflex, sensitive in detection of fine vessels ( sensitivity > 0.87), and fares well on pathological cases.
- 24 March 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 35 (11) , 2369-2380
Top Cited PapersSegmenting Retinal Blood Vessels WithDeep Neural Networks
The condition of the vascular network of human eye is an important diagnostic factor in ophthalmology. Its segmentation in fundus imaging is a nontrivial task due to variable size of vessels, relatively low contrast, and potential presence of pathologies like microaneurysms and hemorrhages. Many algorithms, both unsupervised and supervised, have been proposed for this purpose in the past. We propose a supervised segmentation technique that uses a deep neural network trained on a large (up to 400 \thinspace000) sample of examples preprocessed with global contrast normalization, zero-phase whitening, and augmented using geometric transformations and gamma corrections. Several variants of the method are considered, including structured prediction, where a network classifies multiple pixels simultaneously. When applied to standard benchmarks of fundus imaging, the DRIVE, STARE, and CHASE databases, the networks significantly outperform the previous algorithms on the area under ROC curve measure (up to > 0.99) and accuracy of classification (up to > 0.97). The method is also resistant to the phenomenon of central vessel reflex, sensitive in detection of fine vessels ( sensitivity > 0.87), and fares well on pathological cases.
- 24 March 2016
- Published by Springer Nature
- p. 68-82
Surrogate Fitness via Factorization of Interaction Matrix
- 24 March 2016
- Published by Springer Nature
- p. 68-82
Surrogate Fitness via Factorization of Interaction Matrix
- Open Access
- 1 March 2016
- Published by University of Zielona Góra, Poland in International Journal of Applied Mathematics and Computer Science
- Vol. 26 (1) , 215-229
The performance profile: A multi–criteria performance evaluation method for test–based problems
In test-based problems, solutions produced by search algorithms are typically assessed using average outcomes of interactions with multiple tests. This aggregation leads to information loss, which can render different solutions apparently indifferent and hinder comparison of search algorithms. In this paper we introduce the performance profile, a generic, domain-independent, multi-criteria performance evaluation method that mitigates this problem by characterizing the performance of a solution by a vector of outcomes of interactions with tests of various difficulty. To demonstrate the usefulness of this gauge, we employ it to analyze the behavior of Othello and Iterated Prisoner’s Dilemma players produced by five (co)evolutionary algorithms as well as players known from previous publications. Performance profiles reveal interesting differences between the players, which escape the attention of the scalar performance measure of the expected utility. In particular, they allow us to observe that evolution with random sampling produces players coping well against the mediocre opponents, while the coevolutionary and temporal difference learning strategies play better against the high-grade opponents. We postulate that performance profiles improve our understanding of characteristics of search algorithms applied to arbitrary test-based problems, and can prospectively help design better methods for interactive domains.
- 1 March 2016
- Published by University of Zielona Góra, Poland in International Journal of Applied Mathematics and Computer Science
- Vol. 26 (1) , 215-229
Open AccessThe performance profile: A multi–criteria performance evaluation method for test–based problems
In test-based problems, solutions produced by search algorithms are typically assessed using average outcomes of interactions with multiple tests. This aggregation leads to information loss, which can render different solutions apparently indifferent and hinder comparison of search algorithms. In this paper we introduce the performance profile, a generic, domain-independent, multi-criteria performance evaluation method that mitigates this problem by characterizing the performance of a solution by a vector of outcomes of interactions with tests of various difficulty. To demonstrate the usefulness of this gauge, we employ it to analyze the behavior of Othello and Iterated Prisoner’s Dilemma players produced by five (co)evolutionary algorithms as well as players known from previous publications. Performance profiles reveal interesting differences between the players, which escape the attention of the scalar performance measure of the expected utility. In particular, they allow us to observe that evolution with random sampling produces players coping well against the mediocre opponents, while the coevolutionary and temporal difference learning strategies play better against the high-grade opponents. We postulate that performance profiles improve our understanding of characteristics of search algorithms applied to arbitrary test-based problems, and can prospectively help design better methods for interactive domains.
- 17 March 2015
- Published by Springer Nature
- p. 394-405
The Role of Behavioral Diversity and Difficulty of Opponents in Coevolving Game-Playing Agents
- 17 March 2015
- Published by Springer Nature
- p. 394-405
The Role of Behavioral Diversity and Difficulty of Opponents in Coevolving Game-Playing Agents
- 15 March 2015
- Published by Springer Nature
- p. 53-65
Automatic Derivation of Search Objectives for Test-Based Genetic Programming
- 15 March 2015
- Published by Springer Nature
- p. 53-65
Automatic Derivation of Search Objectives for Test-Based Genetic Programming