Filipe Gouveia
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Filipe Gouveia has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Pedro Tiago Monteiro
INESC-ID, Lisbon, Portugal
6 shared publications - Inês Lynce
Universidade de Lisboa
4 shared publications Inês Lynce
Department of Computer Science and Engineering, INESC-ID/Instituto Superior Técnico, Universidade de Lisboa, Lisboa, Portugal
2 shared publications- Andrew Cropper
University of Oxford
1 shared publication - David Michael Cerna
Czech Academy of Sciences
1 shared publication - José Fragoso Santos
INESC-ID, Lisboa, Portugal
1 shared publication
Latest Publications
- 16 July 2026
- Published by Springer Nature
- p. 56-70
pyModRev: A Python Tool for Model Revision of Boolean Networks
- 16 July 2026
- Published by Springer Nature
- p. 56-70
pyModRev: A Python Tool for Model Revision of Boolean Networks
- Open Access
- 29 April 2026
- Published by AI Access Foundation in Journal of Artificial Intelligence Research
- Vol. 85
Honey, I Shrunk the Hypothesis Space (Through Logical Preprocessing)
Inductive logic programming (ILP) is a form of logical machine learning. The goal is to search a hypothesis space for a hypothesis that generalises training examples and background knowledge. We introduce an approach that shrinks the hypothesis space before an ILP system searches it. Our approach uses background knowledge to find rules that cannot be in an optimal hypothesis regardless of the training examples. For instance, our approach discovers relationships such as even numbers cannot be odd and prime numbers greater than 2 are odd. It then removes violating rules from the hypothesis space. We implement our approach using answer set programming and use it to shrink the hypothesis space of a constraint-based ILP system. Our experiments on multiple domains, including visual reasoning and game playing, show that our approach can substantially reduce learning times whilst maintaining predictive accuracies. For instance, given just 10 seconds of preprocessing time, our approach can reduce learning times from over 10 hours to only 2 seconds.
- 29 April 2026
- Published by AI Access Foundation in Journal of Artificial Intelligence Research
- Vol. 85
Open AccessHoney, I Shrunk the Hypothesis Space (Through Logical Preprocessing)
Inductive logic programming (ILP) is a form of logical machine learning. The goal is to search a hypothesis space for a hypothesis that generalises training examples and background knowledge. We introduce an approach that shrinks the hypothesis space before an ILP system searches it. Our approach uses background knowledge to find rules that cannot be in an optimal hypothesis regardless of the training examples. For instance, our approach discovers relationships such as even numbers cannot be odd and prime numbers greater than 2 are odd. It then removes violating rules from the hypothesis space. We implement our approach using answer set programming and use it to shrink the hypothesis space of a constraint-based ILP system. Our experiments on multiple domains, including visual reasoning and game playing, show that our approach can substantially reduce learning times whilst maintaining predictive accuracies. For instance, given just 10 seconds of preprocessing time, our approach can reduce learning times from over 10 hours to only 2 seconds.
- 16 September 2025
- Published by Springer Nature
- p. 179-191
A MaxSAT Approach for the Train Timetabling Problem with Route Choice and Other Features
- 16 September 2025
- Published by Springer Nature
- p. 179-191
A MaxSAT Approach for the Train Timetabling Problem with Route Choice and Other Features
- Open Access
- 1 May 2025
- Published by Springer Nature
- p. 343-363
Proxy Attribute Discovery in Machine Learning Datasets via Inductive Logic Programming
The issue of fairness is a well-known challenge in Machine Learning (ML) that has gained increased importance with the emergence of Large Language Models (LLMs) and generative AI. Algorithmic bias can manifest during the training of ML models due to the presence of sensitive attributes, such as gender or racial identity. One approach to mitigate bias is to avoid making decisions based on these protected attributes. However, indirect discrimination can still occur if sensitive information is inferred from proxy attributes. To prevent this, there is a growing interest in detecting potential proxy attributes before training ML models. In this case study, we report on the use of Inductive Logic Programming (ILP) to discover proxy attributes in training datasets, with a focus on the ML classification problem. While ILP has established applications in program synthesis and data curation, we demonstrate that it can also advance the state of the art in proxy attribute discovery by removing the need for prior domain knowledge. Our evaluation shows that this approach is effective at detecting potential sources of indirect discrimination, having successfully identified proxy attributes in several well-known datasets used in fairness-awareness studies.
- 1 May 2025
- Published by Springer Nature
- p. 343-363
Open AccessProxy Attribute Discovery in Machine Learning Datasets via Inductive Logic Programming
The issue of fairness is a well-known challenge in Machine Learning (ML) that has gained increased importance with the emergence of Large Language Models (LLMs) and generative AI. Algorithmic bias can manifest during the training of ML models due to the presence of sensitive attributes, such as gender or racial identity. One approach to mitigate bias is to avoid making decisions based on these protected attributes. However, indirect discrimination can still occur if sensitive information is inferred from proxy attributes. To prevent this, there is a growing interest in detecting potential proxy attributes before training ML models. In this case study, we report on the use of Inductive Logic Programming (ILP) to discover proxy attributes in training datasets, with a focus on the ML classification problem. While ILP has established applications in program synthesis and data curation, we demonstrate that it can also advance the state of the art in proxy attribute discovery by removing the need for prior domain knowledge. Our evaluation shows that this approach is effective at detecting potential sources of indirect discrimination, having successfully identified proxy attributes in several well-known datasets used in fairness-awareness studies.
- 29 September 2020
- Published by Springer Nature
- p. 339-348
ModRev - Model Revision Tool for Boolean Logical Models of Biological Regulatory Networks
- 29 September 2020
- Published by Springer Nature
- p. 339-348
ModRev - Model Revision Tool for Boolean Logical Models of Biological Regulatory Networks
- Open Access
Semi-automatic model revision of Boolean regulatory networks: confronting time-series observations with (a)synchronous dynamics
Motivation: Complex cellular processes can be represented by biological regulatory networks. Computational models of such networks have successfully allowed the reprodution of known behaviour and to have a better understanding of the associated cellular processes. However, the construction of these models is still mainly a manual task, and therefore prone to error. Additionally, as new data is acquired, existing models must be revised. Here, we propose a model revision approach of Boolean logical models capable of repairing inconsistent models confronted with time-series observations. Moreover, we account for both synchronous and asynchronous dynamics. Results: The proposed tool is tested on five well known biological models. Different time-series observations are generated, consistent with these models. Then, the models are corrupted with different random changes. The proposed tool is able to repair the majority of the corrupted models, considering the generated time-series observations. Moreover, all the optimal solutions to repair the models are produced. Contact: { filipe.gouveia@tecnico.ulisboa.pt , pedro.tiago.monteiro@tecnico.ulisboa.pt }
Open AccessSemi-automatic model revision of Boolean regulatory networks: confronting time-series observations with (a)synchronous dynamics
Motivation: Complex cellular processes can be represented by biological regulatory networks. Computational models of such networks have successfully allowed the reprodution of known behaviour and to have a better understanding of the associated cellular processes. However, the construction of these models is still mainly a manual task, and therefore prone to error. Additionally, as new data is acquired, existing models must be revised. Here, we propose a model revision approach of Boolean logical models capable of repairing inconsistent models confronted with time-series observations. Moreover, we account for both synchronous and asynchronous dynamics. Results: The proposed tool is tested on five well known biological models. Different time-series observations are generated, consistent with these models. Then, the models are corrupted with different random changes. The proposed tool is able to repair the majority of the corrupted models, considering the generated time-series observations. Moreover, all the optimal solutions to repair the models are produced. Contact: { filipe.gouveia@tecnico.ulisboa.pt , pedro.tiago.monteiro@tecnico.ulisboa.pt }
- 1 February 2020
- Published by SAGE Publications in Journal of Computational Biology
- Vol. 27 (2) , 144-155
Revision of Boolean Models of Regulatory Networks Using Stable State Observations
Models of biological regulatory networks are essential to understand cellular processes. However, the definition of such models is still mostly manually performed, and consequently prone to error. Moreover, as new experimental data are acquired, models need to be revised and updated. Here, we propose a model revision procedure and associated tool, capable of providing the set of minimal repairs to render a model consistent with a set of experimental observations. We consider four possible repair operations, using a lexicographic optimization criterion, giving preference to function repairs over topological ones. Also, we consider observations at stable state discarding the model dynamics. In this article, we extend our previous work to tackle the problem of repairing nodes with multiple reasons of inconsistency. We evaluate our tool on five publicly available logical models. We perform random changes considering several parameter configurations to assess the tool repairing capabilities. Whenever a model is repaired under the time limit, the tool successfully produces the optimal solutions to repair the model. Instances were generated without the previous limitation to validate this extended approach.
- 1 February 2020
- Published by SAGE Publications in Journal of Computational Biology
- Vol. 27 (2) , 144-155
Revision of Boolean Models of Regulatory Networks Using Stable State Observations
Models of biological regulatory networks are essential to understand cellular processes. However, the definition of such models is still mostly manually performed, and consequently prone to error. Moreover, as new experimental data are acquired, models need to be revised and updated. Here, we propose a model revision procedure and associated tool, capable of providing the set of minimal repairs to render a model consistent with a set of experimental observations. We consider four possible repair operations, using a lexicographic optimization criterion, giving preference to function repairs over topological ones. Also, we consider observations at stable state discarding the model dynamics. In this article, we extend our previous work to tackle the problem of repairing nodes with multiple reasons of inconsistency. We evaluate our tool on five publicly available logical models. We perform random changes considering several parameter configurations to assess the tool repairing capabilities. Whenever a model is repaired under the time limit, the tool successfully produces the optimal solutions to repair the model. Instances were generated without the previous limitation to validate this extended approach.
- 9 May 2019
- Published by Springer Nature
- p. 100-112
Model Revision of Boolean Regulatory Networks at Stable State
- 9 May 2019
- Published by Springer Nature
- p. 100-112
Model Revision of Boolean Regulatory Networks at Stable State
- 9 August 2017
- Published by Springer Nature
- p. 657-669
Logic-Based Encodings for Ricochet Robots