Publications

Learning to Solve NP-Complete Problems – A Graph Neural Network for the Decision TSP

Published in Proceedings of the AAAI Conference on Artificial Intelligence, 2019

Graph Neural Networks (GNN) are a promising technique for bridging differential programming and combinatorial domains. GNNs employ trainable modules which can be assembled in different configurations that reflect the relational structure of each problem instance. In this paper, we show that GNNs can learn to solve, with very little supervision, the decision variant of the Traveling Salesperson Problem (TSP).

Recommended citation: Prates, M., Avelar, P. H., Lemos, H., Lamb, L. C., & Vardi, M. Y. (2019, July). Learning to solve NP-complete problems: A graph neural network for decision TSP. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 33, pp. 4731-4738). https://www.aaai.org/ojs/index.php/AAAI/article/view/4399

Assessing Gender Bias in Machine Translation: a Case Study with Google Translate

Published in Neural Computing and Applications, 2019

Recently there has been a growing concern in academia, industrial research laboratories and the mainstream commercial media about the phenomenon dubbed as machine bias, where trained statistical models—unbeknownst to their creators—grow to reflect controversial societal asymmetries, such as gender or racial bias. In this paper, we show that Google Translate exhibits a strong tendency toward male defaults when translation job-related sentences, in particular for fields typically associated to unbalanced gender distribution or stereotypes such as STEM (Science, Technology, Engineering and Mathematics) jobs.

Recommended citation: Prates, M. O., Avelar, P. H., & Lamb, L. C. (2018). Assessing gender bias in machine translation: a case study with Google Translate. Neural Computing and Applications, 1-19. https://link.springer.com/article/10.1007/s00521-019-04144-6

Graph Colouring Meets Deep Learning: Effective Graph Neural Network Models for Combinatorial Problems

Published in arXiv Preprint, 2019

In this work, we showcase how Graph Neural Networks (GNN) can be engineered–with a very simple architecture–to solve the fundamental combinatorial problem of graph colouring. Our results show that the model, which achieves high accuracy upon training on random instances, is able to generalise to graph distributions different from those seen at training time.

Recommended citation: Lemos, H., Prates, M., Avelar, P., & Lamb, L. (2019). Graph Colouring Meets Deep Learning: Effective Graph Neural Network Models for Combinatorial Problems. arXiv preprint arXiv:1903.04598. https://arxiv.org/abs/1903.04598

Neural Networks Models for Analyzing Magic: the Gathering Cards

Published in International Conference on Neural Information Processing, 2018

This work aims to apply neural networks models, including Convolutional Neural Networks and Recurrent Neural Networks, in order to analyze Magic: the Gathering cards, both in terms of card text and illustrations; the card images and texts are used to train the networks in order to be able to classify them into multiple categories. The ultimate goal was to develop a methodology that could generate card text matching it to an input image, which was attained by relating the prediction values of the images and generated text across the different categories.

Recommended citation: Zilio, F., Prates, M., & Lamb, L. (2018, December). Neural Networks Models for Analyzing Magic: the Gathering Cards. In International Conference on Neural Information Processing (pp. 227-239). Springer, Cham. https://link.springer.com/chapter/10.1007/978-3-030-04179-3_20

On Quantifying and Understanding the Role of Ethics in AI Research: A Historical Account of Flagship Conferences and Journals

Published in Proceedings of the AAAI Conference on Artificial Intelligence, 2018

We perform long term, historical corpus-based analyses on a large number of flagship conferences and journals. Our experiments identify the prominence of ethics-related terms in published papers and presents several statistics on related topics. Finally, this research provides quantitative evidence on the pressing ethical concerns of the AI community.

Recommended citation: Prates, M., Avelar, P., & Lamb, L. C. (2018). On Quantifying and Understanding the Role of Ethics in AI Research: A Historical Account of Flagship Conferences and Journals. arXiv preprint arXiv:1809.08328. https://easychair.org/publications/paper/Z7D4

Multitask learning on graph neural networks-learning multiple graph centrality measures with a unified network

Published in arXiv Preprint, 2018

In this work, we showcase how Graph Neural Networks (GNN) can be engineered–with a very simple architecture–to solve the fundamental combinatorial problem of graph colouring. Our results show that the model, which achieves high accuracy upon training on random instances, is able to generalise to graph distributions different from those seen at training time.

Recommended citation: Avelar, P. H., Lemos, H., Prates, M. O., & Lamb, L. (2018). Multitask learning on graph neural networks-learning multiple graph centrality measures with a unified network. arXiv preprint arXiv:1809.07695. https://arxiv.org/abs/1809.07695

Collaboration in social problem-solving: When diversity trumps network efficiency

Published in Proceedings of the AAAI Conference on Artificial Intelligence, 2015

In this paper we analyse a recent social problem-solving model and attempt to address its shortcomings. Specifically, we investigate the effects of separating exploitation from exploration in agent behaviors and explore the concept of diversity in such models. We found out that diverse populations outperform homogeneous ones in both efficient and inefficient networks. Finally, we show that agent diversity is more relevant than the strategic behavioral dynamics. This work contributes towards understanding the role of diverse and dynamic behaviors in social problem-solving as well as the advancement of state-of-art social problem-solving models.

Recommended citation: Noble, D., Prates, M., Bossle, D., & Lamb, L. (2015, February). Collaboration in social problem-solving: When diversity trumps network efficiency. In Twenty-Ninth AAAI Conference on Artificial Intelligence. https://arxiv.org/abs/1903.04598

Leveraging collaboration: A methodology for the design of social problem-solving systems

Published in First AAAI Conference on Human Computation and Crowdsourcing (HCOMP), 2013

We analyse the strategies humans follow when solving a problem, comparing them with alternative ones, and identify the consequences of the employed strategies in the collective performance of the social network. Our results also indicate that copying and guessing are beneficial to the performance of the social networks. We then propose mechanisms that can improve collaborative problem-solving. Finally, we show that our results lead to a methodology for the design of efficient problem-solving systems that can be applied to several kinds of collaborative social systems.

Recommended citation: Tabajara, L. M., Prates, M. O., Noble, D. V., & Lamb, L. C. (2013, November). Leveraging collaboration: A methodology for the design of social problem-solving systems. In First AAAI Conference on Human Computation and Crowdsourcing. https://www.aaai.org/ocs/index.php/HCOMP/HCOMP13/paper/view/7481