Search Publications
Publications
2019 |
Goel, Naman; Faltings, Boi Deep Bayesian Trust : A Dominant and Fair Incentive Mechanism for Crowd Conference Forthcoming 33rd AAAI Conference on Artificial Intelligence, 2019, AAAI, Forthcoming. @conference{Goel2019, title = {Deep Bayesian Trust : A Dominant and Fair Incentive Mechanism for Crowd}, author = {Naman Goel and Boi Faltings}, year = {2019}, date = {2019-01-27}, booktitle = {33rd AAAI Conference on Artificial Intelligence, 2019}, publisher = {AAAI}, keywords = {}, pubstate = {forthcoming}, tppubtype = {conference} } |
2018 |
Goel, Naman; Yaghini, Mohammad; Faltings, Boi Non-Discriminatory Machine Learning through Convex Fairness Criteria Conference Thirty-Second AAAI Conference on Artificial Intelligence, 2018, AAAI, 2018. @conference{Goel2018, title = {Non-Discriminatory Machine Learning through Convex Fairness Criteria}, author = {Naman Goel and Mohammad Yaghini and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/Goel_Naman_AAAI.pdf}, year = {2018}, date = {2018-02-07}, booktitle = {Thirty-Second AAAI Conference on Artificial Intelligence, 2018}, publisher = {AAAI}, keywords = {}, pubstate = {published}, tppubtype = {conference} } |
Goran Radanovic, Boi Faltings Partial Truthfulness in Minimal Peer Prediction Mechanisms with Limited Knowledge Proceeding 32nd AAAI Conference on Artificial Intelligence 2018. @proceedings{radanovic2018a, title = {Partial Truthfulness in Minimal Peer Prediction Mechanisms with Limited Knowledge}, author = {Goran Radanovic, Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/2018/01/Radanovic.pdf}, year = {2018}, date = {2018-02-04}, organization = {32nd AAAI Conference on Artificial Intelligence}, keywords = {}, pubstate = {published}, tppubtype = {proceedings} } |
Goran Radanovic Adish Singla, Andreas Krause Boi Faltings Information Gathering with Peers: Submodular Optimization with Peer-Prediction Constraints Inproceedings Proceedings of the 32nd Conference of the AAAI, 2018. @inproceedings{Radanovic2018b, title = {Information Gathering with Peers: Submodular Optimization with Peer-Prediction Constraints}, author = {Goran Radanovic, Adish Singla, Andreas Krause, Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/2018/01/Radanovic-Singla.pdf }, year = {2018}, date = {2018-02-04}, booktitle = {Proceedings of the 32nd Conference of the AAAI}, keywords = {}, pubstate = {published}, tppubtype = {inproceedings} } |
Kulev, Igor; Pu, Pearl; Faltings, Boi A Bayesian Approach to Intervention-Based Clustering Journal Article ACM Trans. Intell. Syst. Technol., 9 (4), pp. 44:1–44:23, 2018, ISSN: 2157-6904. @article{Kulev:2018:BAI:3183892.3156683, title = {A Bayesian Approach to Intervention-Based Clustering}, author = {Igor Kulev and Pearl Pu and Boi Faltings}, url = {http://doi.acm.org/10.1145/3156683}, doi = {10.1145/3156683}, issn = {2157-6904}, year = {2018}, date = {2018-01-01}, journal = {ACM Trans. Intell. Syst. Technol.}, volume = {9}, number = {4}, pages = {44:1--44:23}, publisher = {ACM}, address = {New York, NY, USA}, keywords = {}, pubstate = {published}, tppubtype = {article} } |
2017 |
Giannakopoulos, Athanasios; Antognini, Diego; Musat, Claudiu; Hossmann, Andreea; Baeriswyl, Michael Dataset Construction via Attention for Aspect Term Extraction with Distant Supervision Proceeding Forthcoming 7th ICDM Workshop on Sentiment Elicitation from Natural Text for Information Retrieval and Extraction (SENTIRE), Forthcoming. @proceedings{giannakopoulos2017b, title = {Dataset Construction via Attention for Aspect Term Extraction with Distant Supervision}, author = {Athanasios Giannakopoulos and Diego Antognini and Claudiu Musat and Andreea Hossmann and Michael Baeriswyl}, url = {https://lia.epfl.ch/wp-content/uploads/publications/Giannakopoulos2017b.pdf}, year = {2017}, date = {2017-11-18}, publisher = {7th ICDM Workshop on Sentiment Elicitation from Natural Text for Information Retrieval and Extraction (SENTIRE)}, abstract = {Aspect Term Extraction (ATE) detects opinionated aspect terms in sentences or text spans, with the end goal of performing aspect-based sentiment analysis. The small amount of available datasets for supervised ATE and the fact that they cover only a few domains raise the need for exploiting other data sources in new and creative ways. Publicly available review corpora contain a plethora of opinionated aspect terms and cover a larger domain spectrum. In this paper, we first propose a method for using such review corpora for creating a new dataset for ATE. Our method relies on an attention mechanism to select sentences that have a high likelihood of containing actual opinionated aspects. We thus improve the quality of the extracted aspects. We then use the constructed dataset to train a model and perform ATE with distant supervision. By evaluating on human annotated datasets, we prove that our method achieves a significantly improved performance over various unsupervised and supervised baselines. Finally, we prove that sentence selection matters when it comes to creating new datasets for ATE. Specifically, we show that, using a set of selected sentences leads to higher ATE performance compared to using the whole sentence set.}, keywords = {}, pubstate = {forthcoming}, tppubtype = {proceedings} } Aspect Term Extraction (ATE) detects opinionated aspect terms in sentences or text spans, with the end goal of performing aspect-based sentiment analysis. The small amount of available datasets for supervised ATE and the fact that they cover only a few domains raise the need for exploiting other data sources in new and creative ways. Publicly available review corpora contain a plethora of opinionated aspect terms and cover a larger domain spectrum. In this paper, we first propose a method for using such review corpora for creating a new dataset for ATE. Our method relies on an attention mechanism to select sentences that have a high likelihood of containing actual opinionated aspects. We thus improve the quality of the extracted aspects. We then use the constructed dataset to train a model and perform ATE with distant supervision. By evaluating on human annotated datasets, we prove that our method achieves a significantly improved performance over various unsupervised and supervised baselines. Finally, we prove that sentence selection matters when it comes to creating new datasets for ATE. Specifically, we show that, using a set of selected sentences leads to higher ATE performance compared to using the whole sentence set. |
Radanovic, Boi Faltings Goran Game Theory for Data Science: Eliciting Truthful Information Book Morgan & Claypool Publishers, 2017, ISBN: 9781627057295. @book{Radanovic2017, title = {Game Theory for Data Science: Eliciting Truthful Information}, author = {Boi Faltings Goran Radanovic}, doi = {10.2200/S00788ED1V01Y201707AIM035}, isbn = {9781627057295}, year = {2017}, date = {2017-09-21}, volume = {35}, publisher = {Morgan & Claypool Publishers}, keywords = {}, pubstate = {published}, tppubtype = {book} } |
Brammert Ottens Christos Dimitrakakis, Boi Faltings DUCT: An Upper Confidence Bound Approach to Distributed Constraint Optimization Problems Journal Article ACM Transactions on Intelligent Systems and Technology (TIST), 8 (5), pp. 69:1-69:27, 2017. @article{Ottens2017, title = { DUCT: An Upper Confidence Bound Approach to Distributed Constraint Optimization Problems }, author = {Brammert Ottens , Christos Dimitrakakis , Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/DUCT-TIST.pdf}, doi = {https://doi.org/10.1145/3066156}, year = {2017}, date = {2017-07-12}, journal = {ACM Transactions on Intelligent Systems and Technology (TIST)}, volume = {8}, number = {5}, pages = {69:1-69:27}, abstract = {We propose a distributed upper confidence bound approach, DUCT, for solving distributed constraint optimization problems. We compare four variants of this approach with a baseline random sampling algorithm, as well as other complete and incomplete algorithms for DCOPs. Under general assumptions, we theoretically show that the solution found by DUCT after T steps is approximately T−1-close to the optimal. Experimentally, we show that DUCT matches the optimal solution found by the well-known DPOP and O-DPOP algorithms on moderate-size problems, while always requiring less agent communication. For larger problems, where DPOP fails, we show that DUCT produces significantly better solutions than local, incomplete algorithms. Overall, we believe that DUCT is a practical, scalable algorithm for complex DCOPs.}, keywords = {}, pubstate = {published}, tppubtype = {article} } We propose a distributed upper confidence bound approach, DUCT, for solving distributed constraint optimization problems. We compare four variants of this approach with a baseline random sampling algorithm, as well as other complete and incomplete algorithms for DCOPs. Under general assumptions, we theoretically show that the solution found by DUCT after T steps is approximately T−1-close to the optimal. Experimentally, we show that DUCT matches the optimal solution found by the well-known DPOP and O-DPOP algorithms on moderate-size problems, while always requiring less agent communication. For larger problems, where DPOP fails, we show that DUCT produces significantly better solutions than local, incomplete algorithms. Overall, we believe that DUCT is a practical, scalable algorithm for complex DCOPs. |
Fei Mi, Boi Faltings Adaptive Sequential Recommendation for Discussion Forums on MOOCs using Context Trees Conference Proceedings of the 10th International Conference on Educational Data Mining, 2017. @conference{miadaptive, title = {Adaptive Sequential Recommendation for Discussion Forums on MOOCs using Context Trees}, author = {Fei Mi, Boi Faltings}, url = {http://educationaldatamining.org/EDM2017/proc_files/papers/paper_25.pdf}, year = {2017}, date = {2017-06-28}, booktitle = {Proceedings of the 10th International Conference on Educational Data Mining}, abstract = {Massive open online courses (MOOCs) have demonstrated growing popularity and rapid development in recent years. Discussion forums have become crucial components for students and instructors to widely exchange ideas and propagate knowledge. It is important to recommend helpful information from forums to students for the benefit of the learning process. However, students or instructors update discussion forums very often, and the student preferences over forum contents shift rapidly as a MOOC progresses. So, MOOC forum recommendations need to be adaptive to these evolving forum contents and drifting student interests. These frequent changes pose a challenge to most standard recommendation methods as they have difficulty adapting to new and drifting observations. We formalize the discussion forum recommendation problem as a sequence prediction problem. Then we compare different methods, including a new method called context tree (CT), which can be effectively applied to online sequential recommendation tasks. The results show that the CT recommender performs better than other methods for MOOCs forum recommendation task. We analyze the reasons for this and demonstrate that it is because of better adaptation to changes in the domain. This highlights the importance of considering the adaptation aspect when building recommender system with drifting preferences, as well as using machine learning in general.}, keywords = {}, pubstate = {published}, tppubtype = {conference} } Massive open online courses (MOOCs) have demonstrated growing popularity and rapid development in recent years. Discussion forums have become crucial components for students and instructors to widely exchange ideas and propagate knowledge. It is important to recommend helpful information from forums to students for the benefit of the learning process. However, students or instructors update discussion forums very often, and the student preferences over forum contents shift rapidly as a MOOC progresses. So, MOOC forum recommendations need to be adaptive to these evolving forum contents and drifting student interests. These frequent changes pose a challenge to most standard recommendation methods as they have difficulty adapting to new and drifting observations. We formalize the discussion forum recommendation problem as a sequence prediction problem. Then we compare different methods, including a new method called context tree (CT), which can be effectively applied to online sequential recommendation tasks. The results show that the CT recommender performs better than other methods for MOOCs forum recommendation task. We analyze the reasons for this and demonstrate that it is because of better adaptation to changes in the domain. This highlights the importance of considering the adaptation aspect when building recommender system with drifting preferences, as well as using machine learning in general. |
Boi Faltings, Michael Schumacher L'intelligence artificielle par la pratique Book PPUR, 2017, ISBN: 978-2-88915-075-5. @book{Faltings2017, title = {L'intelligence artificielle par la pratique}, author = {Boi Faltings, Michael Schumacher}, isbn = {978-2-88915-075-5}, year = {2017}, date = {2017-02-01}, publisher = {PPUR}, keywords = {}, pubstate = {published}, tppubtype = {book} } |
Gujar, Sujit Prakash; Faltings, Boi Auction Based Mechanisms for Dynamic Task Assignments in Expert Crowdsourcing Book Chapter et al., Ceppi S (Ed.): pp. 50-65, Springer, 2017, ISBN: 978-3-319-54229-4. @inbook{Gujar2017, title = {Auction Based Mechanisms for Dynamic Task Assignments in Expert Crowdsourcing}, author = {Sujit Prakash Gujar and Boi Faltings}, editor = {S. Ceppi et al.}, url = {https://lia.epfl.ch/wp-content/uploads/publications/auction-based-mechanisms.pdf}, isbn = {978-3-319-54229-4}, year = {2017}, date = {2017-00-00}, journal = {Agent-Mediated Electronic Commerce}, pages = {50-65}, publisher = {Springer}, series = {LNBIP}, abstract = {Crowdsourcing marketplaces link large populations of workers to an even larger number of tasks. Thus, it is necessary to have mechanisms for matching workers with interesting and suitable tasks. Earlier work has addressed the problem of finding optimal workers for a given set of tasks. However, workers also have preferences and will stay with a platform only if it gives them interesting tasks. We therefore analyze several matching mechanisms that take into account workers’ preferences as well.We propose that the workers pay premiums to get preferred matches and auction-based models where preferences are expressed through variations of the payment for a task. We analyze the properties of two matching different mechanisms: Split Dynamic VCG (SDV) and e-Auction. We compare both the mechanisms with Arrival Priority Serial Dictatorship (APSD) empirically for efficiency.}, keywords = {}, pubstate = {published}, tppubtype = {inbook} } Crowdsourcing marketplaces link large populations of workers to an even larger number of tasks. Thus, it is necessary to have mechanisms for matching workers with interesting and suitable tasks. Earlier work has addressed the problem of finding optimal workers for a given set of tasks. However, workers also have preferences and will stay with a platform only if it gives them interesting tasks. We therefore analyze several matching mechanisms that take into account workers’ preferences as well.We propose that the workers pay premiums to get preferred matches and auction-based models where preferences are expressed through variations of the payment for a task. We analyze the properties of two matching different mechanisms: Split Dynamic VCG (SDV) and e-Auction. We compare both the mechanisms with Arrival Priority Serial Dictatorship (APSD) empirically for efficiency. |
2016 |
Radanovic, Goran; Faltings, Boi; Jurca, Radu Incentives for Effort in Crowdsourcing using the Peer Truth Serum Journal Article Acm Transactions on Intelligent Systems and Technology, 7 (4), pp. 48:1-28, 2016. @article{Radanvoic 2016b, title = {Incentives for Effort in Crowdsourcing using the Peer Truth Serum}, author = {Goran Radanovic and Boi Faltings and Radu Jurca}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ PTS.pdf}, doi = {http://dx.doi.org/10.1145/2856102}, year = {2016}, date = {2016-00-00}, journal = {Acm Transactions on Intelligent Systems and Technology}, volume = {7}, number = {4}, pages = {48:1-28}, abstract = {Crowdsourcing is widely proposed as a method to solve large variety of judgement tasks, such as classifying website content, peer grading in online courses, or collecting real-world data. As the data reported by workers cannot be verified, there is a tendency to report random data without actually solving the task. This can be countered by making the reward for an answer depend on its consistency with answers given by other workers, an approach called peer consistency. However, it is obvious that the best strategy in such schemes is for all workers to report the same answer without solving the task. Dasgupta and Ghosh (WWW 2013) show that in some cases exerting high effort can be encouraged in the highest-paying equilibrium. In this paper we present a general mechanism that implements this idea and is applicable to most crowdsourcing settings. Furthermore, we experimentally test the novel mechanism, and validate its theoretical properties.}, keywords = {}, pubstate = {published}, tppubtype = {article} } Crowdsourcing is widely proposed as a method to solve large variety of judgement tasks, such as classifying website content, peer grading in online courses, or collecting real-world data. As the data reported by workers cannot be verified, there is a tendency to report random data without actually solving the task. This can be countered by making the reward for an answer depend on its consistency with answers given by other workers, an approach called peer consistency. However, it is obvious that the best strategy in such schemes is for all workers to report the same answer without solving the task. Dasgupta and Ghosh (WWW 2013) show that in some cases exerting high effort can be encouraged in the highest-paying equilibrium. In this paper we present a general mechanism that implements this idea and is applicable to most crowdsourcing settings. Furthermore, we experimentally test the novel mechanism, and validate its theoretical properties. |
Chatzopoulos, Dimitris; Gujar, Sujit Prakash; Faltings, Boi; Hui, Pan LocalCoin: An ad-hoc payment scheme for areas with high connectivity: poster. Conference ACM MobiHoc 2016, ACM, 2016. @conference{Faltings2016a, title = {LocalCoin: An ad-hoc payment scheme for areas with high connectivity: poster.}, author = {Dimitris Chatzopoulos and Sujit Prakash Gujar and Boi Faltings and Pan Hui}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ p365-chatzopoulos.pdf}, year = {2016}, date = {2016-00-00}, booktitle = {ACM MobiHoc 2016}, publisher = {ACM}, abstract = {The popularity of digital currencies, especially cryptocurrencies, has been continuously growing since the appearance of Bitcoin. Bitcoin is a peer-to-peer (P2P) cryptocurrency protocol enabling transactions between individuals without the need of a trusted authority. Its network is formed from resources contributed by individuals known as miners. Users of Bitcoin currency create transactions that are stored in a specialised data structure called a block chain. Bitcoin's security lies in a proof-of-work scheme, which requires high computational resources at the miners. These miners have to be synchronised with any update in the network, which produces high data traffic rates. Despite advances in mobile technology, no cryptocurrencies have been proposed for mobile devices. This is largely due to the lower processing capabilities of mobile devices when compared with conventional computers and the poorer Internet connectivity to that of the wired networking. In this work, we propose LocalCoin, an alternative cryptocurrency that requires minimal computational resources, produces low data traffic and works with off-the-shelf mobile devices. LocalCoin replaces the computational hardness that is at the root of Bitcoin's security with the social hardness of ensuring that all witnesses to a transaction are colluders. It is based on opportunistic networking rather than relying on infrastructure and incorporates characteristics of mobile networks such as users' locations and their coverage radius in order to employ an alternative proof-of-work scheme. Localcoin features (i) a lightweight proof-of-work scheme and (ii) a distributed block chain.}, keywords = {}, pubstate = {published}, tppubtype = {conference} } The popularity of digital currencies, especially cryptocurrencies, has been continuously growing since the appearance of Bitcoin. Bitcoin is a peer-to-peer (P2P) cryptocurrency protocol enabling transactions between individuals without the need of a trusted authority. Its network is formed from resources contributed by individuals known as miners. Users of Bitcoin currency create transactions that are stored in a specialised data structure called a block chain. Bitcoin's security lies in a proof-of-work scheme, which requires high computational resources at the miners. These miners have to be synchronised with any update in the network, which produces high data traffic rates. Despite advances in mobile technology, no cryptocurrencies have been proposed for mobile devices. This is largely due to the lower processing capabilities of mobile devices when compared with conventional computers and the poorer Internet connectivity to that of the wired networking. In this work, we propose LocalCoin, an alternative cryptocurrency that requires minimal computational resources, produces low data traffic and works with off-the-shelf mobile devices. LocalCoin replaces the computational hardness that is at the root of Bitcoin's security with the social hardness of ensuring that all witnesses to a transaction are colluders. It is based on opportunistic networking rather than relying on infrastructure and incorporates characteristics of mobile networks such as users' locations and their coverage radius in order to employ an alternative proof-of-work scheme. Localcoin features (i) a lightweight proof-of-work scheme and (ii) a distributed block chain. |
Radanovic, Goran; Faltings, Boi Limiting the Influence of Low Quality Information in Community Sensing Conference Proceedings of the 15th AAMAS, 2016. @conference{Radanovic2016a, title = {Limiting the Influence of Low Quality Information in Community Sensing}, author = {Goran Radanovic and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ radanovic2016a.pdf}, year = {2016}, date = {2016-00-00}, booktitle = {Proceedings of the 15th AAMAS}, pages = {873-881}, abstract = {We consider a community of private sensors that collect mea surements of a physical phenomenon, such as air pollution, and report it to a center. The center should be able to prevent low quality reports from degrading the quality of the aggregated information, as there are numerous reasons for operators to inject false sensor data. Hence, it is necessary to track the quality of the sensors over time in order to filter out low quality and malicious reports. To achieve this, we construct a reputation system with a guaranteed bounds on negative impact that malicious sensors can cause, and we evaluate its performance on a realistic dataset.}, keywords = {}, pubstate = {published}, tppubtype = {conference} } We consider a community of private sensors that collect mea surements of a physical phenomenon, such as air pollution, and report it to a center. The center should be able to prevent low quality reports from degrading the quality of the aggregated information, as there are numerous reasons for operators to inject false sensor data. Hence, it is necessary to track the quality of the sensors over time in order to filter out low quality and malicious reports. To achieve this, we construct a reputation system with a guaranteed bounds on negative impact that malicious sensors can cause, and we evaluate its performance on a realistic dataset. |
Radanovic, Goran; Faltings, Boi Learning to Scale Payments in Crowdsourcing with PropeRBoost Conference Proceedings of the Fourth AAAI conference on Conference on Human Computation and Crowdsourcing (HCOMP'16), 2016. @conference{Radanovic2016c, title = {Learning to Scale Payments in Crowdsourcing with PropeRBoost}, author = {Goran Radanovic and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ Radanovic.pdf}, year = {2016}, date = {2016-00-00}, booktitle = {Proceedings of the Fourth AAAI conference on Conference on Human Computation and Crowdsourcing (HCOMP'16)}, abstract = {Motivating workers to provide significant effort has been recognized as an important issue in crowdsourcing. It is important not only to compensate worker effort, but also to discourage low-quality workers from participating. Several proper incentive schemes have been proposed for this purpose; they are either based on gold tasks or on peer consistency in individual tasks. As the rewards cannot become negative, these schemes have difficulty in achieving zero expected reward for random answers. We describe a novel boosting scheme, ProperRBoost, that improves the efficiency of existing incentive schemes by making a better separation between incentives for high and low quality work, and effectively discourages random answers by assigning them near minimal average rewards. We show the actual performance of the boosting scheme through simulations of various worker strategies.}, keywords = {}, pubstate = {published}, tppubtype = {conference} } Motivating workers to provide significant effort has been recognized as an important issue in crowdsourcing. It is important not only to compensate worker effort, but also to discourage low-quality workers from participating. Several proper incentive schemes have been proposed for this purpose; they are either based on gold tasks or on peer consistency in individual tasks. As the rewards cannot become negative, these schemes have difficulty in achieving zero expected reward for random answers. We describe a novel boosting scheme, ProperRBoost, that improves the efficiency of existing incentive schemes by making a better separation between incentives for high and low quality work, and effectively discourages random answers by assigning them near minimal average rewards. We show the actual performance of the boosting scheme through simulations of various worker strategies. |
Faltings, Boi; Radanovic, Goran Mechanismen zur Beschaffung korrekter Daten Journal Article Informatik Spektrum, 2016. @article{Faltings2016b, title = {Mechanismen zur Beschaffung korrekter Daten}, author = {Boi Faltings and Goran Radanovic}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ Faltings15Sept.pdf}, year = {2016}, date = {2016-00-00}, journal = {Informatik Spektrum}, abstract = {,Smart Cities“ brauchen als Grundlage für Entscheidungen grosse Mengen von Daten, die häufig nicht direkt, sondern über andere Agenten erhoben werden. Es ist daher wichtig, die Qualität durch geeignete Anreize oder Kontrolle sicherzustellen. Bei Ausnutzung der Spieltheorie ist dies erstaunlicherweise selbst dann möglich, wenn die Korrektheit der Daten gar nicht kontrolliert werden kann. Wir zeigen neuartige Mechanismen zur Qualitätssicherung für zwei Arten von Agenten. Für rational handelnde Agenten berechnen wir die Belohnung für die Daten so, dass nur korrekte Daten einen positiven Betrag erwarten können. Für fehlerhafte oder absichlich falsche Daten zeigen wir einen neuen Mechanismus, der den negativen Einfluss auf das erzeugte Modell begrenzt. }, keywords = {}, pubstate = {published}, tppubtype = {article} } ,Smart Cities“ brauchen als Grundlage für Entscheidungen grosse Mengen von Daten, die häufig nicht direkt, sondern über andere Agenten erhoben werden. Es ist daher wichtig, die Qualität durch geeignete Anreize oder Kontrolle sicherzustellen. Bei Ausnutzung der Spieltheorie ist dies erstaunlicherweise selbst dann möglich, wenn die Korrektheit der Daten gar nicht kontrolliert werden kann. Wir zeigen neuartige Mechanismen zur Qualitätssicherung für zwei Arten von Agenten. Für rational handelnde Agenten berechnen wir die Belohnung für die Daten so, dass nur korrekte Daten einen positiven Betrag erwarten können. Für fehlerhafte oder absichlich falsche Daten zeigen wir einen neuen Mechanismus, der den negativen Einfluss auf das erzeugte Modell begrenzt. |
Gujar, Sujit Prakash; Faltings, Boi Online Auctions for Dynamic Assignment: Theory and Empirical Evaluation Conference Proceedings of the European Conference on Artificial Intelligence, 22 , EurAI 2016. @conference{gujar2016a, title = {Online Auctions for Dynamic Assignment: Theory and Empirical Evaluation}, author = {Sujit Prakash Gujar and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ ECAI2016GujarFaltings.pdf}, year = {2016}, date = {2016-00-00}, booktitle = {Proceedings of the European Conference on Artificial Intelligence}, volume = {22}, pages = {1035-1043}, organization = {EurAI}, abstract = {Dynamic resource assignment is a common problem in multi-agent systems.We consider scenarios in which dynamic agents have preferences about assignments and the resources that can be assigned using online auctions.We study the trade-off between the following online auction properties: (i) truthfulness, (ii) expressiveness, (iii) efficiency, and (iv) average case performance. We theoretically and empirically compare four different online auctions: (i) Arrival Priority Serial Dictatorship, (ii) Split Dynamic VCG, (iii) e-Action, and (iv) Online Ranked Competition Auction. The latter is a novel design based on the competitive secretary problem. We show that, in addition to truthfulness and algorithmic efficiency, the degree of competition also plays an important role in selecting the best algorithm for a given context.}, keywords = {}, pubstate = {published}, tppubtype = {conference} } Dynamic resource assignment is a common problem in multi-agent systems.We consider scenarios in which dynamic agents have preferences about assignments and the resources that can be assigned using online auctions.We study the trade-off between the following online auction properties: (i) truthfulness, (ii) expressiveness, (iii) efficiency, and (iv) average case performance. We theoretically and empirically compare four different online auctions: (i) Arrival Priority Serial Dictatorship, (ii) Split Dynamic VCG, (iii) e-Action, and (iv) Online Ranked Competition Auction. The latter is a novel design based on the competitive secretary problem. We show that, in addition to truthfulness and algorithmic efficiency, the degree of competition also plays an important role in selecting the best algorithm for a given context. |
2015 |
Radanovic, Goran; Faltings, Boi Incentives for Subjective Evaluations with Private Beliefs Conference Proceedings of the 29th AAAI Conference on Artificial Intelligence (AAAI'15), 2015. @conference{Radanovic&Faltings:15a, title = {Incentives for Subjective Evaluations with Private Beliefs}, author = {Goran Radanovic and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ 9719-44215-1-PB.pdf}, year = {2015}, date = {2015-00-00}, booktitle = {Proceedings of the 29th AAAI Conference on Artificial Intelligence (AAAI'15)}, keywords = {}, pubstate = {published}, tppubtype = {conference} } |
Radanovic, Goran; Faltings, Boi Incentive Schemes for Participatory Sensing Conference Proceedings of the 14th international conference on autonomous agents and multiagent systems (AAMAS'15), 2015. @conference{Radanovic&Faltings:15b, title = {Incentive Schemes for Participatory Sensing}, author = {Goran Radanovic and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ fp428-radanovic.pdf}, year = {2015}, date = {2015-00-00}, booktitle = {Proceedings of the 14th international conference on autonomous agents and multiagent systems (AAMAS'15)}, keywords = {}, pubstate = {published}, tppubtype = {conference} } |
Bhat, Satyanath; Jain, Shweta; Gujar, Sujit; Narahari, Yadati An Optimal Bidimensional Multi-Armed Bandit Auction for Multi-unit Procurement Conference Proceedings of the 2015 International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2015, Istanbul, Turkey, May 4-8, 2015, 2015. @conference{Sujit15a, title = {An Optimal Bidimensional Multi-Armed Bandit Auction for Multi-unit Procurement}, author = {Satyanath Bhat and Shweta Jain and Sujit Gujar and Yadati Narahari}, url = {http://dl.acm.org/citation.cfm?id=2773438 }, year = {2015}, date = {2015-00-00}, booktitle = {Proceedings of the 2015 International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2015, Istanbul, Turkey, May 4-8, 2015}, pages = {1789--1790}, keywords = {}, pubstate = {published}, tppubtype = {conference} } |
Guajr, Sujit; Faltings, Boi Dynamic Task Assignments: An Online Two Sided Matching Approach Conference Proceedings of the 3rd International workshop on Matching Under Preferences, MATCHUP, 2015. @conference{Gujar15a, title = {Dynamic Task Assignments: An Online Two Sided Matching Approach}, author = {Sujit Guajr and Boi Faltings}, url = {https://docs.google.com/viewer?a=v&pid=sites&srcid=ZGVmYXVsdGRvbWFpbnxzdWppdGd1amFyfGd4Ojc4YTc2OGFkMmYxMzRjNWM }, year = {2015}, date = {2015-00-00}, booktitle = {Proceedings of the 3rd International workshop on Matching Under Preferences, MATCHUP}, keywords = {}, pubstate = {published}, tppubtype = {conference} } |
Guajr, Sujit; Faltings, Boi Auction Based Mechanisms for Dynamic Task Assignments in Expert Crowdsourcing Conference Proceedings of the International workshop on Agent Mediated E-Commerce and Trading Agent Design and Analysis (AMEC/TADA'15) , 2015. @conference{Gujar15b, title = {Auction Based Mechanisms for Dynamic Task Assignments in Expert Crowdsourcing }, author = {Sujit Guajr and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ AMEC_auction-based-mechanisms.pdf}, year = {2015}, date = {2015-00-00}, booktitle = {Proceedings of the International workshop on Agent Mediated E-Commerce and Trading Agent Design and Analysis (AMEC/TADA'15) }, keywords = {}, pubstate = {published}, tppubtype = {conference} } |
Mukherjee, Tridib; Dutta, Partha; Hegde, Vinay G; Gujar, Sujit 29th IEEE International Parallel &Distributed Processing Symposium, IPDPS, 2015. @conference{mukherjeerisc, title = {RISC: Robust Infrastructure over Shared Computing Resources Through Dynamic Pricing and Incentivization}, author = {Tridib Mukherjee and Partha Dutta and Vinay G Hegde and Sujit Gujar}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ riscIPDPS2015.pdf}, year = {2015}, date = {2015-00-00}, booktitle = {29th IEEE International Parallel &Distributed Processing Symposium, IPDPS}, keywords = {}, pubstate = {published}, tppubtype = {conference} } |
Musat, Claudiu Cristian; Faltings, Boi Personalizing Product Rankings Using Collaborative Filtering on Opinion-Derived Topic Profiles Journal Article 24th IJCAI, pp. 830-836, 2015. @article{MusatFaltings2015b, title = {Personalizing Product Rankings Using Collaborative Filtering on Opinion-Derived Topic Profiles}, author = {Claudiu Cristian Musat and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ ijcai15.pdf}, year = {2015}, date = {2015-00-00}, journal = {24th IJCAI}, pages = {830-836}, abstract = {Product review sites such as TripAdvisor, Yelp or Amazon provide a single, non personalized ranking of products. The sparse review data makes personalizing recommendations difficult. Topic Profile Collaborative Filtering exploits review texts to identify user profiles as a basis for similarity. We show that careful use of the available data and separating users into classes can greatly improve the performance of such techniques. We significantly improve MAE, RMSE, and Kendall tau, compared to the previous best results. In addition, we show that personalization does not benefit all the users to the same extent. We propose switching between a personalized and a non personalized method based on the user opinion profile. We show that the user’s opinionatedness is a good indicator of whether the personalization will work or not.}, keywords = {}, pubstate = {published}, tppubtype = {article} } Product review sites such as TripAdvisor, Yelp or Amazon provide a single, non personalized ranking of products. The sparse review data makes personalizing recommendations difficult. Topic Profile Collaborative Filtering exploits review texts to identify user profiles as a basis for similarity. We show that careful use of the available data and separating users into classes can greatly improve the performance of such techniques. We significantly improve MAE, RMSE, and Kendall tau, compared to the previous best results. In addition, we show that personalization does not benefit all the users to the same extent. We propose switching between a personalized and a non personalized method based on the user opinion profile. We show that the user’s opinionatedness is a good indicator of whether the personalization will work or not. |
Maksai, Andrii; Garcin, Florent; Faltings, Boi Predicting Online Performance of News Recommender Systems Through Richer Evaluation Metrics Journal Article ACM Conference on Recommender Systems (Recsys), (9), pp. 179-186, 2015. @article{Maksai2015, title = {Predicting Online Performance of News Recommender Systems Through Richer Evaluation Metrics}, author = {Andrii Maksai and Florent Garcin and Boi Faltings}, url = {https://lia.epfl.ch/wp-content/uploads/publications/ p179.pdf}, year = {2015}, date = {2015-00-00}, journal = {ACM Conference on Recommender Systems (Recsys)}, number = {9}, pages = {179-186}, abstract = {We investigate how metrics that can be measured offline can be used to predict the online performance of recommender systems, thus avoiding costly A-B testing. In addition to accuracy metrics, we combine diversity, coverage, and serendipity metrics to create a new performance model. Using the model, we quantify the tradeoff between different metrics and propose to use it to tune the parameters of recommender algorithms without the need for online testing. Another application for the model is a self-adjusting algorithm blend that optimizes a recommender's parameters over time. We evaluate our findings on data and experiments from news websites.}, keywords = {}, pubstate = {published}, tppubtype = {article} } We investigate how metrics that can be measured offline can be used to predict the online performance of recommender systems, thus avoiding costly A-B testing. In addition to accuracy metrics, we combine diversity, coverage, and serendipity metrics to create a new performance model. Using the model, we quantify the tradeoff between different metrics and propose to use it to tune the parameters of recommender algorithms without the need for online testing. Another application for the model is a self-adjusting algorithm blend that optimizes a recommender's parameters over time. We evaluate our findings on data and experiments from news websites. |