Our paper ''PAC Learning and Stabilizing Hedonic Games: Towards a Unifying Approach'' (by Simone Fioravanti, Michele Flammini, Bojana Kodric, Giovanna Varricchio) was accepted at the 37th AAAI Conference on Artificial Intelligence (AAAI 2023), one of the two international top conferences in artificial intelligence.
Our former secretary Jutta Nadland retired on June 30. We thank her for the many contributions to our team and the excellent service, and we wish her all the best!
At the same time, we welcome Marta Soares as the new secretary in our group.
Several (former) members of our group received awards for oustanding performances among the 2021/22 graduates in Computer Science at Goethe University:
New Paper at IJCAI 2022 20 Apr 2022
Our paper ''Approximate Strategyproof Mechanisms for the Additively
Separable Group Activity Selection Problem'' (by Michele Flammini, Giovanna Varricchio) was accepted at the 31st Int. Joint Conference on Artificial Intelligence (IJCAI 2022), one of the two international top conferences in artificial intelligence.
Visitor 5 Apr 2022
Kevin Schewior is visiting our group Apr 19-21.
Welcome 1 Apr 2022
We welcome Conrad Schecker as a new member of our group.
In our paper ''The Secretary Recommendation Problem'' (by Niklas Hahn, Martin Hoefer, Rann Smorodinsky) we study a basic sequential recommendation problem. A sender (e.g., a social network) decides to recommend taking an action (e.g., clicking on certain media content) to a receiver (user in the network). Actions can have different utilities for sender and receiver, which are known upfront to the sender (but not to the receiver). As such, the receiver must rely on the sender to give informative advice. We analyze the scenario when actions become (un-)available sequentially over time. Our results are near-optimal algorithms to compute good recommendation strategies for the sender. On a technical level, we provide a novel connection between classic stopping theory and persuasion/signaling aspects. The paper has been accepted for publication in Games and Economic Behavior, the top journal in game theory.
New Paper at AAMAS 2022 19 Dec 2021
Our paper ''Asynchronous Opinion Dynamics in Social Networks'' (by Petra Berenbrink, Martin Hoefer, Dominik Kaaser, Pascal Lenzner, Malin Rau, Daniel Schmand) was accepted as full paper at the 21st International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2022), the international top conference on multiagent systems.
New Paper at AAAI 2022 01 Dec 2021
Our paper ''Maximizing Nash Social Welfare in 2-Value Instances'' (by Hannaneh Akrami, Bhaskar Ray Chaudhury, Martin Hoefer, Kurt Mehlhorn, Marco Schmalhofer, Golnoosh Shahkarami, Giovanna Varricchio, Quentin Vermande, Ernest van Wijland) was accepted at the 36th AAAI Conference on Artificial Intelligence (AAAI 2022), one of the two international top conferences in artificial intelligence.
In ''Fair Division of Indivisible Goods for a Class of Concave Valuations'' (by Bhaskar Ray Chaudhury, Yun Kuen Cheung, Jugal Garg, Naveen Garg, Martin Hoefer, Kurt Mehlhorn) we study approximation algorithms for fair allocation of a set of indivisible goods to a set of agents. Our algorithm computes a near-optimal allocation in terms of the geometric mean (also termed Nash social welfare) of agent valuations. In addition, the allocation satisfies approximate versions of envy-freeness, a standard fairness concept in the literature. Our results unify and extend existing techniques for approximating Nash social welfare for additive valuations to a more general, concave and non-separable class of valuation functions. The paper has been accepted for publication in Journal of Artificial Intelligence Research, a top journal in artificial intelligence.
Graduation Award 22 Oct 2021
Lisa Wilhelmi received the award for the best M.Sc. graduate of 2020/21 in Computer Science at Goethe University. Congratulations, well deserved!
In our paper ''Algorithms for Persuasion with Limited Communication'' (by Ronen Gradwohl, Niklas Hahn, Martin Hoefer, Rann Smorodinsky) we study the problem faced by a recommendation engine (e.g., a website for hotel booking or travel) that wants to persuade a user with a recommendation. We analyze restrictions on the signal space such that there are more potential actions by the user than signals that can be issued for recommendation. For symmetric instances, when a priori all actions have similar profits, we characterize and design efficient algorithms for optimal recommendation schemes for the engine. In a subclass of instances, where profits are independent among actions, we obtain an algorithm that computes a near-optimal scheme. The paper has been accepted for publication in Mathematics of Operations Research, a leading journal in operations research.
Tutorial on Algorithms and Information Design at SAGT 2021 27 Aug 2021
Stay tuned for our tutorial on Algorithmic Challenges in Information Design; on 21 Sep 2021 at SAGT in Aarhus (and online).
New Paper at SAGT 2021 29 Jun 2021
Our paper ''When Dividing Mixed Manna is Easier than Dividing Goods: Competitive Equilibria with a Constant Number of Chores'' (by Jugal Garg, Martin Hoefer, Peter McGlaughlin, Marco Schmalhofer) was accepted at the 14th Int. Symposium on Algorithmic Game Theory (SAGT 2021).
New Paper at IJCAI 2021 29 Apr 2021
Our paper ''Stochastic Probing with Increasing Precision'' (by Martin Hoefer, Kevin Schewior, Daniel Schmand) was accepted at the 30th Int. Joint Conference on Artificial Intelligence (IJCAI 2021), one of the two international top conferences in artificial intelligence.
In our paper ''Reaping the Informational Surplus in Bayesian Persuasion'' (by Ronen Gradwohl, Niklas Hahn, Martin Hoefer, Rann Smorodinsky) we study a competitive scenario, in which several experts with knowledge about the state of the world want to influence the choice of a decision maker (e.g., consultants/lobbyists try to influence policy decisions of the government). Even if there is only slight uncertainty for each expert about the interests of other experts, then the decision maker can extract the full informational surplus -- by restricting to listen only to one expert, the emerging competition motivates them to reveal all information about the state of the world. The paper has been accepted for publication in American Economic Journal: Microeconomics, a top journal in economics.
Daniel Schmand moves to University of Bremen 05 Feb 2021
Our group member Daniel Schmand accepted an offer to become assistant professor in discrete optimization at University of Bremen. Congratulations Prof. Schmand!
Welcome 1 Feb 2021
We welcome Tim Koglin as a new member of our group.
WINE 2020 Live Stream 26 Nov 2020
The 16th Conf. Web and Internet Economics (WINE 2020), one of the main events in economics and computation, is happening virtually Dec 7-11, 2020. There will be a free live stream on YouTube with all invited and contributed talks. All talks are scheduled within a (Europe-friendly) time window from 8am to 5pm CET.
New Paper at ITCS 2021 2 Nov 2020
Our paper ''Algorithmic Persuasion with Evidence'' (by Martin Hoefer, Pasin Manurangsi, Alexandros Psomas) was accepted at the 12th Int. Conf. Innovations in Theoretical Computer Science (ITCS 2021), a top conference in theory of computing.
Welcome 1 Nov 2020
We welcome Giovanna Varricchio as a new member of our group.
Welcome 1 Oct 2020
We welcome Lisa Wilhelmi as a new member of our group.
New Paper at SODA 2021 30 Sep 2020
Our paper ''Algorithms for Persuasion with Limited Communication'' (by Ronen Gradwohl, Niklas Hahn, Martin Hoefer, Rann Smorodinsky) was accepted at the 31st Symposium on Discrete Algorithms (SODA 2021), the international top conference in design and analysis of algorithms.
New Paper in Networks 29 Sep 2020
In our paper ''Packing Returning Secretaries'' (by Martin Hoefer, Lisa Wilhelmi) we analyze extensions of the classic secretary problem and relax the assumption that the decision for a specific element must be made instantaneously. Our results are approximation guarantees for a wide class of combinatorial packing problems, such as knapsack, matching or independent set. Moreover, we are interested in scenarios where the decision for an element can be postponed, with the goal of selecting an optimal set of elements with the minimum number of postponements. The paper has been accepted for publication in Networks.
New Paper at COLT 2020 26 May 2020
In the new paper ''A Nearly Optimal Variant of the Perceptron Algorithm for the Uniform Distribution on the Unit Sphere'', Marco Schmalhofer shows that a simple variant of the classic perceptron algorithm provides near-optimal guarantees for PAC-learning of halfspaces, where examples are drawn uniformly at random on the hypersphere. The paper was accepted at the 33rd Annual Conference on Learning Theory (COLT 2020), the international top conference in learning theory.
Two Papers at EC 2020 6 May 2020
Two papers from our group were accepted at the 21st ACM Conference on Economics and Computation (EC 2020), the international top conference at the intersection of computer science and economics:
DFG Research Group 29 Apr 2020
The German Research Foundation grants funding for a new research group on Algorithms, Dynamics, and Information Flow in Networks. The group is headed by Martin Hoefer, the participating PIs are Petra Berenbrink (Hamburg), Nils Bertschinger, Amin Coja Oghlan, Ulrich Meyer (Frankfurt) and Tobias Friedrich (HPI/Potsdam). The goals are to advance our understanding of dynamic processes on networks and their relation to efficient algorithms, e.g., in the analysis of spreading processes, distributed network algorithms, network generation models, as well as application domains such as financial markets.
New Paper at IJCAI 2020 21 Apr 2020
Our paper ''Prophet Inequalities for Bayesian Persuasion'' (by Niklas Hahn, Martin Hoefer, Rann Smorodinsky) was accepted at the 29th Int. Joint Conference on Artificial Intelligence (IJCAI 2020), one of the two international top conferences in artificial intelligence.
Visitor 1 Mar 2020
Kevin Schewior is visiting our group Mar 9-13.
DFG Grant 14 Feb 2020
In a our new three-year project Algorithms for Fair Allocation of Indivisible Goods funded by the German Research Foundation, we study algorithmic problems for fair assignment of indivisble goods, which arise, e.g., when assigning students to schools, splitting an inheritance, sharing rooms in an appartment, etc. Jugal Garg from the University of Illinois at Urbana-Champaign collaborates with us as a Mercator Fellow.
Welcome 16 Jan 2020
We welcome Marco Schmalhofer as a new member of our group.
In our paper ''Efficient Black-Box Reductions for Separable Cost Sharing'' (by Tobias Harks, Martin Hoefer, Anja Schedel, Manuel Surek) we study cost sharing protocols for resource allocation with selfish agents. In scenarios related to facility location and Steiner network design we show how approximation algorithms for optimization can be used in a black-box fashion to construct protocols with good stable states. The paper has been accepted for publication in Mathematics of Operations Research, a leading journal in operations research.
Visitor 3 Dec 2019
Rann Smorodinsky is visiting our group Dec 17-20.
Visitor 15 Nov 2019
Kevin Schewior is visiting our group Nov 21/22.
Visitor 1 Nov 2019
Andreas Abels is visiting our group Nov 12-14.
New Paper at ITCS 2020 31 Oct 2019
Our paper ''Strategic Payments in Financial Networks'' (by Nils Bertschinger, Martin Hoefer, Daniel Schmand) was accepted at the 11th Int. Conf. Innovations in Theoretical Computer Science (ITCS 2020), a top conference in theory of computing.
Welcome 1 Sep 2019
We welcome Steffen Schuldenzucker as a new member of our group.
New Paper at SAGT 2019 1 Jul 2019
Our paper ''The Online Best Reply Algorithm for Resource Allocation Problems'' (by Max Klimm, Daniel Schmand, Andreas Tönnis) was accepted at the 12th Int. Symposium on Algorithmic Game Theory (SAGT 2019).
In our paper ''Earning and Utility Limits in Fisher Markets'' (by Xiaohui Bei, Jugal Garg, Martin Hoefer, Kurt Mehlhorn) we study a generalization of the classic Fisher market model in which participants have limits for their utility. We provide algorihtms for computing equilibria, which have a number of desirable fairness and efficiency guarantees. The paper has been accepted for publication in ACM Transactions on Economics and Computation.
Visitor 6 May 2019
Thomas Kesselheim is visiting our group May 6-7.
New Paper at ICALP 2019 25 Apr 2019
Our paper ''Network Investment Games with Wardrop Followers'' (by Daniel Schmand, Marc Schröder, Alexander Skopalik) was accepted at the 46th Int. Colloquium on Automata, Languages, and Programming (ICALP 2019), the top conference in theory of computing in Europe.
In our paper ''Ascending-Price Algorithms for Unknown Markets'' (by Xiaohui Bei, Jugal Garg, Martin Hoefer) we show how simple price adaptation algorithms can rapidly lead an exchange economy to equilibrium, even if many details such as size or structure of the market are unknown in the beginning. The paper has been accepted for publication in ACM Transactions on Algorithms, a top journal in the field of algorithms.
Upon retirement of an employee, a common practice in many areas is to hire a new person early with a period of overlap. This strategy can have significant advantages over strictly sequential employment, as shown by Daniel Schmand and his colleagues. The paper ''Hiring Secretaries over Time: The Benefit of Concurrent Employment'' (by Yann Disser, John Fearnley, Martin Gairing, Oliver Göbel, Max Klimm, Daniel Schmand, Alexander Skopalik, Andreas Tönnis) has been accepted for publication in Mathematics of Operations Research, a leading journal in operations research.
Visitor 2 Feb 2019
Rann Smorodinsky is visiting our group Feb 25-28.
Visitor 23 Jan 2019
Warut Suksompong is visiting our group Feb 18/19.
New Paper at AAMAS 2019 23 Jan 2019
Our paper ''Tracing Equilibrium in Dynamic Markets via Distributed Adaptation'' (by Yun Kuen Cheung, Martin Hoefer, Paresh Nakhe) has been accepted as a full paper at the 18th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2019), the international top conference on multiagent systems.
Das Wissenschaftsmagazin widmet sich in der aktuellen Ausgabe dem Thema "Mit Unsicherheit rechnen" und unserer Arbeit zu Online-Algorithmen.
As one of the best papers from SAGT 2017, our paper ''Opinion Formation Games with Aggregation and Negative Influence'' (by Markos Epitropou, Dimitris Fotakis, Martin Hoefer, Stratis Skoulakis) has been accepted for publication in the special issue of Theory of Computing Systems.
Welcome 01 Oct 2018
We welcome Daniel Schmand as a new member of our group.
New Paper at FSTTCS 2018 24 Sep 2018
Our paper ''On Fair Division for Indivisible Items'' (by Bhaskar Ray Chaudhuri, Yun Kuen Cheung, Jugal Garg, Naveen Garg, Martin Hoefer, Kurt Mehlhorn) has been accepted at the 38th IARCS Annual Conference on Foundations of Software Technology and Theoretical Computer Science (FSTTCS 2018).
New Paper at ISAAC 2018 31 Aug 2018
Our paper ''Packing Returning Secretaries'' (by Martin Hoefer, Lisa Wilhelmi) has been accepted at the 29th International Symposium on Algorithms and Computation (ISAAC 2018), the top conference in theory of computing in Asia.
Thesis Defense 13 Aug 2018
Paresh Nakhe successfully defended his thesis entitled On Bandit Learning and Pricing in Markets.
Congratulations to Dr. Nakhe!
Thesis Defense 23 Jul 2018
Bojana Kodric successfully defended her thesis entitled Incentives in Dynamic Markets at Saarland University.
Congratulations to Dr. Kodric!
Our paper ''Dynamics in Matching and Coalition Formation Games with Structural Constraints'' (by Martin Hoefer, Daniel Vaz, Lisa Wagner) has been accepted for publication in Artifical Intelligence, the top journal in the field.
Visitor 14 May 2018
Max Klimm is visiting our group May 14-17.
New Paper at ICALP 2018 16 Apr 2018
Our paper ''Efficient Black-Box Reductions for Separable Cost Sharing'' (by Tobias Harks, Martin Hoefer, Anja Huber, Manuel Surek) was accepted at the 45th Int. Colloquium on Automata, Languages, and Programming (ICALP 2018), the top conference in theory of computing in Europe.
Visitor 12 Mar 2018
Rann Smorodinsky is visiting our group March 13-17.
GIF Grant 1 Jan 2018
In a new project funded by the German-Israeli Foundation, we study algorithmic problems with sequential information aggregation, such as recommendation systems and the wisdom of the crowd (e.g., in platforms like TripAdvisor or Booking.com) or sequential matching problems (e.g., in Internet marketing or organ donation). The project runs for three years and is a collaboration with Rann Smorodinsky from the Technion.
Visitors 4 Dec 2017
Tobias Harks is visiting our group December 4-5, and Yun Kuen Cheung is visiting December 5-8.
New Paper at SODA 2018 29 Sep 2017
Our paper ''Approximating the Nash Social Welfare with Budget-Additive Valuations'' (by Jugal, Garg, Martin Hoefer, Kurt Mehlhorn) was accepted at the 29th Symposium on Discrete Algorithms (SODA 2018), the international top conference in design and analysis of algorithms.
New Paper at WINE 2017 26 Sep 2017
In his paper ''Dynamic Pricing in Competitive Markets'', Paresh Nakhe highlights how learning algorithms for revenue maximization can lead markets quickly into equilibrium states. The paper has been accepted at the 13th Conference on Web and Internet Economics (WINE 2017) in Bangalore, India.
Visitor 14 Jul 2017
Xiaohui Bei is visiting our group July 15-30.
Three Papers at SAGT 2017 22 Jun 2017
Three papers of our group were accepted at the 10th Int. Symposium on Algorithmic Game Theory (SAGT 2017):
- Earning Limits in Fisher Markets with Spending-Constraint Utilities
(by Xiaohui Bei, Jugal Garg, Martin Hoefer, Kurt Mehlhorn)
- Opinion Formation Games with Aggregation and Negative Influence
(by Markos Epitropou, Dimitris Fotakis, Martin Hoefer, Stratis Skoulakis)
- On Proportional Allocation in Hedonic Games
(by Martin Hoefer, Wanchote Jiamjitrak)
New Paper at ICALP 2017 18 Apr 2017
Our paper ''Combinatorial Secretary Problems with Ordinal Information'' (by Martin Hoefer, Bojana Kodric) was accepted at the 44th Int. Colloquium on Automata, Languages, and Programming (ICALP 2017), the top conference in theory of computing in Europe. This year's edition is going to take place in Warsaw, Poland, in July 2017.
Website Launch 28 Mar 2017
Our new website is online.
New Paper at WINE 2016 16 Dec 2016
Our paper ''Smoothness for Simultaneous Composition of
Mechanisms with Admission'' (by Martin Hoefer, Bojana Kodric,
Thomas Kesselheim) was presented and published at the 12th Conference on Web and Internet Economics (WINE 2016) in Montreal, Canada.
Invited Talk at SAGT 2017 23 Nov 2016
Martin Hoefer will give an invited talk at the 10th Int. Symposium on Algorithmic Game Theory (SAGT 2017) in L'Aquila, Italy, on September 12-14, 2017.
