News

Multi-Objective Decision Making Workshop - 2023

In recent years there has been a growing awareness of the need for automated and assistive decision making systems to move beyond single-objective formulations when dealing with complex real-world issues, which invariably involve multiple competing objectives. The purpose of this workshop is to promote collaboration and cross-fertilisation of ideas between researchers working in different areas of multi-objective decision making in the context of intelligent systems, and to provide a forum for dissemination of high-quality multi-objective decision making research.

Previous editions of this workshop may be found at the following urls:

The workshop targets high-quality original papers covering all aspects of multi-objective decision making, including, but not limited to, the list of topics below. The following is a non-exhaustive list of topics that we would like to cover in the workshop:

  • Multi-objective/multi-criteria/multi-attribute decision making
  • Multi-objective reinforcement learning
  • Multi-objective planning and scheduling
  • Multi-objective multi-agent decision making
  • Multi-objective game theory
  • Preference elicitation for MODeM
  • Social choice and MODeM
  • Multi-objective decision support systems
  • Multi-objective metaheuristic optimisation (e.g. evolutionary algorithms) for autonomous agents and multi-agent systems
  • Multi-objectivisation
  • Ethical AI through multi-objective modelling
  • Explainable aI through multi-objective modelling
  • Interactive systems for MODeM
  • Applications of MODeM
  • New benchmark problems for MODeM

Extended and revised versions of papers presented at the workshop will be eligible for inclusion in a journal topical collection.

Program Details

The workshop programme will consist of contributed original talks, invited talks, and a panel discussion. We might also ask participants to pre-record talks and make them available outside of the workshop sessions, on a dedicated Youtube channel.

Important Dates

  • Submission Deadline: 30 June 2023   extended to 21 July 2023 (23:59 AoE)
  • Notification of acceptance: 15 August 2023
  • Camera-ready copies: 31 August 2023
  • Workshop: 1 October 2023

Submission Details

Papers should be formatted according to the ECAI 2023 guidelines, and should be a maximum of 7 pages in length (with additional pages containing references only). Additionally, we welcome submission of preliminary results, i.e. work-in-progress, as well as visionary outlook papers that lay out directions for future research in a specific area, both up to 5 pages in length, although shorter papers are very much welcome, and will not be judged differently. Finally, we also accept recently published journal papers in the form of a 2 page abstract.

All submissions will be peer-reviewed (single-blind). Accepted work will be allocated time for poster and/or oral presentation during the workshop. Extended versions of original papers presented at the workshop will also be eligible for inclusion in a post-proceedings journal topical collection.

Papers can be submitted through Microsoft CMT.

Journal Topical Collection

After the workshop, all original contributions presented at MODeM 2023 will be invited to submit substantially improved and extended versions of their work, for consideration to be published in a post-proceedings topical collection of the Springer journal Neural Computing & Applications (NC&A, impact factor 5.102): Topical Collection on Multi-Objective Decision Making 2023.

NCA

Submission of preliminary work to the MODeM 2023 workshop does not confer an automatic entitlement to publish in the NC&A Topical Collection on Multi-Objective Decision Making 2023; all topical collection submissions must pass through the same rigorous NC&A review process and meet the standard NC&A publication criteria.

The Topical Collection on Multi-Objective Decision Making 2023 has an open call for papers; it is not necessary to submit preliminary work to the MODeM 2023 workshop in order to have your manuscript considered for publication in this topical collection. The topical collection is already open for submissions, and will remain open until 31 January 2024. Papers submitted to the topical collection will be sent out for peer review as soon as they are received, and first decisions can be expected within 3 months approx. from the date of submission.

Program

Sunday, 1 October

09:00 - 09:15 Welcome & Opening Remarks
09:15 - 10:30 Session I - Chair: Roxana Rădulescu
09:15 - 10:15 Invited talk
Nawal Benabbou: Combining Preference Elicitation and Search for Solving Multi-objective Optimization Problems
10:15 - 10:30 Alexandra Cimpean, Catholijn M Jonker, Pieter Libin, Ann Nowé
A Multi-objective Framework For Fair Reinforcement Learning
10:30 - 11:00 Coffee break
11:00 - 12:30 Session II - Chair: Izabela Szczech
11:00 - 11:15 William Blanzeisky, Pádraig Cunningham
Beyond Accuracy: Addressing Underestimation Bias in Multi-Label Image Classification through Multi-Objective Optimization
11:15 - 11:30 Timothy J Parker
Anticipating Responsibility for Plan Selection
11:30 - 11:45 Maha Riad, Vinicius Renan de Carvalho, Fatemeh Golpayegani
Multi-Value Alignment in Normative Multi-Agent System: An Evolutionary Optimisation Approach
11:45 - 12:00 Junlin Lu, Patrick Mannion, Karl Mason
Inferring Preferences from Demonstrations in Multi-Objective Residential Energy Management
12:00 - 12:15 Patrick Mannion, Roxana Rădulescu
Comparing utility-based and Pareto-based solution sets in multi-objective normal form games
12:15 - 12:30 José-Luis Pérez-de-la-Cruz, Lawrence Mandow, Sergio Martín-Albo Martín
Multi-objective bandit algorithms with Chebyshev scalarization
12:30 - 13:30 Lunch break
13:30 - 15:00 Session III - Chair: Timothy Parker
13:30 - 13:45 Arundhati Banerjee, Ramina Ghods, Jeff Schneider
Cost-awareness in Multi-Agent Active Search
13:45 - 14:00 Vedat Dogan, Steven D Prestwich
Multi-objective Bilevel Decision Making with Noisy Objectives: A Batch Bayesian Approach
14:00 - 14:15 Robert Susmaga, Izabela Szczech
Visualizing the Inner-Workings of TOPSIS
14:15 - 14:30 Robert Susmaga, Izabela Szczech
Elliptic Generalizations of TOPSIS
14:30 - 14:45 Jesse van Remmerden, Maurice Kenter, Diederik Roijers, Yingqian Zhang, Charalampos Andriotis, Zaharah Bukhsh
A Deep Multi-Objective Reinforcement Learning Approach for Infrastructural Maintenance Planning with Non-Linear Utility Functions
14:45 - 15:00 Andreas H Hamel, Daniel Kostner
MCDM via Cone Ranking Functions
15:00 - 15:30 Coffee break
15:30 - 17:00 Session IV - Chair: Patrick Mannion
15:30 - 15:45 Florian Felten, Daniel Gareev, El-Ghazali Talbi, Gregoire Danoy
Hyperparameter Optimization for Multi-Objective Reinforcement Learning
15:45 - 16:00 Sofyan Ajridi, Willem Röpke, Ann Nowé, Roxana Rădulescu
Deconstructing Reinforcement Learning Benchmarks: Revealing The Objectives
16:00 - 17:00 Invited talk
Nic Wilson: Optimising with Respect to Partially Known Multi-Objective User Preferences
17:00 - 17:10 Closing remarks

Accepted Papers

Paper # Details Title Authors
2[Paper]Multi-objective Bilevel Decision Making with Noisy Objectives: A Batch Bayesian ApproachVedat Dogan (University College Cork); Steven D Prestwich (University College Cork)
3[Paper]Visualizing the Inner-Workings of TOPSISRobert Susmaga (PUT); Izabela Szczech (Poznan University of Technology)
5[Paper]Elliptic Generalizations of TOPSISRobert Susmaga (PUT); Izabela Szczech (Poznan University of Technology)
6[Paper]Beyond Accuracy: Addressing Underestimation Bias in Multi-Label Image Classification through Multi-Objective OptimizationWilliam Blanzeisky (University College Dublin); Pádraig Cunningham (UCD School of Computer Science)
7[Paper][Video]Hyperparameter Optimization for Multi-Objective Reinforcement LearningFlorian Felten (SnT, University of Luxembourg); Daniel Gareev (FSTM/DCS, University of Luxembourg); El-Ghazali Talbi (CNRS/CRIStAL, University of Lille); Gregoire Danoy (FSTM/DCS and SnT, University of Luxembourg)
8[Paper]Inferring Preferences from Demonstrations in Multi-Objective Residential Energy ManagementJunlin Lu (University of Galway); Patrick Mannion (University of Galway); Karl Mason (University of Galway)
9[Paper][Video]Cost-awareness in Multi-Agent Active SearchArundhati Banerjee (Carnegie Mellon University); Ramina Ghods (Carnegie Mellon University); Jeff Schneider (CMU)
11[Paper]Deconstructing Reinforcement Learning Benchmarks: Revealing The ObjectivesSofyan Ajridi (Vrije Universiteit Brussel); Willem Röpke (Vrije Universiteit Brussel); Ann Nowé (Vrije Universiteit Brussel); Roxana Rădulescu (Vrije Universiteit Brussel)
12[Paper]Comparing utility-based and Pareto-based solution sets in multi-objective normal form gamesPatrick Mannion* (University of Galway); Roxana Rădulescu* (Vrije Universiteit Brussel)
13[Paper]MCDM via Cone Ranking FunctionsAndreas H Hamel (Free University of Bozen-Bolzano); Daniel Kostner (Free University of Bozen-Bolzano)
14[Paper]Multi-objective bandit algorithms with Chebyshev scalarizationJosé-Luis Pérez-de-la-Cruz (Universidad de Málaga); Lawrence Mandow (Universidad de Málaga); Sergio Martín-Albo Martín (Universidad de Málaga)
15[Paper]A Deep Multi-Objective Reinforcement Learning Approach for Infrastructural Maintenance Planning with Non-Linear Utility FunctionsJesse van Remmerden (Eindhoven University of Technology); Maurice Kenter (City of Amsterdam); Diederik Roijers (City of Amsterdam); Yingqian Zhang (Eindhoven University of Technology); Charalampos Andriotis (Delft University of Technology); Zaharah Bukhsh (Eindhoven University of Technology)
17[Paper]Anticipating Responsibility for Plan SelectionTimothy J Parker (IRIT)
18[Paper][Video]Multi-Value Alignment in Normative Multi-Agent System: An Evolutionary Optimisation ApproachMaha Riad (University College Dublin); Vinicius Renan de Carvalho (USP); Fatemeh Golpayegani (University College Dublin)
19[Paper]A Multi-objective Framework For Fair Reinforcement LearningAlexandra Cimpean (Vrije Universiteit Brussel); Catholijn M Jonker (Delft University of Technology); Pieter Libin (Vrije Universiteit Brussel); Ann Nowé (Vrije Universiteit Brussel)

Invited Talks

Nawal Benabbou

Affiliation: Sorbonne Université (Paris, France)

Website: https://benabbou.lip6.fr

Talk Title: Combining Preference Elicitation and Search for Solving Multi-objective Optimization Problems

Abstract: The increasing complexity of applications encountered in multi-objective optimization leads us today to apply decision models to combinatorial sets of solutions which are implicitly defined. This significantly complicates the decision process and, in particular, the construction of a preference model fitting the objectives of the decision maker as well as the calculation of the optimal decision. Instead of approaching independently and separately these two subjects, we introduce incremental decision procedures aiming to integrate and combine the elicitation of preferences and the calculation of the preferred solution in order to determine the optimal choice without fully specifying the decision model. In these new interactive resolution schemes, asking preference questions to the decision maker during the exploration of the set of solutions allows to focus the elicitation of preferences on information that is really useful for separating competing solutions and thus reducing the number of questions required. This is the main benefit of the incremental approach to multi-objective decision making.

Bio: Dr. Nawal Benabbou is an assistant professor at Sorbonne Université (Paris), in the DECISION team of the LIP6 laboratory of computer science. Her research interests lie in the area of Algorithmic Decision Theory and its applications in Artificial Intelligence and Operations Research. She is the recipient of the “Artificial Intelligence Dissertation Award”, awarded by the French AI association (AFIA) in 2018, and is one of the vice-presidents of the French OR and Decision Making association (ROADEF). She is also actively reviewing for several top-ranked international AI conferences (e.g., AAAI, IJCAI, AAMAS, ECAI) and international journals (e.g., EJOR, AGNT, ANOR, 4OR).

Nic Wilson

Affiliation: University College Cork (Cork, Ireland)

Website: http://publish.ucc.ie/researchprofiles/D005/nwilson

Talk Title: Optimising with Respect to Partially Known Multi-Objective User Preferences

Abstract: Commonly, a decision support system will have only partial information about the preferences of a user. In multi-objective optimisation problems, user preferences are often modelled with some form of weighted sum of objective values, where the vector of weights is initially unknown. In this talk I will consider reasoning and optimisation with partial knowledge of preferences for such multi-objective optimisation problems. If an alternative is necessarily optimal, i.e., optimal for each consistent candidate preference function, then the system can safely recommend that alternative (even though the knowledge of the user preferences may still only be partial). Otherwise, there are different potentially optimal alternatives. One can increase the knowledge of the user preferences by querying the user, asking the user to compare two or more alternatives. I will discuss different querying strategies, which aim towards finding a necessarily optimal alternative.

Bio: Nic Wilson is a senior research fellow at the Insight Centre for Data Analytics, based in the School of Computer Science and Information Technology, University College Cork. His main research areas are in reasoning and optimisation with preferences and uncertainty, especially on combinatorial problems, and has published around a hundred peer-reviewed papers. Application areas of his work include scheduling, supply chain optimisation, transportation logistics, telecommunications, recommender systems and intelligent buildings.

Program Committee

  • Athirai Irissappane, IBM, US
  • Ben Abramowitz, Tulane University, US
  • Brandon T. Fain, Duke University, US
  • Debora Di Caprio, University of Trento, IT
  • Dimitris Michailidis, University of Amsterdam, NL
  • Florian Felten, SnT, University of Luxembourg, LU
  • Francisco Javier Santos Arteaga, Universidad Complutense de Madrid, ES
  • Fredrik Heintz, Linköping University, SE
  • Gabriel Ramos, Unisinos, BR
  • Johan Källström, Linköping University, SE
  • Junlin Lu, University of Galway, IE
  • Lucas Nunes Alegre, Federal University of Rio Grande do Sul (UFRGS), BR
  • Manel Rodriguez-Soto, IIIA-CSIC, ES
  • Marc Vincent, Thales Land and Air Systems / Sorbonne Université, FR
  • Mathieu Reymond, Vrije Universiteit Brussel, BE
  • Pieter Libin, Vrije Universiteit Brussel, BE
  • Takuya Kanazawa, Hitachi, Ltd., US
  • Thommen Karimpanal George, Deakin University, AU
  • Vincent Corruble, Sorbonne University, FR
  • Willem Röpke, Vrije Universiteit Brussel, BE
  • Yijie Zhang, University of Copenhagen, DK
  • Zhaori Guo, University of Southampton, UK

Organization

This year's workshop is organised by: Senior Advisory Committee:

Contact

If you have any questions about the MODeM workshop, please contact the organizers at:
modem.workshop.2023 AT gmail.com