The theory of belief functions, also referred to as evidence theory or Dempster-Shafer theory, was first introduced by Arthur P. Dempster in the context of statistical inference, and was later developed by Glenn Shafer as a general framework for modelling epistemic uncertainty. These early contributions have been the starting points of many important developments, including the transferable belief model and the theory of hints. The theory of belief functions is now well established as a general framework for reasoning with uncertainty, and has well understood connections to other frameworks such as probability, possibility and imprecise probability theories. It has been applied in diverse areas such as machine learning, information fusion and risk analysis.
The biennal BELIEF conferences, sponsored by the Belief Functions and Applications Society, are dedicated to the confrontation of ideas, the reporting of recent achievements and the presentation of the wide range of applications of this theory. The first edition of this conference series was held in Brest, France, in 2010. Later editions were held in Compiègne, France in 2012, Oxford, UK in 2014, Prague, Czech Republic in 2016, and again in Compiègne, France in 2018. The previous edition was delayed by the global pandemic and took place in 2021 in Beijing, China. The Seventh International Conference on Belief Functions (BELIEF 2022) will be located in Paris, France, and it will take place on October 26-28, 2022.
The expected length of papers is no longer than 10 pages, references included, that should present original contributions with significant results. Springer encourages authors to include their ORCIDs in their papers. In addition, the corresponding author of each accepted paper, acting on behalf of all of the authors of that paper, will have to complete and sign a Consent-to-Publish form. The corresponding author signing the copyright form should match the corresponding author marked on the paper. Once the files have been sent to Springer, changes relating to the authorship of the papers cannot be made.
Authors of selected papers from the BELIEF 2022 conference will be invited to submit extended versions of their papers for possible inclusion in a special issue of the International Journal of Approximate Reasoning.
Original contributions are solicited on theoretical aspects including, but not limited to
as well as on applications to various areas including, but not limited to
The BELIEF 2022 Conference Committee is pleased to invite submissions of abstracts for poster presentations. The poster session aims to encourage discussions either around preliminary works or works in progress. For poster presentation, we also encourage the submission of abstracts corresponding to works in recent papers accepted or submitted elsewhere. The conference or journal where the paper was published or submitted must be specified at submission.
Submissions should include a title and an abstract no longer than 250 words. If accepted, they will be presented at the conference in a special poster session but will not appear in the proceedings. The main objective of the poster session is to foster discussions around preliminary works, or to popularize newly published works, therefore poster presenters are required to participate on site to the conference. The deadline for submissions is: July 31, 2022. Poster presenters will be notified by August 8th.
Abstract submission: email@example.com
Researcher Rémi Bardenet, CRIStAL, CNRS and Lille University, France.
Title: Repulsive point processes for Monte Carlo integration, and why I would like to be a Bayesian
Check the recorded video here
Professor Stéphane Canu, INSA Rouen, France.
Title: Robustness of neural networks and adversarial attacks
Check the slides here
Associate Director Ozgur Erdinc, Raytheon Technologies Research Center, USA.
Title: Challenges in Automating Mission-Critical Decision Making Systems: A Practitioner’s Perspective
Check the slides here
Associate Professor Philippe Xu, Université de Technologie de Compiègne, France.
Title: Fusion of heterogeneous deep neural networks with belief functions
Check the recorded video here
The event will take place at the International Conference Center of Sorbonne University, on Jussieu Campus. For traveling to the conference location inside Paris, please see the practical information paragraph below.
Once you arrive to the closest Metro station, "Jussieu" and to the main entrance of the campus ("ENTREE" in the plan below), head towards the tower number 44, as indicated on the plan. Then, go up to the first floor where you will find Hall 106, where the conference will take place. The itinerary should be indicated by some banners with the conference name and logo.
The welcome reception will take place at the end of the first day at the Sorbonne Center for Artificial Intelligence (SCAI), which is at a 5' walk from Hall 106. We will probably go together to SCAI, but for your convenience its location is indicated below :
The gala dinner will take place inside the Wine Museum, in the heart of Paris and a stone's throw away from the Eiffel Tower. Visiting this location is a great opportunity to discover the long history of this inherent part of French culture and tradition, and we are currently trying to secure a guided visit of the museum for the BELIEF 2022 participants (to be confirmed soon). Update: guided visit is confirmed. Optionally, the wine enthousiasts may subscribe to a blind wine tasting competition organized by the Museum hosts (you need to tick the Cultural Activity option during registration) - the winner gets a prize ! Update: the competition will take place just before the dinner, and will consist in a crash course given by a museum expert presenting the main regions and grape varieties, followed by the actual tasting. Thus, you don't need to be well accustomed to French wine to participate. Note however that if you are not a wine fan or if you don't drink alcohol, the gala dinner is non-alcohol friendly.
The museum is located at a 3' walk from Passy Metro Station (Metro Line 6). Its location is indicated below :
The registration fee for the onsite participants includes the following items:
Online participants benefit from a discounted registration fee, which includes the following items:
Each paper must have at least one author registered with regular on-site registration (i.e., not student registration) with payment received by the author-registration deadline of August 15 to avoid being withdrawn from the conference. A single, full registration may cover up to two (2) papers. When registering, please enter the ID number(s) of the paper(s) that you will be covering with your full registration; for the paper ID number, use the 5-digit Easychair Paper ID number.
If you are a student who will be presenting a paper, and you are covering that paper with your registration, then you must register with the regular on-site rate.
BELIEF 2022 requires that each accepted paper be presented by one of the authors in-person at the conference site according to the schedule published. Any paper accepted into the technical program, but not presented on-site will be withdrawn from the official proceedings archived on Springer. The only exception is in the case of all non-EU authors who might face travel difficulties. In this case, you are required to contact us as early as possible to inform us about your intention to present online.
A student presenting a poster and not covering a paper registration may register with the on-site student rate.
You may find below the registration price depending on the participant category (faculty or student), attendance type (on-site/on-line) and registration period (standard or late registration). Optionally, you may purchase:
|Registration fees (EUR)
|Standard fee* (ending Aug. 31, 2022)
|Late fee* (Sept. 1 - Sept. 30, 2022)
|Additional ticket for welcome reception and gala dinner
|Optional cultural activity
|Please note that the registration fees will switch according to Central European Time (CET)
Please note: for foreigners requiring invitation letters, contact us before attempting to pay the registration fee if you are not a co-author of an accepted paper. Provide a proof of affiliation with an academic institution, contact references and a list of publications on a public indexing engine such as Google Scholar. The organizing committee reserves the right to refuse attendance to individuals who can not demonstrate a clear link with the topic of the conference, and you will be informed. As stated by the registration policy, the registration fee is not refundable.
On-site and on-line participants have to register using this link (Azur-Colloque platform) and have to pay the suitable registration fee mentioned above. Before pursuing, make sure to read carefully the payment guidelines:
Participants are requested to familiarize themselves with their own applicable visa requirements well in advance of the conference. Present worldwide security regulations generally result in more stringent visa requirements and associated longer visa processing time. The BELIEF 2022 organizers cannot intervene with embassies or consulates on behalf of any participant.
In order to obtain a letter of invitation for visa purposes, contact the Organizing Committee directly.
As a general advice, public transportation is generally reliable inside Paris, which has a very dense subway (Metro) network, and it is the most effective way of getting around. If you go for this solution, it is best to get a tourist pass directly upon arrival, at the airport/train station : see here. Make sure the pass includes your starting point. When traveling, watch for pickpockets, unfortunately they exist and tend to target inattentive tourists.
If you arrive to CDG airport, you may use RER B line to get to central Paris. From Orly airport, the simplest option is the OrlyBus. Both are included in the travel passes. Taxi is a good idea if you travel at night or if you have really heavy luggage.
For the daily travel, if you have your accommodation a bit farther - outside Paris ring road, you are OK as long as you are at a walking distance from an RER stop. The bus network is also OK inside Paris for short and medium distances, but traffic is often bad and you are advised to use the subway if possible.
The conference location is at 100 meters from Metro station Jussieu (lines 7 and 10). Among various apps and sites that help you get an itinerary in Paris, we can suggest the official site : RATP. Paris is quite small and dense, so if you do not have luggage and if the weather is nice, do not be afraid to walk for 10-15 minutes from a large station (e.g. Saint-Michel Notre Dame) in order to avoid an additional subway change. There is also a popular shared bike service, if you are big fan of biking, and it is actually quite practical: Velib and there is an option to subscribe to the service for three days.
On-site presenters : the day of your presentation, please provide the slides of your talk in pdf format, in order to avoid compatibility issues. The slides will be copied on the computer that will be provided for presenting (as we need to broadcast the talks to online participants, you will not use your computer for presenting). You are strongly encouraged however if possible to send your slides before your presentation day to firstname.lastname@example.org.
Online presenters : please provide one week before the beginning of your conference, that is by Oct. 19, a 10-15 minute video of your talk, with clearly visible slides, good sound quality and if possible including the presenter in a small window as well. This file is needed as backup in case we experience serious connection issues during your conference presentation (caused by your network or ours). Then you will have two options: you may either present live during the conference and we keep this video as backup, or you may choose to use the video as presentation during the conference and you will answer questions directly (live). Let us know what you prefer when you send us the video, using Google Drive, Dropbox, Baidu drive, wetransfer etc. There are quite a lot of free programs available for recording a presentation, e.g. OBS. The link for downloading the 10-15 minute video file should be sent to email@example.com by Oct. 19, along with your choice for one of the two options above.
(Paris local time UTC+2)
BELIEF proceedings will be available here SpringerLink - Free access granted until end of November 2022.
Note that during coffee breaks and lunch breaks, food service will be provided by the conference.
|Wednesday October 26
|On-site welcome desk open
| Keynote by Philippe Xu
Fusion of heterogeneous deep neural networks with belief functions
session chair Sylvie Le Hégarat-Mascle
|Session 1 (Evidential Clustering)
| Keynote by Ozgur Erdinc
Challenges in Automating Mission-Critical Decision Making Systems: A Practitioner’s Perspective
session chair Anne-Laure Jousselme
|Session 2 (Data and Information Fusion)
|BFAS general assembly (open to conference participants)
|Thursday October 27
| Keynote by Stéphane Canu
Robustness of neural networks and adversarial attacks
session chair Anne-Laure Jousselme
|Session 3 (Machine Learning and Pattern Recognition)
|Session 4 (Algorithms and Evidential Operators)
|Poster session and coffee break
|Guided visit (Wine Museum)
|Oenology course and wine tasting contest (Wine Museum)
|Gala dinner (Wine Museum)
|Friday October 28
| Keynote by Rémi Bardenet
Topics in Monte Carlo computation and Bayesian learning
session chair Anne-Laure Jousselme
|Session 5 (Links with Other Uncertainty Theories or Machine Learning)
|Session 6 (Statistical Inference - Graphical Models)
|Session 7 (Applications)
|Farewell coffee break and conference closure
Instruction for oral presentation
Each presentation is scheduled to be 20 minutes long, including questions. Instruction to chairman is to leave 15 minutes for the presentation itself and 5 minutes for questions/discussions
|Session 1 (Evidential Clustering), Wednesday October 26, session chair Frédéric Pichon
| 11:15-11:35 A Distributional Approach for Soft Clustering Comparison and Evaluation
Andrea Campagner, Davide Ciucci and Thierry Denœux (on-site)
| 11:35-11:55 Causal transfer evidential clustering
Kuang Zhou and Ming Jiang (online)
| 11:55-12:15 Evidential clustering by Competitive Agglomeration
Lulu Xu, Qian Wang, Pei-hong Wang and Zhi-gang Su (online)
| 12:15-12:35 Imperfect Labels with Belief Functions for Active Learning
Arthur Hoarau, Arnaud Martin, Jean-Christophe Dubois and Yolande Le Gall (on-site)
| 12:35-12:55 A variational Bayesian clustering approach to acoustic emission interpretation including soft labels
Martin Mbarga Nkogo, Emmanuel Ramasso, Patrice Le Moal and Gilles Bourbon (on-site)
|Session 2 (Data and Information Fusion), Wednesday October 26, session chair Arnaud Martin
| 15:30-15:50 Heterogeneous Image Fusion for Target Recognition based on Evidence Reasoning
Shuyue Wang, Zhunga Liu, Zuowei Zhang and Yang Li (online)
| 15:50-16:10 Evidential Trustworthiness Estimation for Cooperative Perception
Antoine Lima, Veronique Cherfaoui and Philippe Bonnifait (on-site)
| 16:10-16:30 Cluster Decomposition of the Body of Evidence
Alexander Lepskiy (online)
| 16:30-16:50 An Intelligent System for Managing Uncertain Temporal Flood events
Manel Chehibi, Ahlem Ferchichi and Imed Riadh Farah (online)
|Session 3 (Machine Learning and Pattern Recognition), Thursday October 27, session chair Philippe Xu
| 11:00-11:20 An Evidential Neural Network Model for Regression Based on Random Fuzzy Numbers
Thierry Denœux (on-site)
| 11:20-11:40 Ordinal Classification using Single-model Evidential Extreme Learning Machine
Liyao Ma, Peng Wei and Bin Sun (online)
| 11:40-12:00 Evidential regression by synthesizing feature selection and parameters learning
Chao Liu, Zhi-gang Su and Gang Zhao (online)
| 12:00-12:20 Reliability-based imbalanced data classification with Dempster-Shafer theory
Hongpeng Tian, Zuowei Zhang, Arnaud Martin and Zhunga Liu (online)
|Session 4 (Algorithms and Evidential Operators), Thursday October 27, session chair Prakash P. Shenoy
| 13:30-13:50 Distributed EK-NN classification
Chaoyu Gong, Zhi-gang Su, Qian Wang and Yang You (online)
| 13:50-14:10 On improving a group of evidential sources with different contextual corrections
Siti Mutmainah, Hachour Samir, Frédéric Pichon and David Mercier (on-site)
| 14:10-14:30 Belief functions on ordered frames of discernment
Arnaud Martin (on-site)
| 14:30-14:50 Measure of Information Content of Basic Belief Assignments
Jean Dezert, Albena Tchamova and Deqiang Han (on-site)
| 14:50-15:10 On Modelling and Solving the Shortest Path Problem with Evidential Weights
Tuan-Anh Vu, Sohaib Afifi, Éric Lefèvre and Frédéric Pichon (on-site)
|Poster Session with coffee break, Thursday October 27, session chair Sébastien Destercke
|GCVAE: Generalized-Controllable Variational Autoencoder
presenter: Kenneth Ifeanyi Ezukwoke, Ecole des Mines de Saint-Étienne
|Evidential deep learning-based drivable area detection for autonomous vehicles
presenter: Mihreteab Negash Geletu, Université de Haute Alsace
|Evidence fusion with contextual discounting for multi-modality medical image segmentation
presenter: Ling Huang, Université de Technologie de Compiègne (download poster pdf)
|Handling Noisy Annotations in Deep Supervised Learning
presenter: Ichraq Lemghari, Safran and Paris-Saclay University
|Aggregating Belief functions using Copulas: how and when?
presenter: Roman Malinowski, Université de Technologie de Compiègne (download poster pdf)
|Learning calibrated belief functions from conformal predictions
presenter: Vitor Martin Bordini, Université de Technologie de Compiègne (download poster pdf)
|Belief Functions from Uncertain Evidence: A Computational Complexity Analysis
presenter: Daira Pinto Prieto, University of Amsterdam
|On computing evidential centroid through conjunctive combination: an impossibility theorem
presenter: Yiru Zhang, Cergy University (download poster pdf)
|Session 5 (Links with Other Uncertainty Theories or Machine Learning), Friday October 28, session chair Michael Hanss
| 11:00-11:20 Updating belief functions over Belnap–Dunn logic
Sajad Nazari, Sabine Frittella and Ondrej Majer (on-site) download slides
| 11:20-11:40 The Extension of Dempster’s Combination Rule Based on Generalized Credal Sets
Andrey G. Bronevich and Igor N. Rozenberg (online)
| 11:40-12:00 A qualitative counterpart of belief functions with application to uncertainty propagation in safety cases
Yassir Idmessaoud, Jérémie Guiochet and Didier Dubois (on-site)
| 12:00-12:20 A Correspondence between Credal Partitions and Fuzzy Orthopartitions
Stefania Boffa and Davide Ciucci (on-site) download slides
|Session 6 (Statistical Inference - Graphical Models), Friday October 28, session chair Thierry Denœux
| 13:30-13:50 Valid inferential models offer performance and probativeness assurances
Leonardo Cella and Ryan Martin (online)
| 13:50-14:10 On Conditional Belief Functions in the Dempster-Shafer Theory
Radim Jiroušek, Václav Kratochvíl and Prakash P. Shenoy (on-site)
| 14:10-14:30 A practical strategy for valid partial prior-dependent possibilistic inference
Dominik Hose, Michael Hanss and Ryan Martin (online)
|Session 7 (Applications), Friday October 28, session chair Emmanuel Ramasso
| 14:40-15:00 Addressing ambiguity in randomized reinsurance contracts using belief functions
Davide Petturiti, Gabriele Stabile and Barbara Vantaggi (on-site)
| 15:00-15:20 Real bird dataset with imprecise and uncertain values
Constance Thierry , Arthur Hoarau, Arnaud Martin, Jean-Christophe Dubois and Yolande le Gall (on-site)
| 15:20-15:40 Hybrid Artificial Immune Recognition System with improved belief classification process
Rihab Abdelkhalek and Zied Elouedi (online)
| 15:40-16:00 Evidential filtering and spatio-temporal gradient for micro-movements analysis in the context of bedsores prevention
Nicolas Sutton-Charani, Francis Faux, Didier Delignières, Willy Fagart, Arnaud Dupeyron and Marie Nourrisson (on-site)
We gratefully acknowledge the financial support of the following organizations and research groups: