Work trips.

The sanitary crisis around the COVID-19 pandemic has strongly disturbed the team's activities. Since the beginning of the first lockdown (March 2020), almost all relevant scientific events (conferences, workshops, team seminars. etc.) have been held online. This has drastically reduced the environmental footprint of our scientific missions.

Utilisation of computing resources.

LACODAM contributed last year with a new server (abacus12) to the Igrida computing platform. Being a team specialized in data science and machine learning, a recurrent task in LACODAM is to run CPU intensive algorithms on large data collections, for example, to train deep neural networks. Some of our ongoing PhD research topics (e.g., the theses of Heng Zhang, Simon Corbillé, and Simon Felton) concern deep learning technologies, and the increasing place of eXplanaible AI in our research program will boost our reliance on Igrida (notably with the PhD of Julien Delaunay and Victor Guyomard). This will increase the energetic and environmental footprint of our activities in a non-negligible way. We are therefore willing to collaborate with the institute's direction in any initiative that could mitigate such an impact.

7.1.1 HIPAR

• Name:
  Hierarchical Interpretable Pattern-aided Regression
• Keywords:
  Regression, Pattern extraction
• Functional Description:
  Given a (tabular) dataset with categorical and numerical attributes, HIPAR is a Python library that can extract accurate hybrid rules that offer a trade-off between (a) interpretability, (b) accuracy, and (c) data coverage.
• URL:
  https://­gitlab.­inria.­fr/­lgalarra/­hipar
• Contact:
  Luis Galarraga Del Prado

7.1.2 SNDS Data Generator

• Name:
  SNDS Synthetic Data Generator
• Keywords:
  Databases, Benchmarking
• Functional Description:

  The SNDS, formerly SNIIRAM, is a huge database (consisting of several TBs of data and about 700 tables) that contains information about healthcare reimbursements of about 60 million French insured patients. This database is used to carry out epidemiological and medical-economic studies. Due to its sensitive medical content, identifying information (names, social security numbers) is removed or replaced by spurious information. This repository contains a solution to generate a synthetic version of the database.

  The software generates relational data compliant with the original database schema with realistic distributions. This guarantees privacy preservation thanks to the use of open data only.

  This work is currently under review to the AIME conference.

• URL:
  https://­gitlab.­inria.­fr/­tguyet/­medtrajectory_datagen
• Contact:
  Thomas Guyet

7.1.3 SIERRA

• Name:
  Seasonal tIme sERies foRecaster Api
• Keywords:
  Time Series, Forecasting
• Functional Description:

  SIERRA is a software dedicated to forecasting of seasonal time series. It includes a set of prediction models that are trained from the history of a time series and that can be applied to new data. The input of these models is a seasonal time series. It predicts its evolution over one or several seasons ahead. Models can also be used to describe historical data by providing access to characteristics about typical seasons in the dataset. Two main categories of models are proposed: deterministic models that predict the exact evolution of a time series, and probabilistic models that predict a probability distribution over time.

  The SIERRA software is an API (Applied Programming Interface) which offers functionalities in the form of python module that can be embedded in a data analysis chain. The methods are accompanied by additional tools for the identification of seasonality, for the identification of the best configurations and the visualisation of the (probabilistic) predictions.

• News of the Year:
  During 2021, a version of the SIERRA software has been released. This version has been sell to Orange SA which is now in charge of doing the official APP deposit. Orange SA is currently continuing the development of the sierra software based on this release.
• Publication:
  hal-02371221
• Contact:
  Thomas Guyet
• Participants:
  Thomas Guyet, Colin Leverger
• Partner:
  Orange Labs

8.1.1 Pattern Mining

Participants: Alexandre Termier, Tassadit Bouadi, Elisa Fromont, Luis Galarraga, Christine Largouët, Laurence Rozé, Véronique Masson, Peggy Cellier, Lénaïg Cornanguer, Josie Signe, Erwan Bourrand, Simon Coumes.

Regarding this research sub-axis, our contributions this year cover sequential or timestamped data for the sake of predicting future events, efficient enumeration of skyline groups, and subgroup discovery.

8.1.2 Semantic Web

Participants: Luis Galarraga, Thomas Guyet, Johanne Bakalara.

Our contributions to the domain of Semantic Web cover different topics of interest for knowledge producers, namely archiving and provenance management, as well as one application for data consumers on epidemiology.

Pattern-aided Regression.

HiPaR5 33 is a pattern-aided regression method that mines accurate and compact sets of hybrid rules. These are rules of the form $p\Rightarrow y=f\left(X\right)$ where $p$ is a set of conditions on categorical attributes and the right-hand side is a regression model that is only valid for the points that match $p$. HiPaR consists of two phases: an enumeration stage that explores the space of hybrid rules efficiently, and a selection phase that picks a small set of rules that guarantees both joint prediction accuracy and coverage. Our experimental evaluation shows that this method strikes a best trade-off between model simplicity (measured by the number of rules and their length) and prediction performance among existing interpretable and non-interpretable regression methods.

Explainable Multivariate Time Series (MTS) Classification.

The current state-of-the-art MTS classifier is a heavyweight deep learning black-box approach that outperforms the second-best MTS classifier only on large datasets. To account for the need for explainability in MTS classification, 1 presents XCM, an eXplainable Convolutional neural network for MTS classification that extracts information relative to the observed variables and time directly from the input data. Thus, XCM architecture enables a good generalization ability on both large and small datasets, while allowing the full exploitation of a faithful post hoc model-specific explainability method (Gradient-weighted Class Activation Maps) by precisely identifying the observed variables and timestamps of the input data that are important for predictions. Experiments on real and synthetic data show the superiority of XCM in terms of performance and interpretability.

Performance-Explainability Benchmarking.

23 proposes a new performance-explainability analytical framework to assess and benchmark machine learning methods. The framework details a set of characteristics that systematize the performance-explainability assessment of existing machine learning methods. The framework has been applied to MTS classifiers.

User Impact of Performance Predictions for Term Planning.

13 offers a user study on the impact of showing academic performance predictions to students using the iCoRA term planning tool  39. This tool allows students to plan the courses to be taken the next semester, and provides ML-based recommendations in the form of GPA performance predictions for each course based on different factors, e.g., workload, course stringency factors, the student's academic history, etc. The study sheds light on the impact of showing those predictions according to three criteria: behavior, decisions, and preferences. Moreover the article studies the effect of the visual representation used to convey the predictions, as well as the effect of providing global GPA predictions and feature-attribution explanations for the system's outcomes. The study shows, among other things, that users praise simplicity and credibility for predictions, and that specific visual representation (e.g., value ranges, punctual predictions) encourages the users to invest more time and effort in the task.

8.3.1 Computer Vision and Robotics

Participants: Elisa Fromont, Samuel Felton, Heng Zhang.

8.3.2 Recommender Systems

Participants: Elisa Fromont, Romaric Gaudel, Camille Sovanneary Gauthier.

8.3.3 Forecasting

Participants: Thomas Guyet, Alexandre Termier, Colin Leverger.

8.3.4 Logistics

Participants: Alexandre Termier, Elisa Fromont, Laurence Rozé, Grégory Martin.

Interference Attacks on Aggregated Time Series.

Aggregation is widely used as a privacy protection method. Aggregate membership inference attacks aim to determine whether or not a given target participated in the computation of the aggregate under attack. In 28, the authors study the vulnerability of aggregated time series—where each point is a time-stamped aggregate—to membership inference attacks. The considered attacker has auxiliary knowledge about a superset of the aggregated data (e.g., from a data leak). 28 proposes a new attack leveraging this type of auxiliary knowledge and the multiple points forming the aggregate time series. The attack is modeled as an integer linear optimization problem, allowing the attacker to benefit from the power of dedicated solvers (e.g., Gurobi). This attack, tested on public datasets, shows the vulnerability of an aggregated time series publication if the number of aggregated series is too small compared to the number of points constituting the series.

Temporal Phenotyping applied to Care Pathways for COVID-19 patients.

During the COVID19 crisis, Intensive Care Units admitted many patients with breathing disorders up to respiratory insufficiency. The care strategy of such patients was difficult to find and preventing patients to drift away toward a critical situation was one of the first challenges of physicians. In 21, the authors characterize care pathways of patients that required a mechanical ventilation (invasive treatment for the most critical respiratory insufficiencies). Through the analysis of the sequence of cares, the goal of this work is to support physicians in understanding the evolution of patients, and let them propose new medical procedures for preventing ventilation. This article proposes a method which combines a tensor factorization and sequence clustering. The tensor factorization enables to represent the care sequences as a sequence of daily phenotypes. Then, the sequences of phenotypes is clustered to extract typical care trajectories. This method is experimented on real data from Greater Paris university Hospital and is compared to a direct clustering of the sequences. The results show that the outputs are more easily interpretable with the proposed method.

Benchmarking for Healthcare Systems.

The recent development of data analysis provides opportunities for improving healthcare systems through analysis of health databases. However this thirst for data conflicts with the preservation of the privacy of individuals. In that vibe, the generation of synthetic datasets may foster research on healthcare data analytics. Synthetic data is mostly based on generative statistical models fitted on real data. 25 proposes a probabilistic relational model fitted on publicly available datasets. More specifically, the authors propose to generate a synthetic version of the national database of French insured patients. The authors do not only provide synthetic datasets, but a generator of datasets that can be used without any data access request. Experiments compare official statistics with those computed on synthetic datasets and show the potential use of plausible synthetic data.

8.5.1 Agriculture

Participants: Christine Largouët, Raphaël Gauthier.

Fishing and River Resources.

6 studies the perceptions of fishers on the interactions between their practices and their environment. The study was performed in the upper section of the Maroni River in French Guiana, a relatively remote region in the tropical rainforest where subsistence fishing still occurs. We assessed the fishers' perceptions of their relationship with the river by asking them about the state of natural resources, their fishing practices, nearby activities, and their way of life. Cognitive mapping was used to capture their individual viewpoints, especially those that formed a consensus with the other fishers. Regardless of their ethnic group (Aluku vs. Amerindian) or way of life (subsistence vs. commercial fishing), most fishers generally shared the same views. The main perception was that fishing is threatened by illegal gold mining, increasing use of fishing nets, and a loss of knowledge of fish behavior by younger generations of villagers. Furthermore, fishers perceived an ongoing shift in their role and relationship with the fish resource, which is becoming increasingly commercially oriented, and since the river is no longer the only source of food. Detailed analysis of arguments put forward to explain these threats shows that this process originates from ill-managed Westernization, which has caused painful changes in lifestyles of local populations, especially Amerindians. This analysis can also provide local governments with mechanisms for action. Our results raise questions about the future of this region and suggest ways to protect its natural resources better. They can help decision makers respond to poorly understood informal fisheries and motivate local residents to contribute to sustainable management of a river's natural resources.

This work is complemented by a survey 7 that sheds light on the fishing practices in the Maroni River in French Guiana. The authors surveyed 754 boat landings in seven villages located in the upper half of the watershed, representing > 6300 fish during the study period (November 2013 - September 2014). Fishers used canoes with outboard engines almost exclusively (75 %) and fished within 32 km of their villages. Most fish were caught in trammel nets (81 %); the 20 most-landed species represented more than 87% of catches. Depending on the village, daily catches and biomass averaged 6–14 fish and 1.7–13 kg per boat landing, respectively. Seven control sites located outside of the fishing grounds were fished to identify potential differences in catch per unit effort and fish size. Per 100 m2 of trammel net, mean catches ranged from 4 to 13 and 8–29 fish in the villages and control sites, respectively, while fish biomass ranged from 0.9 to 4 and 3.2–7 kg in villages and control sites, respectively. For all species combined, fish caught at control sites were bigger than those landed in villages. This difference was significant for nine of the most-landed species. Differences in fishing techniques and fish catches between villages illustrated the gradual disappearance of the ancestral subsistence fishing. Our results support indications that the fish community in the upper Maroni River is harvested intensively, address the issue of sustainability of the fishery there, and call attention to the need to conserve the river's remarkable biodiversity.

Edge-to-Cloud Computing for Earthquake Detection.

The growth of the Internet of Things is resulting in an explosion of data volumes at the Edge of the Internet. To reduce costs incurred due to data movement and centralized cloud-based processing, it is becoming increasingly important to process and analyze such data closer to the data sources. Exploiting Edge computing capabilities for stream-based processing is however challenging. It requires addressing the complex characteristics and constraints imposed by all the resources along the data path, as well as the large set of heterogeneous data processing and management frameworks. Consequently, the community needs tools that can facilitate the modeling of this complexity and can integrate the various components involved. The work in 10 introduces MDSC, a hierarchical approach for modeling distributed stream-based applications on Edge-to-Cloud continuum infrastructures. The article demonstrates how MDSC can be applied to a concrete real-life ML-based application—early earthquake warning—to help answer questions such as: when is it worth decentralizing the classification load from the Cloud to the Edge and how?

10.1.1 FP7 & H2020 projects

TAILOR (Foundations of Trustworthy AI – Integrating Reasoning, Learning and Optimization) H2020-ICT-2019-3 Elisa Fromont, Luis Galárraga and Alexandre Termier are all involved in the European project in particular in WP3 about Trustworthy AI.

10.2.1 ANR

• FAbLe: Framework for Automatic Interpretability in Machine Learning

  Participants: L. Galárraga (holder), C. Largouët

  Participants: Luis Galárraga (holder), Christine Largouët, Julien Delaunay.

  How can we fully automatically choose the best explanation for a given use case in classification?. Answering this question is the raison d’être of the JCJC ANR project FAbLe. By “best explanation” we mean the explanation that yields the best trade-off between interpretability and fidelity among a universe of possible explanations. While fidelity is well-defined as the accuracy of the explanation w.r.t the answers of the black-box, interpretability is a subjective concept that has not been formalized yet. Hence, in order to answer our prime question we first need to answer the question: “How can we formalize and quantify interpretability across models?”. Much like research in automatic machine learning has delegated the task of accurate model selection to computers  41, FAbLe aims at fully delegating the selection of interpretable explanations to computers. Our goal is to produce a suite of algorithms that will compute suitable explanations for ML algorithms based on our insights of what is interpretable. The algorithms will choose the best explanation method based on the data, the use case, and the user’s background. We will implement our algorithms so that they are fully compatible with the body of available software for data science (e.g., Scikit-learn).

11.1.1 Scientific events: organisation

11.1.2 Scientific events: selection

11.1.3 Journal

11.1.4 Invited talks

Elisa Fromont did the following invited talks:

• 2/10/2021: Invited speaker (in French) for the "Journée Recherche du Labex DigiCosme" on the topic "Explainable Time Series Classification" (virtual).
• 05/10/2021: Keynote speaker (in French) for the Conférence sur l'Intelligence Artificielle organized by ESIR, ENSSAT et IGR-IAE Rennes.
• 29/09/2021: Invited speaker (in French) GDR IA fall school on "(Deep) neural networks: basic principles and explanation issues", Paris.
• 16/09/2021: Keynote speaker (in French) for the conference ORASIS (Journées francophones des jeunes chercheurs en vision par ordinateur) on the topic "Multispectral Object Detection" (virtual).
• 29/06/2021: Invited speaker (in French) for JSI (Journées Scientifiques INRIA) on the topic "Challenges in computer vision and a few contributions" (virtual).
• 21/06/2021: Invited speaker (in French) Colloque DRIAS (Droit, IA et Santé) "Apprentissage Machine Explicable", Rennes (virtual).
• 22/04/2021: Invited speaker (in French) Cycle IA & IMAGES du Breizh Data Club "Analyse de scènes extérieures par des méthodes de deep learning: problèmes de fusion de données, de déséquilibres de classes et d'adaptation de domaine", Rennes (virtual).
• 08/03/2021: Invited speaker for MIAI legality & inclusion day, Grenoble (virtual).

Christine Largouët did the following talk:

• 27/03/2021 Invited Talk : "Data vs. Concept driven models" in "Journée ATMOSPHASE", INRAE organized by AgroParisTech.

Alexandre Termier did the following invited talks:

• 25/11/2021: Invited speaker at the workshop "Explainable and Responsible AI" of the Multidisciplinary Institute in Artificial Intelligence of Grenoble-Alpes University, title: “HiPaR: Hierarchical Pattern-Aided Regression”
• 08/12/2021: Invited speaker at ENS Rennes seminar for 2nd year students, title: “Pattern Mining: a biased perspective”.

11.1.5 Scientific expertise

• Elisa Fromont is a member of the scientific council of the GDR IA, the Machine Learning College at AFIA, and of the “Société Savante Francophone d'Apprentissage Machine” (SSFAM).
• Elisa Fromont was a member of a recruitement committee in Leuven, Belgium; Position MCF_27MCF0950 at (LIFO; University); Position PR_4641 at Nantes (LS2N, University).
• Alexandre Termier and Christine Largouët were members of an associate professor recruitement committee in Rennes (Institut Agro, Agrocampus Ouest).
• Peggy Cellier was member of an associate professor recruitement committee in Orléans (Université d'Orléans).
• Alexandre Termier was member of the recruitement committees for permanent researcher (CRCN) positions at Inria Renne and for the MCF_1261 associate professor position at ESIR, University of Rennes 1.

11.1.6 Research administration

• Since September 2018, Peggy Cellier is in charge of the Irisa Ph.D. students at IRISA, i.e. she is involved in the "commission du personnel" and organizes the selection of Ph.D. students for ministerial grants (contrats doctoraux). She is also an elected member of the “Conseil de Composante IRISA/INSA” at INSA and an elected member of the "Conseil de laboratoire" at IRISA. Peggy Cellier is "secrétaire" of "Revue de Traitement automatique des langues" since 2019.
• Since 2017, Elisa Fromont is elected at IRISA lab council (she is a member of the gender equality group and responsible for the anti-harassment group). She is elected, since 2020, at the scientific council of the University (UR1) and is a member of the HDR committee for the University. She is the head of the D7 scientific departement at IRISA since Sept 2021 (and part of the direction scienfic board).

11.2.1 Teaching

Some members of the project-team LACODAM are also faculty members and are actively involved in computer science teaching programs in ISTIC, INSA, Agrocampus-Ouest, and ENSAI. Besides these usual teachings LACODAM is involved in the following programs:

• Alexandre Termier is responsible for the following courses at ISTIC (Univ. Rennes 1): Object Programming (L2 info, elec, maths), AI (M1 info), Data Mining and Visualization (M2 SIF).
• ENSAI's third year (M2 level): Machine Learning modules (Deep Learning, large-scale Learning, Recommender Systems), 21h, ENSAI (R. Gaudel)
• Master Smart Data: Bandits Theory et Recommender Systems, 9h, ENSAI (R. Gaudel)
• At INSA, Peggy Cellier is responsible of four courses: "Databases and web development" (Licence 3 INFO), "Databases" (Licence 3 Math), "Data Mining" (Licence 3) and "Advanced Database and Semantic Web" (Master 2). She also teaches some other courses: "Database" (Licence 2), "Use and functionalities of an operating system" (Licence 3). At master 2 SIF, she teaches in English 4 hours in the data mining course (DMV). In addition she gives a lecture of 2 hours also in master 2 SIF about "Qu’est-ce qu’une thèse, un doctorat, un·e doctorant·e ?".
• Participation in the module “Case Study in Data Science” with the seminar “Interpretable AI”, M2 EIT DSC, ISTIC, Univ. Rennes I (L. Galárraga, 4h).

11.2.2 Supervision

Postdocts

• Alexandre Termier and Elisa Fromont were supervisors of the post-doc of Neetu Kushwaha (until February 2021) and the M2 Internship of Ezanin Christian Bile (Polytechnique, from April to September 2021)
• Luis Galárraga co-supervises the following post-doctoral fellows: Daniel Hernández working on how-provenance for SPARQL queries in collaboration with Katja Hose from Aalborg University, and Mohit Mihal post-doctoral fellow from the Sequel team at Inria Lille in collaboration with Philippe Preux.

PhD. Students

• Elodie Germani, 2021-2024; supervisors: Élisa Fromont and Camille Maumet; title: Metric learning for robust FMRI pipelines.
• Abderaouf Nassim Amalou (PhD, UR1/Projects) 2021-2024; supervisors: Élisa Fromont and Isabelle Puaut; title: Machine Learning for Timing Estimation.
• Olivier Gauriau, (Inria, DigitAg, Acta Toulouse) 2021-2024; supervisors: Alexandre Termier, Luis Galárraga, François Brun, and David Makowski; title: Numerical Rule Mining for the Prediction of the Dynamics of Crop Diseases.
• Antonin Voyez, (PhD, CIFRE Enedis) 2020-2023; supervisors: Élisa Fromont, Tristan Allard and Gildas Avoine; title: Privacy-preserving Power Consumption Time-series Publishing.
• Julie Boudebs, 2021-2024; supervisors: Peggy Cellier and Sébastien Ferré (SemLIS), title: Un assistant en langue naturelle pour interroger le Web sémantique, ED MathStic.
• Simon Corbillé, (PhD, UR1) 2019-2022; supervisors: Élisa Fromont and Eric Anquetil; title: Explainable Deep-learning-based Methods for Children Handwriting Analysis in Education.
• Samuel Felton, (PhD, UR1) 2019-2022; supervisors: Élisa Fromont and Erich Marchand; title: Deep Learning for End-to-end Visual Servoing.
• Gregory Martin, (PhD, CIFRE PSA) 2019-2022 supervisors: Élisa Fromont, Alexandre Termier, and Laurence Rozé; title: Data mining to Optimize a Free-floating Car Sharing Service.
• Camille-Sovanneary Gauthier, (PhD, CIFRE Vuitton) 2019-2022; supervisors: Romaric Gaudel and Élisa Fromont, title: Bandit-based Recommender Systems.
• Heng Zhang, (PhD, CIFRE Atermès) 2018-2021; supervisors: Élisa Fromont and Sébastien Lefèvre. Rémi Emonet, Romain Tavenard, and Simon Malinowski; title: Multispectral Object detection
• Yichang Wang, (PhD, CSC) 2018-2021; supervisors: Élisa Fromont, Rémi Emonet, Romain Tavenard and, Simon Malinowski; title: Interpretable Time Series Classification
• Victor Guyomard, 2020-2023; supervisors: Tassadit Bouadi, Thomas Guyet, Françoise Fessant (Orange Labs) and Alexandre Termier, title: Explaining individual decisions made by an AI algorithm.
• Josie Signe, 2020-2023; supervisors: Peggy Cellier, Yannick Le Cozler (Inrae), Véronique Masson and Alexandre Termier, title: Animal Welfare, Characterizing the Diversity between and within Livestock Farming Situations with Data Mining Methods used on Information from Dairy Herd Sensors, ED MathStic.
• Lénaïg Cornanguer, 2020-2023, supervisors: Christine Largouët, Alexandre Termier and Laurence Rozé; title: Timed Automata Learning, ED MathSTIC
• Julien Delaunay, (Inria, ANR) 2020-2023; supervisors: Christine Largouët and Luis Galárraga; title: Automatic Construction of Explanations for AI Models, ED MathSTIC.
• Maëva Durand, 2020-2023; supervisors: Christine Largouët, Charlotte Gaillard (INRAE) and Jean-Yves Dourmad (INRAE); title: Real-time Integration of Gestating Sow Welfare and Health from Heterogeneous Data for Precision Feeding, ED EGAAL.
• Hugo Ayats, 2020-2023; supervisors: Peggy Cellier and Sébastien Ferré (SemLIS), title: De la prédiction à l'automatisation avec une IA explicable et centrée-utilisateur – application à la construction de graphes de connaissances, ED MathStic.
• Olivier Pelgrin, (PhD, AAU) 2019-2022; supervisors: Katja Hose and Luis Galárraga; title: Fully-fledged Archiving for RDF Datasets.
• Johanne Bakalara, 2018-2021; supervisors: Thomas Guyet, E. Oger, O. Dameron, A. Happe; title: Temporal Models of Care Sequences for the Exploration of Medico-administrative Data.
• Francesco Bariatti, 2018-2021; supervisors: Peggy Cellier and Sébastien Ferré (SemLIS); title: Mining Tractable Sets of Graph Patterns with the Minimum Description Length Principle, ED MathStic.
• Raphaël Gauthier, 2017-2021, supervisors: Christine Largouët and Jean-Yves Dourmad (INRAE), title: "Precision feeding system for lactating sows, using modeling and machine learning", ED EGAAL.

11.2.3 Juries

• Alexandre Termier was a member of the following PhD juries in 2021: Yichang Wang, 20/09 Rennes (committee member, president); Amaury Bouchra Pilet, 10/11 Rennes (committee member, president); Francesco Bariatti, 23/11 Rennes (committee member, president); Luong Nguyen, 10/12 Tours (reviewer). He was a member of the mid-term evaluation juries of Grégory Siekaniec (University of Rennes 1) and Daniel Rosendo (University of Rennes 1).
• Elisa Fromont was a member of the following PhD juries in 2021: Ian Jeantet, 5/01 Rennes (committee member, president); Corentin Lonjarret, 12/01 Lyon (committee member, president); Eduardo Hugo Sanchez, 17/02 Toulouse (committee member, president); Yann-Raphael Lifchitz, 20/04, Rennes (committee member, president); Ahmed Nassar, 21/05 Vannes (committee member, president); Alexandre Araujo, 1/06, Paris (committee member, president); Loïck Bonniot, 10/06, Rennes (committee member, president); Luca Erculiani, 10/06 Trento (reviewer); Tatiana Makhalova, 23/06 Nancy (committee member); Elies Gherbi, 5/07 Evry (reviewer); Benjamin Lucas, Monash, Australia (reviewer); Guillaume Renton, 08/07 Rouen (committee member); Yichang Wang, 20/09 Rennes (co-surpervisor); Nitika Verma, 20/10 Grenoble (committee member); Vincent Le Guen, 30/11 Paris (committee member); Tanguy Kerdoncuff, 9/12 Saint-Etienne (committee member, president); Grégoire Siekaniec, 10/12 Rennes (committee member, president); Victor Bouvier, 13/12 Paris (committee member); Heng Zhang, 14/12 Rennes (co-supervisor).
• Christine Largouët was a member of the following PhD juries: Raphaël Gauthier 05/03/2021 (co-supervisor), Yang Su 22/10/2021 (Committee member).
• Luis Galárraga was a member of mid-term evaluation committee of the following PhD candidates: Armand Boschin (Télécom ParisTech, 05/2021) working on learning and exploiting knowledge graph representations, Hugo Ayats (Univ. Rennes I, 05/2021) working on relation extraction via concepts of neighbors, Louis Béziaud (Univ. Rennes I/UAQM, 05/2021) working on privacy and fairness in law, and Armita KHajeh Nassiri (Univ. Paris Saclay, 11/2021) working on discovering expressive rules in knowledge graphs
• Peggy Cellier was a member of the following PhD juries: Cheikh Brahim EL VAIGH, 07/01/2021, Univ. Rennes I (committee member); Francesco Bariatti, 23/11 Rennes (co-supervisor). She was a member of mid-term evaluation committee of the following PhD candidates: Cyrielle Mallart (Univ. Rennes I), Priscilla Keip (Université de Montpellier), Albeiro Espinal (IMT Atlantique), Grégory Martin (Univ. Rennes I).

11.2.4 Education

• Veronique Masson is the head of the L3 studies in Computer Science at University of Rennes 1
• Since September 2021, Alexandre Termier is co-head of Master 2 SIF (Science Informatique - research master in Computer Science) at University of Rennes 1, with Bertrand Coüasnon (INSA Rennes).
• Elisa Fromont is reponsible for the Master MEEF NSI (a master dedicated to future high school teachers of computer sciences)
• Christine Largouët is responsible for the computer-science educational program in Institut Agro, Agrocampus Ouest (2 engineering schools)
• Romaric Gaudel is responsible for the department of computer-science at ENSAI (École nationale de la statistique et de l'analyse de l'information).
• Since October 2021, Peggy Cellier is responsible of the last year at Computer Science Department at INSA (master 2 level, about 70 students).

11.2.5 Interventions

• Camille-Sovanneary Gauthier participated to the challenge "ma thèse en 180 secondes" and was selected for the inter-regional final.
• Lénaïg Cornangueur participated to the program L codent L créent that raises awareness of programming among middle school girls during creative workshops.

International journals

• 1 articleK.Kevin Fauvel, T.Tao Lin, V.Véronique Masson, É.Élisa Fromont and A.Alexandre Termier. XCM: An Explainable Convolutional Neural Network for Multivariate Time Series Classification.Mathematics 923December 2021, 1-20
  HALDOIback to text
• 2 articleS.Samuel Felton, P.Pascal Brault, E.Elisa Fromont and E.Eric Marchand. Visual Servoing in Autoencoder Latent Space.IEEE Robotics and Automation Letters2022
  HALDOIback to text
• 3 articleC.Charlotte Gaillard, M.Maëva Durand, C.Christine Largouët, J.-Y.Jean-Yves Dourmad and C.Céline Tallet. Effects of the environment and animal behavior on nutrient requirements for gestating sows: Future improvements in precision feeding.Animal Feed Science and Technology2792021, 1-17
  HALDOIback to text
• 4 articleL.Luis Galárraga, D.Daniel Hernández and K.Katja Hose. Computing How-provenance for SPARQL queries via Query Rewriting.Proceedings of the VLDB Endowment (PVLDB)1413September 2021, 3389-3401
  HALDOIback to text
• 5 articleR.Raphaël Gauthier, C.Christine Largouët, L.Laurence Rozé and J.-Y.Jean-Yves Dourmad. Online forecasting of daily feed intake in lactating sows supported by offline time-series clustering, for precision livestock farming.Computers and Electronics in Agriculture1882021, 106329
  HALDOIback to text
• 6 articleG.Guillaume Longin, L.Louis Bonneau De Beaufort, G.Guy Fontenelle, R.Raphaelle Rinaldo, J.-M.Jean-Marc Roussel and P.-Y.Pierre-Yves Le Bail. Fishers’ perceptions of river resources: case study of French Guiana native populations using contextual cognitive mapping.Cybium : Revue Internationale d’Ichtyologie4512021, 5-20
  HALDOIback to text
• 7 articleG.Guillaume Longin, G.Guy Fontenelle, L.Louis Bonneau De Beaufort, C.Chrystelle Delord, S.Sophie Launay, R.Raphaëlle Rinaldo, G.Gilles Lassalle, P.-Y.Pierre-Yves Le Bail and J.-M.Jean-Marc Roussel. When subsistence fishing meets conservation issues: Survey of a small fishery in a neotropical river with high biodiversity value.Fisheries Research241September 2021, 1-9
  HALDOIback to text
• 8 articleO.Olivier Pelgrin, L.Luis Galárraga and K.Katja Hose. Towards Fully-fledged Archiving for RDF Datasets.Semantic Web – Interoperability, Usability, Applicability126April 2021, 903-925
  HALDOIback to text
• 9 articleY.Yiru Zhang, T.Tassadit Bouadi, Y.Yewan Wang and A.Arnaud Martin. A distance for evidential preferences with application to group decision making.Information Sciences5682021, 113-132
  HALDOIback to textback to text

International peer-reviewed conferences

• 10 inproceedingsD.Daniel Balouek-Thomert, P.Pedro Silva, K.Kevin Fauvel, A.Alexandru Costan, G.Gabriel Antoniu and M.Manish Parashar. MDSC: Modelling Distributed Stream Processing across the Edge-to-Cloud Continuum.DML-ICC 2021 workshop (held in conjunction with UCC 2021)Leicester, United KingdomDecember 2021
  HALback to text
• 11 inproceedingsL.Lénaïg Cornanguer. Passive Learning of Timed Automata from Logs (Student Abstract).Proceedings of the 35th AAAI Conference on Artificial Intelligence (AAAI-21)AAAI 2021 - 35th AAAI Conference on Artificial IntelligenceVancouver (virtual), CanadaFebruary 2021, 1-2
  HALback to text
• 12 inproceedingsS.Samuel Felton, E.Elisa Fromont and E.Eric Marchand. Siame-se(3): regression in se(3) for end-to-end visual servoing.ICRA 2021 - IEEE International Conference on Robotics and AutomationXi'an, ChinaIEEEMay 2021, 14454-14460
  HALback to text
• 13 inproceedingsL.Luis Galárraga, G.Gonzalo Mendez and K.Katherine Chiluiza. Showing Academic Performance Predictions during Term Planning: Effects on Students’ Decisions, Behaviors, and Preferences.CHI 2021 - ACM CHI Conference on Human Factors in Computing SystemsYokohama, JapanACMMay 2021, 1-17
  HALDOIback to text
• 14 inproceedingsC.-S.Camille-Sovanneary Gauthier, R.Romaric Gaudel and E.Elisa Fromont. Bandit Algorithm for Both Unknown Best Position and Best Item Display on Web Pages.IDA 2021 - 19th International Symposium on Intelligent Data AnalysisAdvances in Intelligent Data Analysis XIXPorto (virtual), Portugal2021, 12
  HALback to text
• 15 inproceedingsC.-S.Camille-Sovanneary Gauthier, R.Romaric Gaudel, E.Elisa Fromont and B. A.Boammani Aser Lompo. Parametric Graph for Unimodal Ranking Bandit.ICML 2021 - International Conference on Machine Learning139Proceedings of the 38th International Conference on Machine LearningVirtual, Canada2021, 3630--3639
  HALback to textback to textback to text
• 16 inproceedingsR.Raphaël Gauthier, C.Christine Largouët, L.Laurence Rozé and J.-Y.Jean-Yves Dourmad. Algorithm for real-time prediction of daily feed intake in lactating sows.Journées de la Recherche Porcine en France53. Journées de la Recherche Porcine5353èmes Journées de la recherche porcineEn ligne, FranceIfipFebruary 2021, 127-132
  HAL
• 17 inproceedingsC.Colin Leverger, T.Thomas Guyet, S.Simon Malinowski, V.Vincent Lemaire, A.Alexis Bondu, L.Laurence Rozé, A.Alexandre Termier and R.Régis Marguerie. Probabilistic forecasting of seasonal time series Combining clustering and classification for forecasting.ITISE 2021 - 7th International Conference on Time Series and ForecastingGran Canaria, SpainJuly 2021, 1-13
  HALback to text
• 18 inproceedingsO.Olivier Pelgrin, L.Luis Galárraga and K.Katja Hose. TrieDF: Efficient In-memory Indexing for Metadata-augmented RDF.CEUR-WS.org 2021 - CEUR Workshop ProceedingsVirtual Event, FranceOctober 2021, 1-10
  HALback to text

Conferences without proceedings

• 19 inproceedingsJ.Johanne Bakalara, T.Thomas Guyet, O.Olivier Dameron, A.André Happe and E.Emmanuel Oger. An extension of chronicles temporal model with taxonomies -Application to epidemiological studies.HEALTHINF 2021 - 14th International Conference on Health Informaticsonline, FranceFebruary 2021, 1-10
  HALback to text
• 20 inproceedingsE.Erwan Bourrand, L.Luis Galárraga, E.Esther Galbrun, E.Elisa Fromont and A.Alexandre Termier. Discovering Useful Compact Sets of Sequential Rules in a Long Sequence.ICTAI 2021 - 33rd IEEE International Conference on Tools with Artificial IntelligenceVirtual, United StatesIEEENovember 2021, 1-5
  HALback to text
• 21 inproceedingsM.Mathieu Chambard, T.Thomas Guyet, Y.-L.Yên-Lan Nguyen and E.Etienne Audureau. Temporal phenotyping for characterisation of hospital care pathways of COVID19 patients.AALTD 2021 - The 6th International Workshop on Advanced Analytics and Learning on Temporal DataBilbao / Virtual, SpainSeptember 2021, 1-16
  HALback to text
• 22 inproceedingsS.Simon Coumes, T.Tassadit Bouadi, L.Lhouari Nourine and A.Alexandre Termier. Skyline Groups Are Ideals. An Efficient Algorithm for Enumerating Skyline Groups.IWOCA 2021 - 32nd International Workshop on Combinatorial Algorithms12757Lecture Notes in Computer ScienceOttawa, CanadaSpringer International PublishingJune 2021, 223-236
  HALDOIback to textback to text
• 23 inproceedingsK.Kevin Fauvel, V.Véronique Masson and E.Elisa Fromont. A Performance-Explainability Framework to Benchmark Machine Learning Methods: Application to Multivariate Time Series Classifiers.IJCAI-PRICAI 2020 - Workshop on Explainable Artificial Intelligence (XAI)Yokohama, JapanJanuary 2021, 1-8
  HALback to text
• 24 inproceedingsC.-S.Camille-Sovanneary Gauthier, R.Romaric Gaudel, E.Elisa Fromont and A. B.Aser Boammani Lompo. Ordonnancement d'objets par bandits unimodaux sur des graphes paramétriques.CAp 2021 - Conférence sur l'Apprentissage automatiqueSaint-Etienne (en ligne), FranceJune 2021, 1
  HALback to text
• 25 inproceedingsT.Thomas Guyet, T.Tristan Allard, J.Johanne Bakalara and O.Olivier Dameron. An open generator of synthetic administrative healthcare databases.Actes de l'atelier Intelligence Artificielle et Santé (IAS)IAS 2021 - Atelier Intelligence Artificielle et SantéBordeaux (virtuel), FranceJune 2021, 1-8
  HALback to text
• 26 inproceedingsG.Gregory Martin, M.Matthieu Donain, E.Elisa Fromont, T.Tias Guns, L.Laurence Roze and A.Alexandre Termier. Prediction-Based Fleet Relocation for Free Floating Car Sharing Services.ICTAI 2021 - 33rd International Conference on Tools with Artificial IntelligenceVirtual, United StatesNovember 2021, 1-5
  HALback to text
• 27 inproceedingsJ.Josie Signe. Extraction de sous-groupes exceptionnels de séries temporelles.RJCIA 2021 - Rencontres des Jeunes Chercheurs en Intelligence ArtificielleBordeaux / Virtual, FranceJuly 2021, 89-90
  HALback to text
• 28 inproceedingsA.Antonin Voyez, T.Tristan Allard, G.Gildas Avoine, P.Pierre Cauchois, E.Elisa Fromont and M.Matthieu Simonin. Attaque par inférence d'appartenance sur des séries temporelles agrégées en utilisant la programmation par contraintes.BDA 2021 - 37ème Conférence sur la Gestion de Données – Principes, Technologies et ApplicationsParis, FranceOctober 2021, 1
  HALback to textback to text
• 29 inproceedingsH.Heng Zhang, E.Elisa Fromont, S.Sébastien Lefevre and B.Bruno Avignon. Deep Active Learning from Multispectral Data Through Cross-Modality Prediction Inconsistency.ICIP 2021 - 28th IEEE International Conference on Image ProcessingAnchorage, United StatesIEEESeptember 2021, 1-5
  HALback to text
• 30 inproceedingsH.Heng Zhang, E.Elisa Fromont, S.Sébastien Lefèvre and B.Bruno Avignon. Guided Attentive Feature Fusion for Multispectral Pedestrian Detection.WACV 2021 - IEEE Winter Conference on Applications of Computer VisionWaikoloa /Virtual, United StatesIEEEJanuary 2021, 1-9
  HALback to text
• 31 inproceedingsH.Heng Zhang, E.Elisa Fromont, S.Sébastien Lefèvre and B.Bruno Avignon. Low-cost Multispectral Scene Analysis with Modality Distillation.IEEE Winter Conference on Applications of Computer Vision (WACV)Waikoloa /Virtual, United StatesJanuary 2022
  HALback to text
• 32 inproceedingsH.Heng Zhang, E.Elisa Fromont, S.Sébastien Lefèvre and B.Bruno Avignon. PDF-Distil: including Prediction Disagreements in Feature-based Distillation for object detection.British Machine Vision Conference (BMVC) 2021 proceedingsBMVC 2021 - 32nd British Machine Vision ConferenceVirtual, United KingdomNovember 2021, 1-13
  HALback to text

Scientific book chapters

• 33 inbookL.Luis Galárraga, O.Olivier Pelgrin and A.Alexandre Termier. HiPaR: Hierarchical Pattern-Aided Regression.12712Advances in Knowledge Discovery and Data MiningLecture Notes in Computer ScienceSpringer International PublishingMay 2021, 320-332
  HALDOIback to text

Doctoral dissertations and habilitation theses

• 34 thesisY.Yichang Wang. Interpretable time series classification.Université Rennes 1September 2021
  HAL
• 35 thesisH.Heng Zhang. Multispectral object detection.Rennes 1December 2021
  HAL

Reports & preprints

• 36 miscS.Samuel Felton, P.Pascal Brault, E.Elisa Fromont and E.Eric Marchand. Visual Servoing in Autoencoder Latent Space: Supplementary material.November 2021
  HAL

Other scientific publications

• 37 articleG.Guillaume Gravier, E.Elisa Fromont, N.Nicolas Courty, T.Teddy Furon, C.Christine Guillemot and P.Paolo Robuffo Giordano. Rennes - une IA souveraine au service de la vie publique.Bulletin de l'Association Française pour l'Intelligence Artificielle2021
  HAL
• 38 inproceedingsT.Thomas Guyet. Génération d'un SNDS synthétique à partir de données ouvertes.EGC 2021 - Actes de la conférence Extraction et Gestion des connaissancesMontpellier, FranceJanuary 2021, 1-2
  HAL