2023Activity reportProject-TeamSCOOL

7.1.1 rlberry

• Keywords:
  Reinforcement learning, Simulation, Artificial intelligence
• Functional Description:
  rlberry is a reinforcement learning (RL) library in Python for research and education. The library provides implementations of several RL agents for you to use as a starting point or as baselines, provides a set of benchmark environments, very useful to debug and challenge your algorithms, handles all random seeds for you, ensuring reproducibility of your results, and is fully compatible with several commonly used RL libraries like OpenAI gym and Stable Baselines.
• URL:
  https://­github.­com/­rlberry-py/­rlberry
• Contact:
  Timothee Mathieu

7.1.2 gym-DSSAT

• Keywords:
  Reinforcement learning, Crop management, Sequential decision making under uncertainty, Mechanistic modeling
• Functional Description:
  gym-DSSAT let you (learn to) manage a crop parcel, from seed selection, to daily activity in the field, to harvesting.
• URL:
  https://­gitlab.­inria.­fr/­rgautron/­gym_dssat_pdi
• Contact:
  Romain Gautron
• Partners:
  CIRAD, Cgiar

7.1.3 Weight Trajectory Predictor : algorithm

• Name:
  Weight Trajectory Predictor : algorithm
• Keywords:
  Medical applications, Machine learning
• Scientific Description:
  We performed a retrospective study of clinical data collected prospectively on patients with up to five years postoperative follow-up (ABOS cohort, CHU Lille) and trained a supervised model to predict the relative total weight loss (“%TWL”) of a patient 1, 3, 12, 24 and 60 months after surgery. This model consists in a decision tree, written in python, taking as input a selected subset of preoperative attributes (weight, height, type of intervention, age, presence or absence of type 2 diabetes or impaired glucose tolerance, diabetes duration, smoking habits) and returns an estimation of %TWL as well as a prediction interval based on the interquartile range of %TWL observed on similar patients. The predictions of this tool have been validated both internally and externally (on French and Dutch cohorts).
• Functional Description:

  The “Weight Tracjectory Predictor” algorithm is part of a larger project, whose goal is to leverage artificial intelligence techniques to improve patient care. This code is the product of a collaboration between Inria SCOOL and the UMR 1190-EGID team of the CHU Lille. It aims to predict the weight loss trajectory of a patient following bariatric surgery (treatment of severe obesity) from a set of preoperative characteristics.

  We performed a retrospective study of clinical data collected prospectively on patients with up to five years postoperative follow-up (ABOS cohort, CHU Lille) and trained a supervised model to predict the relative total weight loss (“%TWL”) of a patient 1, 3, 12, 24 and 60 months after surgery. This model consists in a decision tree, written in python, taking as input a selected subset of preoperative attributes (weight, height, type of intervention, age, presence or absence of type 2 diabetes or impaired glucose tolerance, diabetes duration, smoking habits) and returns an estimation of %TWL as well as a prediction interval based on the interquartile range of %TWL observed on similar patients. The predictions of this tool have been validated both internally and externally (on French and Dutch cohorts).

  The goal of this software is to improve patient follow-up after bariatric surgery: - during preoperative visits, by providing clinicians with a quantitative tool to inform the patient regarding potential weight loss outcome. - during postoperative control visits, by comparing the predicted and realized weight trajectories, which may facilitate early detection of complications.

  This software component will be embedded in a web app for ease of use.

• Release Contributions:
  Initial version
• URL:
  https://­bariatric-weight-trajectory-prediction.­univ-lille.­fr/
• Contact:
  Julien Teigny
• Participants:
  Pierre Bauvin, Francois Pattou, Philippe Preux, Violeta Raverdy, Patrick Saux, Tomy Soumphonphakdy, Julien Teigny, Hélène Verkindt
• Partner:
  CHU de Lille

7.1.4 Adastop

• Keywords:
  Hypothesis testing, Reinforcement learning, Reproducibility
• Functional Description:
  This package contains the AdaStop algorithm. AdaStop implements a statistical test to adaptively choose the number of runs of stochastic algorithms necessary to compare these algorithms and be able to rank them with a theoretically controlled family-wise error rate. One particular application for which AdaStop was created is to compare Reinforcement Learning algorithms. Please note, that what we call here an algorithm is really a certain implementation of an algorithm.
• URL:
  https://­github.­com/­TimotheeMathieu/­adastop
• Contact:
  Timothee Mathieu

7.1.5 average-reward-reinforcement-learning

• Keywords:
  Mutli-armed bandits, Reinforcement learning, Python
• Functional Description:
  Library of RL and Bandit algorithms.
• URL:
  https://­gitlab.­inria.­fr/­omaillar/­average-reward-reinforcement-learning
• Contact:
  Odalric-Ambrym Maillard
• Participant:
  Odalric-Ambrym Maillard

7.1.6 FarmGym

• Name:
  Farming Environment Gym factory for Reinforcement Learning
• Keywords:
  Reinforcement learning, Simulator, Agroecology
• Functional Description:
  Farming Environment Gym factory for Reinforcement Learning
• Release Contributions:
  This version is an entire rewriting by Odlaric-Ambrym Maillard of the prototype V1 created by Thomas Carta. Version V2 features modular creation of farms, specifications of various entities and monitoring facilities. Recent additions include slight adjustment of the entities dynamics. It also features unit tests and an automatic generation of base policies done by Brahim Driss.
• News of the Year:

  Nov 2022: Déploiement de FarmGym sur gitlab, auparavant en développement interne par Odalric-Ambrym Maillard (2021-2022). Nov-Dec 2022: Organisation de compétition interne par Timothée Mathieu.

  Printemps 2023: Arrivée de Brahim Driss sur le projet, pour mettre en place tests unitaires/fonctionnels et assister Odalric-Ambrym Maillard dans le développement de fonctionnalité.

• URL:
  https://­github.­com/­farm-gym/
• Publication:
  hal-03960683
• Contact:
  Odalric-Ambrym Maillard
• Participants:
  Odalric-Ambrym Maillard, Brahim Driss, Timothee Mathieu
• Partner:
  Inria

8.4.1 Methodology

AdaStop: sequential testing for efficient and reliable comparisons of Deep RL Agents, 45

The reproducibility of many experimental results in Deep Reinforcement Learning (RL) is under question. To solve this reproducibility crisis, we propose a theoretically sound methodology to compare multiple Deep RL algorithms. The performance of one execution of a Deep RL algorithm is random so that independent executions are needed to assess it precisely. When comparing several RL algorithms, a major question is how many executions must be made and how can we assure that the results of such a comparison is theoretically sound. Researchers in Deep RL often use less than 5 independent executions to compare algorithms: we claim that this is not enough in general. Moreover, when comparing several algorithms at once, the error of each comparison accumulates and must be taken into account with a multiple tests procedure to preserve low error guarantees. To address this problem in a statistically sound way, we introduce AdaStop, a new statistical test based on multiple group sequential tests. When comparing algorithms, AdaStop adapts the number of executions to stop as early as possible while ensuring that we have enough information to distinguish algorithms that perform better than the others in a statistical significant way. We prove both theoretically and empirically that AdaStop has a low probability of making an error (Family-Wise Error). Finally, we illustrate the effectiveness of AdaStop in multiple use-cases, including toy examples and difficult cases such as Mujoco environments.

8.4.2 Digital health

Impact of Robotic Assistance on Complications in Bariatric Surgery at Expert Laparoscopic Surgery Centers: A Retrospective Comparative Study With Propensity Score, 12

Objective: To investigate the way robotic assistance affected rate of complications in bariatric surgery at expert robotic and laparoscopic surgery facilities. Background: While the benefits of robotic assistance were established at the beginning of surgical training, there is limited data on the robot's influence on experienced bariatric laparoscopic surgeons. Methods: We conducted a retrospective study using the BRO clinical database (2008–2022) collecting data of patients operated on in expert centers. We compared the serious complication rate (defined as a Clavien score$\ge$3) in patients undergoing metabolic bariatric surgery with or without robotic assistance. We used a directed acyclic graph to identify the variables adjustment set used in a multivariable linear regression, and a propensity score matching to calculate the average treatment effect (ATE) of robotic assistance. Results: The study included 35,043 patients [24,428 sleeve gastrectomy (SG); 10,452 Roux-en-Y gastric bypass (RYGB); 163 single anastomosis duodenal-ileal bypass with sleeve gastrectomy (SADI-S)], with 938 operated on with robotic assistance (801 SG; 134 RYGB; 3 SADI-S), among 142 centers. Overall, we found no benefit of robotic assistance regarding the risk of complications (average treatment effect=-0.05, P =0.794), with no difference in the RYGB+SADI group (P =0.322) but a negative trend in the SG group (more complications, P =0.060). Length of hospital stay was decreased in the robot group ($3.7\pm 11.1$ vs $4.0\pm 9.0$ days, $P<0.001$). Conclusions: Robotic assistance reduced the length of stay but did not statistically significantly reduce postoperative complications (Clavien score$\ge 3$) following either GBP or SG. A tendency toward an elevated risk of complications following SG requires more supporting studies.

Development and validation of an interpretable machine learning-based calculator for predicting 5-year weight trajectories after bariatric surgery: a multinational retrospective cohort SOPHIA study, 14

Background Weight loss trajectories after bariatric surgery vary widely between individuals, and predicting weight loss before the operation remains challenging. We aimed to develop a model using machine learning to provide individual preoperative prediction of 5-year weight loss trajectories after surgery. Methods In this multinational retrospective observational study we enrolled adult participants (aged $\ge$18 years) from ten prospective cohorts (including ABOS [NCT01129297], BAREVAL [NCT02310178], the Swedish Obese Subjects study, and a large cohort from the Dutch Obesity Clinic [Nederlandse Obesitas Kliniek]) and two randomised trials (SleevePass [NCT00793143] and SM-BOSS [NCT00356213]) in Europe, the Americas, and Asia, with a 5 year followup after Roux-en-Y gastric bypass, sleeve gastrectomy, or gastric band. Patients with a previous history of bariatric surgery or large delays between scheduled and actual visits were excluded. The training cohort comprised patients from two centres in France (ABOS and BAREVAL). The primary outcome was BMI at 5 years. A model was developed using least absolute shrinkage and selection operator to select variables and the classification and regression trees algorithm to build interpretable regression trees. The performances of the model were assessed through the median absolute deviation (MAD) and root mean squared error (RMSE) of BMI. Findings10 231 patients from 12 centres in ten countries were included in the analysis, corresponding to 30 602 patient-years. Among participants in all 12 cohorts, 7701 (75.3%) were female, 2530 (24.7%) were male. Among 434 baseline attributes available in the training cohort, seven variables were selected: height, weight, intervention type, age, diabetes status, diabetes duration, and smoking status. At 5 years, across external testing cohorts the overall mean MAD BMI was 2.8 kg/m${}^2$ (95% CI 2.6-3.0) and mean RMSE BMI was 4.7 kg/m${}^2$ (4.4-5.0), and the mean difference between predicted and observed BMI was -0.3 kg/m${}^2$ (SD 4.7). This model is incorporated in an easy to use and interpretable web-based prediction tool to help inform clinical decision before surgery. InterpretationWe developed a machine learning-based model, which is internationally validated, for predicting individual 5-year weight loss trajectories after three common bariatric interventions.

Elbow trauma in children: development and evaluation of radiological artificial intelligence models, 13

Rationale and Objectives: To develop a model using artificial intelligence (A.I.) able to detect post-traumatic injuries on pediatric elbow X-rays then to evaluate its performances in silico and its impact on radiologists' interpretation in clinical practice. Material and Methods: A total of 1956 pediatric elbow radiographs performed following a trauma were retrospectively collected from 935 patients aged between 0 and 18 years. Deep convolutional neural networks were trained on these X-rays. The two best models were selected then evaluated on an external test set involving 120 patients, whose X-rays were performed on a different radiological equipment in another time period. Eight radiologists interpreted this external test set without then with the help of the A.I. models. Results: Two models stood out: model 1 had an accuracy of 95.8% and an AUROC of 0.983 and model 2 had an accuracy of 90.5% and an AUROC of 0.975. On the external test set, model 1 kept a good accuracy of 82.5% and AUROC of 0.916 while model 2 had a loss of accuracy down to 69.2% and of AUROC to 0.793. Model 1 significantly improved radiologist's sensitivity (0.82 to 0.88, P = 0.016) and accuracy (0.86 to 0.88, P = 0,047) while model 2 significantly decreased specificity of readers (0.86 to 0.83, P = 0.031). Conclusion: End-to-end development of a deep learning model to assess post-traumatic injuries on elbow Xray in children was feasible and showed that models with close metrics in silico can unpredictably lead radiologists to either improve or lower their performances in clinical settings.

8.4.3 Sailboat digital twin

Reinforcement-learning robotic sailboats: simulator and preliminary results, 33

This work focuses on the main challenges and problems in developing a virtual oceanic environment reproducing real experiments using Unmanned Surface Vehicles (USV) digital twins. We introduce the key features for building virtual worlds, considering using Reinforcement Learning (RL) agents for autonomous navigation and control. With this in mind, the main problems concern the definition of the simulation equations (physics and mathematics), their effective implementation, and how to include strategies for simulated control and perception (sensors) to be used with RL. We present the modeling, implementation steps, and challenges required to create a functional digital twin based on a real robotic sailing vessel. The application is immediate for developing navigation algorithms based on RL to be applied on real boats.

General System Architecture and COTS Prototyping of an AIoT-Enabled Sailboat for Autonomous Aquatic Ecosystem Monitoring, 11

Unmanned vehicles keep growing attention as they facilitate innovative commercial and civil applications within the Internet of Things (IoT) realm. In this context, autonomous sailing boats are becoming important marine platforms for performing different tasks, such as surveillance, water, and environmental monitoring. Most of these tasks heavily depend on artificial intelligence (AI) technologies, such as visual navigation and path planning, and comprise the so-called Artificial Intelligence of Things (AIoT). In this paper, we propose (i) the OpenBoat, an automating system architecture for AIoTenabled sailboats with application-agnostic autonomous environment monitoring capability, and (ii) the F-Boat, a fully functional prototype of OpenBoat built with Commercial Off-The-Shelf (COTS) components on a real sailboat. F-Boat includes low-level control strategies for autonomous path following, communication infrastructure for remote operation and cooperation with other systems, edge computing with AI accelerator, modular support for application-specific monitoring systems, and navigation aspects. F-Boat is also designed and built for robustness situations to guarantee its operation under extreme events, such as high temperatures and bad weather, through extended periods of time. We show the results of field experiments running in Guanabara Bay, an important aquatic ecosystem in Brazil, that demonstrate the functionalities of the prototype and demonstrate the AIoT capability of the proposed architecture.

8.4.4 Computer-assisted mathematics

A Formalization of Doob's Martingale Convergence Theorems in mathlib, 34

We present the formalization of Doob's martingale convergence theorems in the mathlib library for the Lean theorem prover. These theorems give conditions under which (sub)martingales converge, almost everywhere or in $L^1$. In order to formalize those results, we build a definition of the conditional expectation in Banach spaces and develop the theory of stochastic processes, stopping times and martingales. As an application of the convergence theorems, we also present the formalization of Lévy's generalized Borel-Cantelli lemma. This work on martingale theory is one of the first developments of probability theory in mathlib, and it builds upon diverse parts of that library such as topology, analysis and most importantly measure theory.

8.4.5 Interpretability

Optimal Interpretability-Performance Trade-off of Classification Trees with Black-Box Reinforcement Learning, 44

Interpretability of AI models allows for user safety checks to build trust in these models. In particular, decision trees (DTs) provide a global view on the learned model and clearly outlines the role of the features that are critical to classify a given data. However, interpretability is hindered if the DT is too large. To learn compact trees, a Reinforcement Learning (RL) framework has been recently proposed to explore the space of DTs. A given supervised classification task is modeled as a Markov decision problem (MDP) and then augmented with additional actions that gather information about the features, equivalent to building a DT. By appropriately penalizing these actions, the RL agent learns to optimally trade-off size and performance of a DT. However, to do so, this RL agent has to solve a partially observable MDP. The main contribution of this paper is to prove that it is sufficient to solve a fully observable problem to learn a DT optimizing the interpretability-performance trade-off. As such any planning or RL algorithm can be used. We demonstrate the effectiveness of this approach on a set of classical supervised classification datasets and compare our approach with other interpretability-performance optimizing methods.

“How Biased are Your Features?": Computing Fairness Influence Functions with Global Sensitivity Analysis, 24

Fairness in machine learning has attained significant focus due to the widespread application in high-stake decision-making tasks. Unregulated machine learning classifiers can exhibit bias towards certain demographic groups in data, thus the quantification and mitigation of classifier bias is a central concern in fairness in machine learning. In this paper, we aim to quantify the influence of different features in a dataset on the bias of a classifier. To do this, we introduce the Fairness Influence Function (FIF). This function breaks down bias into its components among individual features and the intersection of multiple features. The key idea is to represent existing group fairness metrics as the difference of the scaled conditional variances in the classifier’s prediction and apply a decomposition of variance according to global sensitivity analysis. To estimate FIFs, we instantiate an algorithm that applies variance decomposition of classifier’s prediction following local regression. Experiments demonstrate that captures FIFs of individual feature and intersectional features, provides a better approximation of bias based on FIFs, demonstrates higher correlation of FIFs with fairness interventions, and detects changes in bias due to fairness affirmative/punitive actions in the classifier. The code is available online.

8.4.6 Privacy

From Noisy Fixed-Point Iterations to Private ADMM for Centralized and Federated Learning, 22

We study differentially private (DP) machine learning algorithms as instances of noisy fixed-point iterations, in order to derive privacy and utility results from this well-studied framework. We show that this new perspective recovers popular private gradient-based methods like DP-SGD and provides a principled way to design and analyze new private optimization algorithms in a flexible manner. Focusing on the widely-used Alternating Directions Method of Multipliers (ADMM) method, we use our general framework to derive novel private ADMM algorithms for centralized, federated and fully decentralized learning. For these three algorithms, we establish strong privacy guarantees leveraging privacy amplification by iteration and by subsampling. Finally, we provide utility guarantees using a unified analysis that exploits a recent linear convergence result for noisy fixed-point iterations.

Marich: A Query-efficient Distributionally Equivalent Model Extraction Attack using Public Data, 28

We study design of black-box model extraction attacks that can send minimal number of queries from a publicly available dataset to a target ML model through a predictive API with an aim to create an informative and distributionally equivalent replica of the target. First, we define distributionally equivalent and Max-Information model extraction attacks, and reduce them into a variational optimisation problem. The attacker sequentially solves this optimisation problem to select the most informative queries that simultaneously maximise the entropy and reduce the mismatch between the target and the stolen models. This leads to an active sampling-based query selection algorithm, Marich, which is model-oblivious. Then, we evaluate Marich on different text and image data sets, and different models, including CNNs and BERT. Marich extracts models that achieve $\sim 60-95\%$ of true model's accuracy and uses $\sim 1,000-8,500$ queries from the publicly available datasets, which are different from the private training datasets. Models extracted by Marich yield prediction distributions, which are $\sim 2-4\times$ closer to the target's distribution in comparison to the existing active sampling-based attacks. The extracted models also lead to $84-96\%$ accuracy under membership inference attacks. Experimental results validate that Marich is query-efficient, and capable of performing task-accurate, high-fidelity, and informative model extraction.

10.1.1 Inria associate team not involved in an IIL or an international program

10.1.2 Participation in other International Programs

International collaborators:

• C. Dimitrakakis, Professor, Université de Neuchâtel, safe reinforcement learning, Bayesian reinforcement learning, Online learning in games.
• M. Alibeigi, Researcher, Zenseact AB of Volvo, safe reinforcement learning.
• B. Ghosh, Research Scientist, ASTAR Singapore, fairness in machine learning.
• K. Meel, Associate Professor, National University of Singapore and University of Toronto, Formal methods for machine learning.
• S. Bressan, Associate Professor, National University of Singapore, Machine learning for quantum physics.

10.2.1 Visits of international scientists

10.2.2 Visits to international teams

10.3.1 Other european programs/initiatives

• Title:
  CausalXRL
• Duration:
  2021 $\to$ 2024
• Coordinator:
  Aditya Gilra, U. Amsterdam
• Partners:
  • U. Amsterdam
  • U. Sheffield
  • U. Vienna
  • Inria Scool
• Inria contact:
  Philippe Preux
• Summary:
  Deep reinforcement learning systems are approaching or surpassing human-level performance in specific domains, from games to decision support to continuous control, albeit in non-critical environments. Most of these systems require random exploration and state-action-value-based exploitation of the environment. However, in important real-life domains, like medical decision support or patient rehabilitation, every decision or action must be fully justified and certainly not random. We propose to develop neural networks that learn causal models of the environment relating action to effect, initially using offline data. The models will then be interfaced with reinforcement learning and decision support networks, so that every action taken online can be explained or justified based on its expected effect. The causal model can then be refined iteratively, enabling to better predict future cascading effects of any action chain. The system, subsequently termed CausalXRL, will only propose actions that can be justified on the basis of beneficial effects. When the immediate benefit is uncertain, the system will propose explorative actions that generate most-probable future benefit. CausalXRL thus supports the user in choosing actions based on specific expected outcomes, rather than as prescribed by a black box.

10.4.1 ANR projects

Scool is involved in 4 ANR projects:

• ANR Bold, headed by V. Perchet (ENS Paris-Saclay, ENSAE), local head: E. Kaufmann, 2019–2023.
• ANR JCJC FATE, PI: R. Degenne, 2023–2026
• ANR JCJC REPUBLIC, PI: D. Basu, 2023–2026
• ANR BIP-UP, partnership: Scool/Inserm (CHU de Lille), PI: Ph. Preux, 2023–2026.

10.4.2 PEPR projects

Scool is involved in 2 PEPR:

• PEPR AI: project FOUNDRY, local head: E. Kaufmann (description below);
• PEPR « Agroécologie et numérique », Pl@ntAgroEco, local head: O.-A. Maillard.

• Title:
  FOUNDRY
• Duration:
  July 2024 $\to$ June 2028
• Coordinator:
  Panayotis Mertikopoulous, Polaris, Univ. Grenoble Alpes
• Partners:
  • POLARIS: a joint research team between the CNRS, Inria, and Univ. Grenoble Alpes.
  • ENS Lyon: faculty from the pure and applied mathematics department of ENS Lyon.
  • Inria FAIRPLAY: a joint team between Criteo, IP Paris (ENSAE and Ecole Polytechnique), and Inria.
  • LTCI: the informations and communications laboratory of Télécom Paris.
  • MILES: the machine intellligence and learning systems of the LAMSADE lab at Paris Dauphine.
  • Inria Scool
• Inria contact:
  Emilie Kaufmann
• Summary:
  From automated hospital admission systems powered by machine learning (ML), to flexible chatbots capable of fluent conversations and self-driving cars, the wildfire spread of artificial intelligence (AI) has brought to the forefront a crucial question with far-reaching ramifications for the society at large: Can ML systems and models be relied upon to provide trustworthy output in high-stakes, mission- critical environments? These questions invariably revolve around the notion of robustness, an operational desideratum that has eluded the field since its nascent stages. One of the main reasons for this is the fact that ML models and systems are typically data-hungry and highly sensitive to their training input, so they tend to be brittle, narrow-scoped, and unable to adapt to situations that go beyond their training envelope. On that account, the core vision of the proposed research is that robustness cannot be achieved by blindly throwing more data and computing power to larger and larger models with exponentially growing energy requirements (and a commensurate carbon footprint to boot). Instead, our proposal intends to rethink and develop the core theoretical and methodological FOUNDations of Robustness and reliabilitY (FOUNDRY) that are needed to build and instill trust in ML-powered technologies and systems from the ground up.

• Title:
  Pl@ntAgroEco
• Duration:
  July 2024 $\to$ June 2028
• Coordinator:
  Alexis Joly, Inria Zenith, and Pierre Bonnet CIRAD, AMAP.
• Partners:
  • INRAE
  • INRIA
  • IRD
  • CIRAD
  • Tela Botanica
  • Université de Montpellier
  • Université Paris-Saclay
• Inria contact:
  Odalric-Ambrym Maillard
• Summary:

  Agroecology necessarily involves crop diversification, but also the early detection of diseases, deficiencies and stresses (hydric, etc.), as well as better management of biodiversity. The main stumbling block is that this paradigm shift in agricultural practices requires expert skills in botany, plant pathology and ecology that are not generally available to those working in the field, such as farmers or agri-food technicians. Digital technologies, and artificial intelligence in particular, can play a crucial role in overcoming this barrier to access to knowledge.

  The aim of the Pl@ntAgroEco project will be to design, experiment with and develop new high-impact agro-ecology services within the Pl@ntNet platform. This includes :

  • research in AI and plant sciences ;
  • agile development of new components within the platform;
  • organization of participatory science programs and animation of the Pl@ntNet user community.

  Ce programme de travail a pour but de produire une amélioration de la détection et reconnaissance des maladies végétales, de l’identification des niveaux infraspécifiques. Il permettra le développement d’outils d’estimation de la sévérité des symptômes, carences, stades de déclin et stress hydrique ou de caractérisation des associations d’espèces à partir d’images multi-spécimens. Il améliorera la connaissance des espèces.

  Le projet Pl@ntAgroEco rassemble des forces complémentaires en matière de recherche, de développement et d'animation. S’ajouteront à l'équipe pluridisciplinaire chargée de la plateforme Pl@ntNet de nouvelles forces de recherche ayant une expertise reconnue dans les sciences participatives. Le consortium rassemblera 10 partenaires incluant des organismes de recherche, des universités, des acteurs de la société civile et des partenaires internationaux

10.4.3 Inria challenge

Scool has been involved in the Challenge HY_AIAI. In this challenge, we collaborated with L. Gallaraga, CR Inria Rennes, about the combination of statistical and symbolic approaches in machine learning.

10.4.4 Other projects in France

Scool is involved in the Regalia pilot-project.

Other collaborations:

• A. Bellet, CR, Inria Lille-Nord Europe (Équipe Magnet).
• B. De-Saporta, Université de Montpellier, piecewise-deterministic Markov processes.
• A. Garivier, Professor, ENS Lyon, PhD co-supervisor of Aymen Al-Marjani.
• R. Gautron, Post-doctoral researcher, CGIAR, agricultural practices recommendation, gym-DSSAT software.
• B. Raffin, DR Inria Grenoble, continuous-time reinforcement learning.
• L. Soulier, Associate Professor, Sorbonne Université, reinforcement learning for information retrieval.
• L. Richert, R. Thiébaut, Inria SISTM, Bordeaux, bandits for vaccine clinical trials.
• A. Tirinzoni, Meta AI, instance-dependent sample complexity of reinforcement learning.

11.1.1 Scientific events: organisation

11.1.2 Scientific events: selection

11.1.3 Journal

11.1.4 Invited talks

• D. Basu: invited speaker, ACMU seminars, Indian Statistical Institute (Kolkata, India).
• D. Basu: invited speaker, Journée Security at CRIStAL (Lille, France).
• D. Basu: invited speaker, JSPS Japan-Singapore Joint Seminars (National Institute of Informatics, Japan).
• D. Basu: invited speaker, Descartes seminars, CNRS@CREATE, National University of Singapore (Singapore).
• R. Degenne: invited talk, Probability and Statistics seminar, University of Freiburg (Freiburg, Germany).
• R. Degenne: invited speaker, Lean for the Curious Mathematician 2023 (Düsseldorf, Germany)
• R. Degenne: invited speaker, Workshop on Bandits and Statistical Tests (Potsdam, Germany).
• E. Kaufmann: invited talk at the Machine Learning Theory Bootcamp in CWI (Amsterdam).
• E. Kaufmann: invited speaker at the European Workshop on Reinforcement Learning (Brussels).
• E. Kaufmann: invited talk at the Statistics seminar of University Paris-Saclay (Orsay).
• E. Kaufmann: invited speaker at the Workshop on Bandits and Statistical Tests (Potsdam).
• O.-A. Maillard, invited speaker, PMSMA: Processus markoviens, semi-markoviens et leurs applications
• T. Mathieu, invited speaker, Conference on Statistical estimation.
• Ph. Preux, invited speaker, Machine Learning in Poland (ML in PL).
• Ph. Preux, keynote speaker, Inria-Chile Scientific days.

11.1.5 Tutorials

• D. Basu gave a tutorial on Auditing Bias of Machine Learning Algorithms: Tools and Overview at IJCAI, 2023.
• D. Basu gave a tutorial on Artificial Intelligence for Agricultural Sciences at Bihar Agricultural University, Sabour.
• É. Kaufmann gave a class on bandits at the Reinforcement Learning Summer Scool, Barcelona.
• O.-A. Maillard gave a class on reinforcement learning at Ecole Polytechnique International Executive Master (May) and Executive Master (October)
• Ph. Preux gave a class on reinforcement learning at the INSA-Rouen Summer Scool.

11.1.6 Scientific expertise

Ph. Preux is:

• a member of the IRD CSS 5 (data science and models),
• a member of the scientific committee of the PEPR « agro-écologie et numérique »,
• a member of the scientific committee on ethics of the Health Data warehouse of the CHU de Lille.

11.1.7 Research administration

Ph. Preux is the scientific head of the CornelIA CPER project.

11.2.1 Teaching

• R. Akrour: « Option découverte: Machine Learning », L3 in Computer Science, Université de Lille
• R. Akrour: « Perception et motricité 2 », L2 MIASHS, Université de Lille
• R. Akrour: « Perception et motricité 1 », L1 MIASHS, Université de Lille
• D. Basu: Sequential Decision Making, M2 in Data Science, Centrale Lille and Université de Lille
• D. Basu: Research Reading Group, M2 in Data Science, Centrale Lille and Université de Lille
• D. Basu: Advanced Machine Learning and Decision Making, Centrale Lille
• R. Degenne: Sequential learning, M2 MVA, ENS Paris-Saclay
• R. Degenne: Sequential learning, Centrale Lille
• R. Degenne: Sciences des données 3, L3 MIASHS, Université de Lille
• E. Kaufmann: Sequential Decision Making (24h), M2 Data Science, Ecole Centrale Lille.
• O.-A. Maillard: Statistical Reinforcement Learning (48h), MAP/INF641, Master Artificial Intelligence and advanced Visual Computing, École Polytechnique.
• Ph. Preux: « Prise de décision séquentielle dans l'incertain », M2 in Computer Science, Université de Lille.
• Ph. Preux: « Apprentissage par renforcement », M2 in Computer Science, Université de Lille.
• Ph. Preux: « Science des données II », L3 MIASHS, Université de Lille.
• Ph. Preux: « IA et apprentissage automatique », DU IA & Santé, Université de Lille.

11.2.2 Supervision

• D. Basu and O.-A. Maillard supervised Thomas Michel's M2 internship.
• D. Basu supervised Priyank Agarwal's PhD internship.
• D. Basu supervised Udvas Das's masters thesis.
• E. Kaufmann and R.Degenne co-supervised the M2 internship of Adrienne Tuynman.
• O.-A. Maillard supervised Waris Radji's M2 internship.
• A. Shilova, Ph. Preux and B. Raffin (Inria Grenoble) co-supervised Thomas Delliaux's M2 internship.

11.2.3 Juries

D. Basu was a member of the Ph.D. defense committee for:

• Paolo Recchia, Université Côte d'Azur

E. Kaufmann was a member of the Ph.D. defense committees for:

• Chloé Rouyer, University of Copenhagen (reviewer), January
• Hannes Errikson, Chalmers University, September
• Kaito Ariu, KTH, November
• Aymen Al-Marjani, ENS Lyon (supervisor), December
• Otmane Sahki, Criteo/ENSAE (reviewer), December

E. Kaufmann was a reviewer of the HDR defense of Raphaël Feraud (Orange Labs, Université Paris-Saclay) who defended in Feburary.

O.-A. Maillard was a member of the Ph.D. defense committees for:

• Réda Ouhamma, Université de Lille (supervisor), April.
• Antoine Barrier, Ecole normale supérieure de Lyon (reviewer), July.
• Fabien Pesquerel, Université de Lille (supervisor), December.

Ph. Preux was a member of the Ph.D. defense committees for:

• Hélène Plisnier, VUB (reviewer), March.
• Nathan Grinsztajn, Université de Lille (supervisor), June.
• Hannes Eriksson, Chalmers University, September.
• Meyssa Zouambi, Université de Lille (chair), December.
• Étienne Ménager, Université de Lille (chair), December.

11.3.1 Articles and contents

O.-A. Maillard was interviewed for an article for Inria Demain, un compagnon d’aide à la décision pour l’agriculture ?

11.3.2 Education

E. Kaufmann did 4 one hour sessions of “CHICHE: Un Scientifique, Une Classe” in the Lycée Kernanec, Marcq-en-Barœul.

Ph. Preux made a presentation on AI for a group of pupils at Collège M. Yourcenar, April, Marchiennes.

11.3.3 Interventions

E. Kaufmann gave a presentation at the “Rencontre des Jeunes Mathématiciennes et Informaticiennes” (RJMI) organized at Inria. A. Tuynman (PhD student in Scool) organized a problem session at RJMI for a group a high school female students.

O.-A. Maillard gave a presentation at table ronde “Decision making in a uncertain environment” organized for all master students of graduate programs of Université de Lille, December.

Ph. Preux:

• was part of the Merlin project of Université de Lille that led to “The big investigation on AI” TV program produced by « L'esprit Sorcier TV » channel.
• was part of the scientific committee of the « Forum des Sciences » in Villeneuve d'Ascq regarding the season on AI.
• gave a talk on AI to the ASAP, Université de Lille: “A short history of AI”.
• gave a talk during the Robotik exhibition in Orchies: “AI: from myths to reality”.
• participated to the TV program “We tell you more” on AI on the Wéo channel.
• made a presentation about AI during a meeting of the Comité d'Orientation Stratégique et Pédagogique (COSP) of the INSPE HdF regarding how AI impacts teaching and how future (primary, elementary, high) school teachers should be trained on AI.

International journals

• 11 articleA. P.André P D Araújo, D.Dickson Daniel, R.Raphael Guerra, D.Diego Brandão, E. C.Eduardo Charles Vasconcellos, A.Alvaro Negreiros, E.Esteban Clua, L.Luiz Goncalves and P.Philippe Preux. General System Architecture and COTS Prototyping of an AIoT-Enabled Sailboat for Autonomous Aquatic Ecosystem Monitoring.IEEE Internet of Things Journal2023HALDOIback to text
• 12 articleR.Robert Caiazzo, P.Pierre Bauvin, C.Camille Marciniak, P.Patrick Saux, G.Geoffrey Jacqmin, R.Raymond Arnoux, S.Salomon Benchetrit, J.Jerome Dargent, J.-M.Jean-Marc Chevallier, V.Vincent Frering, J.Jean Gugenheim, D.David Lechaux, S.Simon Msika, A.Adrien Sterkers, P.Philippe Topart, G.Grégory Baud and F.François Pattou. Impact of Robotic Assistance on Complications in Bariatric Surgery at Expert Laparoscopic Surgery Centers: A Retrospective Comparative Study With Propensity Score.Annals of Surgery2784September 2023, 489-496HALDOIback to text
• 13 articleC.Clémence Rozwag, F.Franck Valentini, A.Anne Cotten, X.Xavier Demondion, P.Philippe Preux and T.Thibaut Jacques. Elbow trauma in children: development and evaluation of radiological artificial intelligence models.Research in Diagnostic and Interventional Imaging6April 2023HALDOIback to text
• 14 articleP.Patrick Saux, P.Pierre Bauvin, V.Violeta Raverdy, J.Julien Teigny, H.Hélène Verkindt, T.Tomy Soumphonphakdy, M.Maxence Debert, A.Anne Jacobs, D.Daan Jacobs, V.Valerie Monpellier, P. C.Phong Ching Lee, C. H.Chin Hong Lim, J. C.Johanna C Andersson-Assarsson, L.Lena Carlsson, P.-A.Per-Arne Svensson, F.Florence Galtier, G.Guelareh Dezfoulian, M.Mihaela Moldovanu, S.Severine Andrieux, J.Julien Couster, M.Marie Lepage, E.Erminia Lembo, O.Ornella Verrastro, M.Maud Robert, P.Paulina Salminen, G.Geltrude Mingrone, R.Ralph Peterli, R. V.Ricardo V Cohen, C.Carlos Zerrweck, D.David Nocca, C. W.Carel W Le Roux, R.Robert Caiazzo, P.Philippe Preux and F.François Pattou. Development and validation of an interpretable machine learning-based calculator for predicting 5-year weight trajectories after bariatric surgery: a multinational retrospective cohort SOPHIA study.The Lancet Digital HealthAugust 2023HALDOIback to textback to text

International peer-reviewed conferences

• 15 inproceedingsA.Aymen Al-Marjani, A.Andrea Tirinzoni and E.Emilie Kaufmann. Active Coverage for PAC Reinforcement Learning.Proceedings of Machine Learning Research (PMLR)Conference on Learning Theory 2023Bangalore, India2023HALback to text
• 16 inproceedingsA.Achraf Azize and D.Debabrota Basu. Interactive and Concentrated Differential Privacy for Bandits.EWRL 2023 – European Workshop on Reinforcement LearningBrussels (Belgium), BelgiumSeptember 2023HALback to text
• 17 inproceedingsA.Achraf Azize, M.Marc Jourdan, A. A.Aymen Al Marjani and D.Debabrota Basu. On the Complexity of Differentially Private Best-Arm Identification with Fixed Confidence.NeurIPS 2023 – Conference on Neural Information Processing SystemsNew Orleans (US), United StatesDecember 2023HALback to text
• 18 inproceedingsD.Dorian Baudry, F.Fabien Pesquerel, R.Rémy Degenne and O.-A.Odalric-Ambrym Maillard. Fast Asymptotically Optimal Algorithms for Non-Parametric Stochastic Bandits.Advances in Neural Information Processing Systems 36 (NeurIPS 2023)NeurIPS 2023 - Thirty-seventh Conference on Neural Information Processing SystemsNew Orleans (Louisiana), United StatesDecember 2023HALback to text
• 19 inproceedingsE.Emil Carlsson, D.Debabrota Basu, F. D.Fredrik D. Johansson and D.Devdatt Dubhashi. Pure Exploration in Bandits with Linear Constraints.EWRL 2023 – European Workshop on Reinforcement LearningBrussels, BelgiumSeptember 2023HALback to text
• 20 inproceedingsM.Matheus Centa and P.Philippe Preux. Soft Action Priors: Towards Robust Policy Transfer.AAAI 2023 - Thirty-Seventh AAAI Conference on Artificial IntelligenceWashington DC, United StatesFebruary 2023HALback to text
• 21 inproceedingsS. R.Sayak Ray Chowdhury, P.Patrick Saux, O.-A.Odalric-Ambrym Maillard and A.Aditya Gopalan. Bregman Deviations of Generic Exponential Families.Conference On Learning Theory (COLT)Bangalore, IndiaJuly 2023HALback to text
• 22 inproceedingsE.Edwige Cyffers, A.Aurélien Bellet and D.Debabrota Basu. From Noisy Fixed-Point Iterations to Private ADMM for Centralized and Federated Learning.Proceedings of the 40th International Conference on Machine Learning (ICML)Honolulu, United StatesJuly 2023HALback to text
• 23 inproceedingsR.Rémy Degenne. On the Existence of a Complexity in Fixed Budget Bandit Identification.Proceedings of Machine Learning ResearchThirty Sixth Conference on Learning Theory195Bengaluru (Bangalore), IndiaJune 2023HALback to text
• 24 inproceedingsB.Bishwamittra Ghosh, D.Debabrota Basu and K.Kuldeep Meel. "How Biased are Your Features?": Computing Fairness Influence Functions with Global Sensitivity Analysis.FAccT '23: the 2023 ACM Conference on Fairness, Accountability, and TransparencyChicago IL, United StatesACMJune 2023, 138-148HALDOIback to text
• 25 inproceedingsM.Marc Jourdan, R.Rémy Degenne and E.Emilie Kaufmann. An ε-Best-Arm Identification Algorithm for Fixed-Confidence and Beyond.Advances in Neural Information Processing Systems (NeurIPS)New Orleans, United StatesDecember 2023HALback to text
• 26 inproceedingsM.Marc Jourdan, R.Rémy Degenne and E.Emilie Kaufmann. Dealing with Unknown Variances in Best-Arm Identification.Proceedings of Machine Learning Research (PMLR)Algorithmic Learning Theory (ALT)Singapore (SG), SingaporeFebruary 2023HALback to text
• 27 inproceedingsM.Marc Jourdan and R.Rémy Degenne. Non-Asymptotic Analysis of a UCB-based Top Two Algorithm.Advances in Neural Information Processing Systems 36 (NeurIPS 2023)Thirty-seventh Conference on Neural Information Processing SystemsNew Orleans (Louisiana), United StatesDecember 2023HALback to text
• 28 inproceedingsP.Pratik Karmakar and D.Debabrota Basu. Marich: A Query-efficient Distributionally Equivalent Model Extraction Attack using Public Data.Advances in Neural Information Processing Systems (NeurIPS)New orleans, USA, United StatesDecember 2023HALback to text
• 29 inproceedingsC.Cyrille Kone, E.Emilie Kaufmann and L.Laura Richert. Adaptive Algorithms for Relaxed Pareto Set Identification.NeurIPS 2023 - 37th Conference on Neural Information Processing SystemsLa Nouvelle Orléans, LA, United StatesDecember 2023HALback to text
• 30 inproceedingsR.Reda Ouhamma, D.Debabrota Basu and O.-A.Odalric-Ambrym Maillard. Bilinear Exponential Family of MDPs: Frequentist Regret Bound with Tractable Exploration & Planning.Proceedings of the AAAI Conference on Artificial Intelligence37Proceedings of the AAAI Conference on Artificial Intelligence, Volume 38Washignton DC, United StatesJune 2023, 9336-9344HALDOIback to text
• 31 inproceedingsP.Patrick Saux and O.-A.Odalric-Ambrym Maillard. Risk-aware linear bandits with convex loss.International Conference on Artificial Intelligence and Statistics (AISTATS)PMLR206Valencia, SpainApril 2023HALback to text
• 32 inproceedingsA.Andrea Tirinzoni, A.Aymen Al-Marjani and E.Emilie Kaufmann. Optimistic PAC Reinforcement Learning: the Instance-Dependent View.Proceedings of Machine Learning Research (PMLR)Algorithmic Learning Theory (ALT)Singapore (SG), SingaporeFebruary 2023HALback to text
• 33 inproceedingsE. C.Eduardo Charles Vasconcellos, R. M.Ronald M Sampaio, A. P.André P D Araújo, E. W.Esteban Walter Gonzales Clua, P.Philippe Preux, R.Raphael Guerra, L. M.Luiz M G Gonçalves, L.Luis Martí, H.Hernan Lira and N.Nayat Sanchez-Pi. Reinforcement-learning robotic sailboats: simulator and preliminary results.NeurIPS 2023 Workshop on Robot Learning Workshop: Pretraining, Fine-Tuning, and Generalization with Large Scale ModelsNew Orelans, United StatesDecember 2023HALback to text
• 34 inproceedingsK.Kexing Ying and R.Rémy Degenne. A Formalization of Doob's Martingale Convergence Theorems in mathlib.Proceedings of the 12th ACM SIGPLAN International Conference on Certified Programs and Proofs12th ACM SIGPLAN International Conference on Certified Programs and ProofsBoston (Massachusetts), United StatesJanuary 2023HALDOIback to text

Conferences without proceedings

• 35 inproceedingsR.Romain Gautron, E. J.Emilio J Padrón, P.Philippe Preux, J.Julien Bigot, O.-A.Odalric-Ambrym Maillard, G.Gerrit Hoogenboom and J.Julien Teigny. Learning crop management by reinforcement: gym-DSSAT.AIAFS 2023 - 2nd AAAI Workshop on AI for Agriculture and Food SystemsWashignton DC, United StatesFebruary 2023HALback to text
• 36 inproceedingsO.-A.Odalric-Ambrym Maillard, T.Timothée Mathieu and D.Debabrota Basu. Farm-gym: A modular reinforcement learning platform for stochastic agronomic games.AIAFS 2023 - Artificial Intelligence for Agriculture and Food SystemsWahington DC, United StatesFebruary 2023HALback to text

Doctoral dissertations and habilitation theses

• 37 thesisN.Nathan Grinsztajn. Reinforcement learning for combinatorial optimization : leveraging uncertainty, structure and priors.Université de LilleJune 2023HAL
• 38 thesisR.Reda Ouhamma. Toward realistic reinforcement learning.Université de LilleApril 2023HAL
• 39 thesisF.Fabien Pesquerel. Information per unit of interaction in stochastic sequential decision making.Université de LilleDecember 2023HAL

Reports & preprints

• 40 miscS.Shubhada Agrawal, T.Timothée Mathieu, D.Debabrota Basu and O.-A.Odalric-Ambrym Maillard. CRIMED: Lower and Upper Bounds on Regret for Bandits with Unbounded Stochastic Corruption.September 2023HALback to text
• 41 miscA.Aymen Al-Marjani, A.Andrea Tirinzoni and E.Emilie Kaufmann. Towards Instance-Optimality in Online PAC Reinforcement Learning.October 2023HAL
• 42 miscR.Riccardo Della Vecchia and D.Debabrota Basu. Online Instrumental Variable Regression: Regret Analysis and Bandit Feedback.February 2023HALback to text
• 43 miscH.Hannes Eriksson, D.Debabrota Basu, T.Tommy Tram, M.Mina Alibeigi and C.Christos Dimitrakakis. Reinforcement Learning in the Wild with Maximum Likelihood-based Model Transfer.February 2023HALback to text
• 44 reportH.Hector Kohler, R.Riad Akrour and P.Philippe Preux. Optimal Interpretability-Performance Trade-off of Classification Trees with Black-Box Reinforcement Learning.RR-9503Inria Lille Nord Europe - Laboratoire CRIStAL - Université de LilleApril 2023HALback to text
• 45 reportT.Timothée Mathieu, R.Riccardo Della Vecchia, A.Alena Shilova, M.Matheus Centa de Medeiros, H.Hector Kohler, O.-A.Odalric-Ambrym Maillard and P.Philippe Preux. AdaStop: sequential testing for efficient and reliable comparisons of Deep RL Agents.RR-9513Inria Lille Nord Europe - Laboratoire CRIStAL - Université de LilleJune 2023HALback to text
• 46 reportA.Alena Shilova, T.Thomas Delliaux, P.Philippe Preux and B.Bruno Raffin. Learning HJB Viscosity Solutions with PINNs for Continuous-Time Reinforcement Learning.RR-9541Inria Lille - Nord Europe, CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189; Univ. Lille, CNRS, Centrale Lille, Inria UMR 9189 - CRIStAL,INRIA Lille Nord Europe, Villeneuve d’Ascq, France; Univ. Grenoble Alps, CNRS, Inria, Grenoble INP, LIG, 38000 Grenoble, FranceFebruary 2024, 1-30HAL