Portail HAL des publications du LIRMM
Not a member yet
13279 research outputs found
Sort by
Robotic co-manipulation of deformable linear objects for large deformation tasks
No full textInternational audienceThis research addresses the challenge of large/complex deformation in the shape control tasks of Deformable Linear Objects (DLO). We propose a collaborative approach using two manipulators to achieve shape control of a DLO in 2D workspace. The proposed methodology introduces an innovative Intermediary Shapes Generation (ISG) algorithm which outputs a series of intermediary shapes to guide the DLO towards the desired shape. The robot controller is formulated as an optimization problem, where the main objective is to minimize the error between the current shape and the desired shape, while ensuring the diminishing rigidity property of the DLO as a constraint. We conduct extensive simulations and real-life experiments to evaluate the effectiveness of our approach. We consider various scenarios of basic shapes, as well as complex deformations with opposite concavities between initial and final shapes. The outcomes demonstrate the robustness and high accuracy of the proposed system, in achieving complex deformations. This capability represents the primary contribution of our research. The optimization-based control framework, coupled with the ISG algorithm, enables effective shape control without the need for extensive modeling nor training, and offers a promising solution for practical applications requiring precise shape control of DLOs. Moreover, we carry out a thorough review and comparative analysis encompassing the latest literature in DLO shape control, and the techniques for DLO modeling
Using Dynamic Laser Speckle Imaging for Plant Breeding: A Case Study of Water Stress in Sunflowers
No full textInternational audienceThis study focuses on the promising use of biospeckle technology to detect water stress in plants, a complex physiological mechanism. This involves monitoring the temporal activity of biospeckle pattern to study the occurrence of stress within the leaf. The effects of water stress in plants can involve physical and biochemical changes. Some of these changes may alter the optical scattering properties of leaves. The present study therefore proposes to test the potential of a biospeckle measurement to observe the temporal evolution in different varieties of sunflower plants under water stress. An experiment applying controlled water stress with osmotic shock using polyethylene glycol 6000 (PEG) was conducted on two sunflower varieties: one sensitive, and the other more tolerant to water stress. Temporal monitoring of biospeckle activity in these plants was performed using the average value of difference (AVD) indicator. Results indicate that AVD highlights the difference in biospeckle activity between day and night, with lower activity at night for both varieties. The addition of PEG entailed a gradual decrease in values throughout the experiment, particularly for the sensitive variety. The results obtained are consistent with the behaviour of the varieties submitted to water stress. Indeed, a few days after the introduction of PEG, a stronger decrease in AVD indicator values was observed for the sensitive variety than for the resistant variety. This study highlights the dynamics of biospeckle activity for different sunflower varieties undergoing water stress and can be considered as a promising phenotyping tool
The Hippo pathway terminal effector TAZ/WWTR1 mediates oxaliplatin sensitivity in p53 proficient colon cancer cells
No full textInternational audienceYAP and TAZ, the Hippo pathway terminal transcriptional activators, are frequently upregulated in cancers. In tumor cells, they have been mainly associated with increased tumorigenesis controlling different aspects from cell cycle regulation, stemness, or resistance to chemotherapies. In fewer cases, they have also been shown to oppose cancer progression, including by promoting cell death through the action of the p73/YAP transcriptional complex, in particular after chemotherapeutic drug exposure. Using HCT116 cells, we show here that oxaliplatin treatment led to core Hippo pathway down-regulation and nuclear accumulation of TAZ. We further show that TAZ was required for the increased sensitivity of HCT116 cells to oxaliplatin, an effect that appeared independent of p73, but which required the nuclear relocalization of TAZ. Accordingly, Verteporfin and CA3, two drugs affecting the activity of YAP and TAZ, showed antagonistic effects with oxaliplatin in co-treatments. Importantly, using several colorectal cell lines, we show that the sensitizing action of TAZ to oxaliplatin is dependent on the p53 status of the cells. Our results support thus an early action of TAZ to sensitize cells to oxaliplatin, consistent with a model in which nuclear TAZ in the context of DNA damage and p53 activity pushes cells towards apoptosis
Multiple-base Logarithmic Quantization and Application in Reduced Precision AI Computations
No full textInternational audienceThe power of logarithmic quantizations and computations has been recognized as a useful tool in optimizing the performance of large ML models. In this article, we provide results that demonstrate significantly better quantization signal-to-noise ratio performance thanks to multiple-base logarithmic number systems (MDLNS) in comparison with the floating-point quantizations that use the same number of bits. On a hardware level, we present details about our Xilinx VCU-128 FPGA design for dot product and matrix-vector computations. The MDLNS matrix-vector design significantly outperforms equivalent fixed-point binary designs in terms of area (A) and time (T) complexity and power consumption as evidenced by a 4x scaling of AT2 metric for VLSI performance, and 57% increase in computational throughput per watt compared to fixed-point arithmetic
Dispositif d'apprentissage automatique collaboratif pour la pratique du débat
No full textInternational audienceThe AREN-DIA project (ARgumentation and Digital - Didactics & Artificial Intelligence) aims to raise awareness of debate practices within both educational contexts and civil society discussions. The project is realized through the creation and testing of a debate platform. This platform allows for engaging in structured debates based on text, renewing the traditional approach to argumentative exchanges. It incorporates collaborative Natural Language Processing (NLP) technology to enhance the efficiency of the debate process. Our paper focuses on the challenges of the AI axis of the project, specifically, how to design a reinforcement mechanism that encourages users to participate in improving the AI system producing a structured representation of debate discourse. Thus, an automatic procedure complements the debate by suggesting key terms that synthesize the statements made. This indexing serves as the starting point for the machine's analysis and support of the debate. It is subject to completion by users who are invited to validate, invalidate, or supplement these key terms. To resolve semantic ambiguity, we employ an enrichment step for the terms to prepare them for the knowledge extraction operation based on Formal Concept Analysis (FCA). The knowledge, expressed in the form of implications, is utilized to update the relationships in the knowledge base that are leveraged during these processes.Le projet AREN-DIA (ARgumentation Et Numérique - Didactique & Intelligence Artificielle) vise à sensibiliser à la pratique du débat dans le cadre de l’éducation et au sein de la société civile. Le projet se matérialise à travers la création et la mise à l’épreuve d’une plateforme de débat. Cette dernière offre la possibilité d’engager des débats structurés à partir d’un texte, renouvelant l’approche traditionnelle des échanges argumentatifs. Elle intègre une technologie collaborative de Traitement Automatique du Langage afin d’augmenter l’efficacité du processus de débat. Notre article se consacrera aux enjeux de l’axe IA du projet, à savoir : comment concevoir un mécanisme de renforcement incitant les utilisateurs à participer à l’amélioration du système d’IA produisant une représentation structurée des propos d’un débat ? Ainsi, une procédure automatique vient compléter le débat en suggérant des termes-clés synthétisant les propos tenus. Cette indexation est le point de départ de l’analyse et de l’accompagnement du débat par la machine. Elle est soumise à une interaction avec les utilisateurs, qui seront invités à valider, invalider ou compléter ces termes-clés. Afin de lever l’ambiguïté sémantique, nous avons recours à une étape d’enrichissement des termes pour les préparer à l’opération d’extraction de connaissances basée sur l’analyse formelle de concepts (AFC). Ces connaissances, sous forme d’implications, sont utilisées pour mettre à jour les relations dans la base de connaissances exploitée
Decoding Simultaneous Rational Evaluation Codes
No full textInternational audienceIn this paper, we deal with the problem of simultaneous reconstruction of a vector of rational numbers, given modular reductions containing errors (SRNRwE). Our methods apply as well to the simultaneous reconstruction of rational functions given evaluations containing errors (SRFRwE), improving known results [7 , 9]. In the latter case, one can take advantage of techniques from coding theory [4, 10] and provide an algorithm that extends classical Reed-Solomon decoding. In recent works [7 , 9], interleaved Reed-Solomon codes [3 , 19 ] are used to correct beyond the unique decoding capability in the case of random errors at the price of positive but small failure probability. Our first contribution is to extend these works to the simultaneous reconstruction with errors of rational numbers instead of functions. Thus considering rational number codes [ 16 ], we provide an algorithm decoding beyond the unique decoding capability and, as a central result of this paper, we analyze in detail its failure probability. Our analysis generalizes for the first time the best known analysis for interleaved Reed-Solomon codes [19] to SRFRwE, improving on the existing bound [8], to interleaved Chinese remainder codes, also improving the known bound [1], and finally for the first time to SRNRwE
IMGT/mAb-KG: the knowledge graph for therapeutic monoclonal antibodies
No full textInternational audienceIntroduction: Therapeutic monoclonal antibodies (mAbs) have demonstrated promising outcomes in diverse clinical indications, including but not limited to graft rejection, cancer, and autoimmune diseases lately.Recognizing the crucial need for the scientific community to quickly and easily access dependable information on monoclonal antibodies (mAbs), IMGT®, the international ImMunoGeneTics information system®, provides a unique and invaluable resource: IMGT/mAb-DB, a comprehensive database of therapeutic mAbs, accessible via a user-friendly web interface. However, this approach restricts more sophisticated queries and segregates information from other databases.Methods To connect IMGT/mAb-DB with the rest of the IMGT databases, we created IMGT/mAb-KG, a knowledge graph for therapeutic monoclonal antibodies connected to IMGT structures and genomics databases. IMGT/mAb-KG is developed using the most effective methodologies and standards of semantic web and acquires data from IMGT/mAb-DB. Concerning interoperability, IMGT/mAb-KG reuses terms from biomedical resources and is connected to related resources.Results and discussion In February 2024, IMGT/mAb-KG, encompassing a total of 139,629 triplets, provides access to 1,489 mAbs, approximately 500 targets, and over 500 clinical indications. It offers detailed insights into the mechanisms of action of mAbs, their construction, and their various products and associated studies. Linked to other resources such as Thera-SAbDab (Therapeutic Structural Antibody Database), PharmGKB (a comprehensive resource curating knowledge on the impact of genetic variation on drug response), PubMed, and HGNC (HUGO Gene Nomenclature Committee), IMGT/mAb-KG is an essential resource for mAb development. A user-friendly web interface facilitates the exploration and analyse of the content of IMGT/mAb-KG
Robots marins et sous-marins - Perception, modélisation, commande (deuxième édition)
No full textCet article présente les principales catégories de robots marins et sous-marins, en précisant leurs domaines d'application ainsi que les aspects technologiques et scientifiques associés. L’article dresse ensuite l’inventaire des moyens de localisation, de perception et de communication, et en résume les caractéristiques techniques essentielles. La troisième section établit le modèle dynamique des véhicules sous-marins et détaille les phénomènes et forces en jeu. Enfin, trois approches de commande sont proposées et comparées expérimentalement
Scalable Enforcement of Geometric Non-interference Constraints for Gradient-Based Optimization
No full textInternational audienceMany design optimization problems include constraints to prevent intersection of the geometric shape being optimized with other objects or with domain boundaries. When applying gradient-based optimization to such problems, the constraint function must provide an accurate representation of an object's boundary and be smooth, amenable to numerical differentiation, and fast-to-evaluate for a large number of points. We propose the use of tensor product Bsplines to generate an efficient-to-evaluate implicit level set function that locally approximates the signed distance function for representing geometric non-interference constraints. Adapting ideas from the field of surface reconstruction, we formulate an energy minimization problem to compute the implicit function values at B-spline control points given an oriented point set sampled over a geometric shape. Unlike previous non-interference constraint formulations, our method requires an initial setup, but results in a more efficient and scalable representation of geometric non-interference constraints. This paper presents the results of accuracy and scaling studies performed on our formulation. We demonstrate our method by solving a medical robot design optimization problem with non-interference constraints. We achieve constraint evaluation times on the order of 10 −6 seconds per point on a modern desktop workstation, and a maximum on-surface error of less than 1.0% of the minimum bounding box diagonal for all examples studied. Overall, our method provides an effective formulation for non-interference constraint enforcement with high computational efficiency for gradient-based design optimization problems whose solutions may require hundreds of evaluations of constraints and their derivatives
Digraph redicolouring
No full textInternational audienceIn this work, we generalize several results on graph recolouring to digraphs. Given two k-dicolourings of a digraph D, we prove that it is PSPACE-complete to decide whether we can transform one into the other by recolouring one vertex at each step while maintaining a dicolouring at any step even for k = 2 and for digraphs with maximum degree 5 or oriented planar graphs with maximum degree 6. A digraph is said to be k-mixing if there exists a transformation between any pair of k-dicolourings. We show that every digraph D is k-mixing for all k ≥ δ * min (D) + 2, generalizing a result due to Dyer et al. We also prove that every oriented graph ⃗ G is k-mixing for all k ≥ δ * max (⃗ G) + 1 and for all k ≥ δ * avg (⃗ G) + 1. Here δ * min , δ * max , and δ * avg denote the min-degeneracy, the max-degeneracy, and the average-degeneracy respectively. We pose as a conjecture that, for every digraph D, the dicolouring graph of D on k ≥ δ * min (D) + 2 colours has diameter at most O(|V (D)| 2). This is the analogue of Cereceda's conjecture for digraphs. We generalize to digraphs two results supporting Cereceda's conjecture. We first prove that the dicolouring graph of any digraph D on k ≥ 2δ * min (D) + 2 colours has linear diameter, extending a result from Bousquet and Perarnau. We also prove that the analogue of Cereceda's conjecture is true when k ≥ 3 2 (δ * min (D) + 1), which generalizes a result from Bousquet and Heinrich. Restricted to the special case of oriented graphs, we prove that the dicolouring graph of any subcubic oriented graph on k ≥ 2 colours is connected and has diameter at most 2n. We conjecture that every non 2-mixing oriented graph has maximum average degree at least 4, and we provide some support for this conjecture by proving it on the special case of 2-freezable oriented graphs. More generally, we show that every k-freezable oriented graph on n vertices must contain at least kn + k(k − 2) arcs, and we give a family of k-freezable oriented graphs that reach this bound. In the general case, we prove as a partial result that every non 2-mixing oriented graph has maximum average degree at least 7 2