Índice
Literatura académica sobre el tema "Apprentissage frugal"
Crea una cita precisa en los estilos APA, MLA, Chicago, Harvard y otros
Consulte las listas temáticas de artículos, libros, tesis, actas de conferencias y otras fuentes académicas sobre el tema "Apprentissage frugal".
Junto a cada fuente en la lista de referencias hay un botón "Agregar a la bibliografía". Pulsa este botón, y generaremos automáticamente la referencia bibliográfica para la obra elegida en el estilo de cita que necesites: APA, MLA, Harvard, Vancouver, Chicago, etc.
También puede descargar el texto completo de la publicación académica en formato pdf y leer en línea su resumen siempre que esté disponible en los metadatos.
Tesis sobre el tema "Apprentissage frugal"
Leconte, Louis. "Compression and federated learning : an approach to frugal machine learning". Electronic Thesis or Diss., Sorbonne université, 2024. http://www.theses.fr/2024SORUS107.
Texto completo“Intelligent” devices and tools are gradually becoming the standard, as the implementation of algorithms based on artificial neural networks is experiencing widespread development. Neural networks consist of non-linear machine learning models that manipulate high-dimensional objects and obtain state-of-the-art performances in various areas, such as image recognition, speech recognition, natural language processing, and recommendation systems.However, training a neural network on a device with lower computing capacity can be challenging, as it can imply cutting back on memory, computing time or power. A natural approach to simplify this training is to use quantized neural networks, whose parameters and operations use efficient low-bit primitives. However, optimizing a function over a discrete set in high dimension is complex, and can still be prohibitively expensive in terms of computational power. For this reason, many modern applications use a network of devices to store individual data and share the computational load. A new approach, federated learning, considers a distributed environment: Data is stored on devices and a centralized server orchestrates the training process across multiple devices.In this thesis, we investigate different aspects of (stochastic) optimization with the goal of reducing energy costs for potentially very heterogeneous devices. The first two contributions of this work are dedicated to the case of quantized neural networks. Our first idea is based on an annealing strategy: we formulate the discrete optimization problem as a constrained optimization problem (where the size of the constraint is reduced over iterations). We then focus on a heuristic for training binary deep neural networks. In this particular framework, the parameters of the neural networks can only have two values. The rest of the thesis is about efficient federated learning. Following our contributions developed for training quantized neural network, we integrate them into a federated environment. Then, we propose a novel unbiased compression technique that can be used in any gradient based distributed optimization framework. Our final contributions address the particular case of asynchronous federated learning, where devices have different computational speeds and/or access to bandwidth. We first propose a contribution that reweights the contributions of distributed devices. Then, in our final work, through a detailed queuing dynamics analysis, we propose a significant improvement to the complexity bounds provided in the literature onasynchronous federated learning.In summary, this thesis presents novel contributions to the field of quantized neural networks and federated learning by addressing critical challenges and providing innovative solutions for efficient and sustainable learning in a distributed and heterogeneous environment. Although the potential benefits are promising, especially in terms of energy savings, caution is needed as a rebound effect could occur
Cazorla, Clément. "Analyse d'images en microscopie par réseaux de neurones dans un contexte frugal". Electronic Thesis or Diss., Université de Toulouse (2023-....), 2024. http://www.theses.fr/2024TLSES028.
Texto completoIn this thesis manuscript, we explore the use of neural networks in a frugal context, i.e. with little training data, applied to microscopy image analysis tasks. In the course of this work, I have carried out substantial integration work in ergonomic user interfaces. These take the form of plugins for Napari, an open-source software package for multi-dimensional image visualization. More broadly, this work brings microscope users and deep learning technologies, currently complex to grasp and implement, closer together through interactive learning software. Biologists are making increasing use of various imaging modalities, with the aim of accessing new sources of information in both basic and applied research contexts. This intensification in the use of microscopes has led to large volumes of data, the analysis of which today poses real technical and scientific questions. Data processing is often carried out manually by biologists (e.g. cell counting), leading to problems of inter- and intra-operator reproducibility of results, as well as processing times that can be very long. The development of deep learning over the last few years has opened up analysis possibilities that were unattainable just a few years ago. Nevertheless, at a time when energy consumption and storage are becoming crucial issues, these approaches pose certain problems. Collecting and annotating large quantities of images is costly and complex. Their reliability is sometimes called into question, as these neural models are frequently considered difficult, if not impossible, to interpret. Today, numerous initiatives are underway to pool models and computational resources, and to research the interpretability of these networks, in order to make them accessible to a wider audience. This thesis focuses on two main areas of image analysis : segmentation and classification. First, we will review the state of the art in learning-based analysis methods for microscopy, and then present the frugal methods we have developed for these two tasks. We will then present their integration as open-source Napari plug-ins. Finally, we'll use concrete examples from both academic and industrial collaborations to demonstrate the value of these methods and the results they deliver
Deschamps, Sébastien. "Apprentissage actif profond pour la reconnaissance visuelle à partir de peu d’exemples". Electronic Thesis or Diss., Sorbonne université, 2023. http://www.theses.fr/2023SORUS199.
Texto completoAutomatic image analysis has improved the exploitation of image sensors, with data coming from different sensors such as phone cameras, surveillance cameras, satellite imagers or even drones. Deep learning achieves excellent results in image analysis applications where large amounts of annotated data are available, but learning a new image classifier from scratch is a difficult task. Most image classification methods are supervised, requiring annotations, which is a significant investment. Different frugal learning solutions (with few annotated examples) exist, including transfer learning, active learning, semi-supervised learning or meta-learning. The goal of this thesis is to study these frugal learning solutions for visual recognition tasks, namely image classification and change detection in satellite images. The classifier is trained iteratively by starting with only a few annotated samples, and asking the user to annotate as little data as possible to obtain satisfactory performance. Deep active learning was initially studied with other methods and suited our operational problem the most, so we chose this solution. In this thesis, we have developed an interactive approach, where we ask the most informative questions about the relevance of the data to an oracle (annotator). Based on its answers, a decision function is iteratively updated. We model the probability that the samples are relevant, by minimizing an objective function capturing the representativeness, diversity and ambiguity of the data. Data with high probability are then selected for annotation. We have improved this approach, using reinforcement learning to dynamically and accurately weight the importance of representativeness, diversity and ambiguity of the data in each active learning cycle. Finally, our last approach consists of a display model that selects the most representative and diverse virtual examples, which adversely challenge the learned model, in order to obtain a highly discriminative model in subsequent iterations of active learning. The good results obtained against the different baselines and the state of the art in the tasks of satellite image change detection and image classification have demonstrated the relevance of the proposed frugal learning models, and have led to various publications (Sahbi et al. 2021; Deschamps and Sahbi 2022b; Deschamps and Sahbi 2022a; Sahbi and Deschamps2022)
Prouteau, Thibault. "Graphs,Words, and Communities : converging paths to interpretability with a frugal embedding framework". Electronic Thesis or Diss., Le Mans, 2024. http://www.theses.fr/2024LEMA1006.
Texto completoRepresentation learning with word and graph embedding models allows distributed representations of information that can in turn be used in input of machine learning algorithms. Through the last two decades, the tasks of embedding graphs’ nodes and words have shifted from matrix factorization approaches that could be trained in a matter of minutes to large models requiring ever larger quantities of training data and sometimes weeks on large hardware architectures. However, in a context of global warming where sustainability is a critical concern, we ought to look back to previous approaches and consider their performances with regard to resources consumption. Furthermore, with the growing involvement of embeddings in sensitive machine learning applications (judiciary system, health), the need for more interpretable and explainable representations has manifested. To foster efficient representation learning and interpretability, this thesis introduces Lower Dimension Bipartite Graph Framework (LDBGF), a node embedding framework able to embed with the same pipeline graph data and text from large corpora represented as co-occurrence networks. Within this framework, we introduce two implementations (SINr-NR, SINr-MF) that leverage community detection in networks to uncover a latent embedding space where items (nodes/words) are represented according to their links to communities. We show that SINr-NR and SINr-MF can compete with similar embedding approaches on tasks such as predicting missing links in networks (link prediction) or node features (degree centrality, PageRank score). Regarding word embeddings, we show that SINr-NR is a good contender to represent words via word co-occurrence networks. Finally, we demonstrate the interpretability of SINr-NR on multiple aspects. First with a human evaluation that shows that SINr-NR’s dimensions are to some extent interpretable. Secondly, by investigating sparsity of vectors, and how having fewer dimensions may allow interpreting how the dimensions combine and allow sense to emerge
Tuo, Aboubacar. "Extraction d'événements à partir de peu d'exemples par méta-apprentissage". Electronic Thesis or Diss., université Paris-Saclay, 2023. http://www.theses.fr/2023UPASG098.
Texto completoInformation Extraction (IE) is a research field with the objective of automatically identifying and extracting structured information within a given domain from unstructured or minimally structured text data. The implementation of such extractions often requires significant human efforts, either in the form of rule development or the creation of annotated data for systems based on machine learning. One of the current challenges in information extraction is to develop methods that minimize the costs and development time of these systems whenever possible. This thesis focuses on few-shot event extraction through a meta-learning approach that aims to train IE models from only few data. We have redefined the task of event extraction from this perspective, aiming to develop systems capable of quickly adapting to new contexts with a small volume of training data. First, we propose methods to enhance event trigger detection by developing more robust representations for this task. Then, we tackle the specific challenge raised by the "NULL" class (absence of events) within this framework. Finally, we evaluate the effectiveness of our proposals within the broader context of event extraction by extending their application to the extraction of event arguments
Cherdo, Yann. "Détection d'anomalie non supervisée sur les séries temporelle à faible coût énergétique utilisant les SNNs". Electronic Thesis or Diss., Université Côte d'Azur, 2024. http://www.theses.fr/2024COAZ4018.
Texto completoIn the context of the predictive maintenance of the car manufacturer Renault, this thesis aims at providing low-power solutions for unsupervised anomaly detection on time-series. With the recent evolution of cars, more and more data are produced and need to be processed by machine learning algorithms. This processing can be performed in the cloud or directly at the edge inside the car. In such a case, network bandwidth, cloud services costs, data privacy management and data loss can be saved. Embedding a machine learning model inside a car is challenging as it requires frugal models due to memory and processing constraints. To this aim, we study the usage of spiking neural networks (SNNs) for anomaly detection, prediction and classification on time-series. SNNs models' performance and energy costs are evaluated in an edge scenario using generic hardware models that consider all calculation and memory costs. To leverage as much as possible the sparsity of SNNs, we propose a model with trainable sparse connections that consumes half the energy compared to its non-sparse version. This model is evaluated on anomaly detection public benchmarks, a real use-case of anomaly detection from Renault Alpine cars, weather forecasts and the google speech command dataset. We also compare its performance with other existing SNN and non-spiking models. We conclude that, for some use-cases, spiking models can provide state-of-the-art performance while consuming 2 to 8 times less energy. Yet, further studies should be undertaken to evaluate these models once embedded in a car. Inspired by neuroscience, we argue that other bio-inspired properties such as attention, sparsity, hierarchy or neural assemblies dynamics could be exploited to even get better energy efficiency and performance with spiking models. Finally, we end this thesis with an essay dealing with cognitive neuroscience, philosophy and artificial intelligence. Diving into conceptual difficulties linked to consciousness and considering the deterministic mechanisms of memory, we argue that consciousness and the self could be constitutively independent from memory. The aim of this essay is to question the nature of humans by contrast with the ones of machines and AI