样本是否足够丰富，至少在本地确定神经网络的参数？为了回答这个问题，我们通过固定其某些权重的值来介绍给定深层神经网络的新局部参数化。这使我们能够定义本地提升操作员，其倒置是高维空间的平滑歧管的图表。Deep Relu神经网络实现的函数由依赖样本的线性操作员组成局部提升。我们从这种方便的表示中得出了局部可识别性的几何必要条件。查看切线空间，几何条件提供了：1/可识别性的尖锐而可测试的必要条件以及2/可识别局部可识别性的尖锐且可测试的足够条件。可以使用反向传播和矩阵等级计算对条件的有效性进行数值测试。

我们研究神经网络表达能力的基本限制。给定两组$ f $，$ g $的实值函数，我们首先证明了$ f $中的功能的一般下限，可以在$ l^p（\ mu）$ norm中通过$ g中的功能近似$，对于任何$ p \ geq 1 $和任何概率度量$ \ mu $。下限取决于$ f $的包装数，$ f $的范围以及$ g $的脂肪震动尺寸。然后，我们实例化了$ g $对应于分段的馈电神经网络的情况，并详细描述了两组$ f $：h {\“ o} lder balls和多变量单调函数。除了匹配（已知或新的）上限与日志因素外，我们的下限还阐明了$ l^p $ Norm或SUP Norm中近似之间的相似性或差异，解决了Devore等人的开放问题（2021年））。我们的证明策略与SUP Norm案例不同，并使用了Mendelson（2002）的关键概率结果。

由于其在输入空间子集上的功能的知识，因此可以根据情况，诅咒或祝福来恢复神经网络的参数权重和偏差的可能性。一方面，恢复参数允许更好的对抗攻击，并且还可以从用于构造网络的数据集中披露敏感信息。另一方面，如果可以恢复网络的参数，它可以保证用户可以解释潜在空间中的特征。它还提供基础，以获得对网络性能的正式保障。因此，表征可以识别其参数的网络以及其参数不能的网络是很重要的。在本文中，我们在深度全连接的前馈recu网络上提供了一组条件，在该馈电中，网络的参数是唯一识别的模型置换和正重型 - 从其实现输入空间的子集。

For long-term simultaneous planning, localization and mapping (SPLAM), a robot should be able to continuously update its map according to the dynamic changes of the environment and the new areas explored. With limited onboard computation capabilities, a robot should also be able to limit the size of the map used for online localization and mapping. This paper addresses these challenges using a memory management mechanism, which identifies locations that should remain in a Working Memory (WM) for online processing from locations that should be transferred to a Long-Term Memory (LTM). When revisiting previously mapped areas that are in LTM, the mechanism can retrieve these locations and place them back in WM for online SPLAM. The approach is tested on a robot equipped with a short-range laser rangefinder and a RGB-D camera, patrolling autonomously 10.5 km in an indoor environment over 11 sessions while having encountered 139 people.

Traditional approaches to RL have focused on learning decision policies directly from episodic decisions, while slowly and implicitly learning the semantics of compositional representations needed for generalization. While some approaches have been adopted to refine representations via auxiliary self-supervised losses while simultaneously learning decision policies, learning compositional representations from hand-designed and context-independent self-supervised losses (multi-view) still adapts relatively slowly to the real world, which contains many non-IID subspaces requiring rapid distribution shift in both time and spatial attention patterns at varying levels of abstraction. In contrast, supervised language model cascades have shown the flexibility to adapt to many diverse manifolds, and hints of self-learning needed for autonomous task transfer. However, to date, transfer methods for language models like few-shot learning and fine-tuning still require human supervision and transfer learning using self-learning methods has been underexplored. We propose a self-supervised loss policy called contrastive distillation which manifests latent variables with high mutual information with both source and target tasks from weights to tokens. We show how this outperforms common methods of transfer learning and suggests a useful design axis of trading off compute for generalizability for online transfer. Contrastive distillation is improved through sampling from memory and suggests a simple algorithm for more efficiently sampling negative examples for contrastive losses than random sampling.

As of 2022, greenhouse gases (GHG) emissions reporting and auditing are not yet compulsory for all companies and methodologies of measurement and estimation are not unified. We propose a machine learning-based model to estimate scope 1 and scope 2 GHG emissions of companies not reporting them yet. Our model, specifically designed to be transparent and completely adapted to this use case, is able to estimate emissions for a large universe of companies. It shows good out-of-sample global performances as well as good out-of-sample granular performances when evaluating it by sectors, by countries or by revenues buckets. We also compare our results to those of other providers and find our estimates to be more accurate. Thanks to the proposed explainability tools using Shapley values, our model is fully interpretable, the user being able to understand which factors split explain the GHG emissions for each particular company.

In intensively managed forests in Europe, where forests are divided into stands of small size and may show heterogeneity within stands, a high spatial resolution (10 - 20 meters) is arguably needed to capture the differences in canopy height. In this work, we developed a deep learning model based on multi-stream remote sensing measurements to create a high-resolution canopy height map over the "Landes de Gascogne" forest in France, a large maritime pine plantation of 13,000 km$^2$ with flat terrain and intensive management. This area is characterized by even-aged and mono-specific stands, of a typical length of a few hundred meters, harvested every 35 to 50 years. Our deep learning U-Net model uses multi-band images from Sentinel-1 and Sentinel-2 with composite time averages as input to predict tree height derived from GEDI waveforms. The evaluation is performed with external validation data from forest inventory plots and a stereo 3D reconstruction model based on Skysat imagery available at specific locations. We trained seven different U-net models based on a combination of Sentinel-1 and Sentinel-2 bands to evaluate the importance of each instrument in the dominant height retrieval. The model outputs allow us to generate a 10 m resolution canopy height map of the whole "Landes de Gascogne" forest area for 2020 with a mean absolute error of 2.02 m on the Test dataset. The best predictions were obtained using all available satellite layers from Sentinel-1 and Sentinel-2 but using only one satellite source also provided good predictions. For all validation datasets in coniferous forests, our model showed better metrics than previous canopy height models available in the same region.

Purpose: The purpose of this paper is to present a method for real-time 2D-3D non-rigid registration using a single fluoroscopic image. Such a method can find applications in surgery, interventional radiology and radiotherapy. By estimating a three-dimensional displacement field from a 2D X-ray image, anatomical structures segmented in the preoperative scan can be projected onto the 2D image, thus providing a mixed reality view. Methods: A dataset composed of displacement fields and 2D projections of the anatomy is generated from the preoperative scan. From this dataset, a neural network is trained to recover the unknown 3D displacement field from a single projection image. Results: Our method is validated on lung 4D CT data at different stages of the lung deformation. The training is performed on a 3D CT using random (non domain-specific) diffeomorphic deformations, to which perturbations mimicking the pose uncertainty are added. The model achieves a mean TRE over a series of landmarks ranging from 2.3 to 5.5 mm depending on the amplitude of deformation. Conclusion: In this paper, a CNN-based method for real-time 2D-3D non-rigid registration is presented. This method is able to cope with pose estimation uncertainties, making it applicable to actual clinical scenarios, such as lung surgery, where the C-arm pose is planned before the intervention.

Human behavior understanding requires looking at minute details in the large context of a scene containing multiple input modalities. It is necessary as it allows the design of more human-like machines. While transformer approaches have shown great improvements, they face multiple challenges such as lack of data or background noise. To tackle these, we introduce the Forced Attention (FAt) Transformer which utilize forced attention with a modified backbone for input encoding and a use of additional inputs. In addition to improving the performance on different tasks and inputs, the modification requires less time and memory resources. We provide a model for a generalised feature extraction for tasks concerning social signals and behavior analysis. Our focus is on understanding behavior in videos where people are interacting with each other or talking into the camera which simulates the first person point of view in social interaction. FAt Transformers are applied to two downstream tasks: personality recognition and body language recognition. We achieve state-of-the-art results for Udiva v0.5, First Impressions v2 and MPII Group Interaction datasets. We further provide an extensive ablation study of the proposed architecture.

White matter bundle segmentation is a cornerstone of modern tractography to study the brain's structural connectivity in domains such as neurological disorders, neurosurgery, and aging. In this study, we present FIESTA (FIber gEneration and bundle Segmentation in Tractography using Autoencoders), a reliable and robust, fully automated, and easily semi-automatically calibrated pipeline based on deep autoencoders that can dissect and fully populate WM bundles. Our framework allows the transition from one anatomical bundle definition to another with marginal calibrating time. This pipeline is built upon FINTA, CINTA, and GESTA methods that demonstrated how autoencoders can be used successfully for streamline filtering, bundling, and streamline generation in tractography. Our proposed method improves bundling coverage by recovering hard-to-track bundles with generative sampling through the latent space seeding of the subject bundle and the atlas bundle. A latent space of streamlines is learned using autoencoder-based modeling combined with contrastive learning. Using an atlas of bundles in standard space (MNI), our proposed method segments new tractograms using the autoencoder latent distance between each tractogram streamline and its closest neighbor bundle in the atlas of bundles. Intra-subject bundle reliability is improved by recovering hard-to-track streamlines, using the autoencoder to generate new streamlines that increase each bundle's spatial coverage while remaining anatomically meaningful. Results show that our method is more reliable than state-of-the-art automated virtual dissection methods such as RecoBundles, RecoBundlesX, TractSeg, White Matter Analysis and XTRACT. Overall, these results show that our framework improves the practicality and usability of current state-of-the-art bundling framework