The goal of the Mars Sample Return campaign is to collect soil samples from the surface of Mars and return them to Earth for further study. The samples will be acquired and stored in metal tubes by the Perseverance rover and deposited on the Martian surface. As part of this campaign, it is expected that the Sample Fetch Rover will be in charge of localizing and gathering up to 35 sample tubes over 150 Martian sols. Autonomous capabilities are critical for the success of the overall campaign and for the Sample Fetch Rover in particular. This work proposes a novel system architecture for the autonomous detection and pose estimation of the sample tubes. For the detection stage, a Deep Neural Network and transfer learning from a synthetic dataset are proposed. The dataset is created from photorealistic 3D simulations of Martian scenarios. Additionally, the sample tubes poses are estimated using Computer Vision techniques such as contour detection and line fitting on the detected area. Finally, laboratory tests of the Sample Localization procedure are performed using the ExoMars Testing Rover on a Mars-like testbed. These tests validate the proposed approach in different hardware architectures, providing promising results related to the sample detection and pose estimation.

Spacecraft pose estimation is a key task to enable space missions in which two spacecrafts must navigate around each other. Current state-of-the-art algorithms for pose estimation employ data-driven techniques. However, there is an absence of real training data for spacecraft imaged in space conditions due to the costs and difficulties associated with the space environment. This has motivated the introduction of 3D data simulators, solving the issue of data availability but introducing a large gap between the training (source) and test (target) domains. We explore a method that incorporates 3D structure into the spacecraft pose estimation pipeline to provide robustness to intensity domain shift and we present an algorithm for unsupervised domain adaptation with robust pseudo-labelling. Our solution has ranked second in the two categories of the 2021 Pose Estimation Challenge organised by the European Space Agency and the Stanford University, achieving the lowest average error over the two categories.

Privacy-preserving machine learning in data-sharing processes is an ever-critical task that enables collaborative training of Machine Learning (ML) models without the need to share the original data sources. It is especially relevant when an organization must assure that sensitive data remains private throughout the whole ML pipeline, i.e., training and inference phases. This paper presents an innovative framework that uses Representation Learning via autoencoders to generate privacy-preserving embedded data. Thus, organizations can share the data representation to increase machine learning models' performance in scenarios with more than one data source for a shared predictive downstream task.

In this work, we propose a framework relying solely on chat-based customer support (CS) interactions for predicting the recommendation decision of individual users. For our case study, we analyzed a total number of 16.4k users and 48.7k customer support conversations within the financial vertical of a large e-commerce company in Latin America. Consequently, our main contributions and objectives are to use Natural Language Processing (NLP) to assess and predict the recommendation behavior where, in addition to using static sentiment analysis, we exploit the predictive power of each user's sentiment dynamics. Our results show that, with respective feature interpretability, it is possible to predict the likelihood of a user to recommend a product or service, based solely on the message-wise sentiment evolution of their CS conversations in a fully automated way.

We introduce pyGSL, a Python library that provides efficient implementations of state-of-the-art graph structure learning models along with diverse datasets to evaluate them on. The implementations are written in GPU-friendly ways, allowing one to scale to much larger network tasks. A common interface is introduced for algorithm unrolling methods, unifying implementations of recent state-of-the-art techniques and allowing new methods to be quickly developed by avoiding the need to rebuild the underlying unrolling infrastructure. Implementations of differentiable graph structure learning models are written in PyTorch, allowing us to leverage the rich software ecosystem that exists e.g., around logging, hyperparameter search, and GPU-communication. This also makes it easy to incorporate these models as components in larger gradient based learning systems where differentiable estimates of graph structure may be useful, e.g. in latent graph learning. Diverse datasets and performance metrics allow consistent comparisons across models in this fast growing field. The full code repository can be found on https://github.com/maxwass/pyGSL.

We present the GPry algorithm for fast Bayesian inference of general (non-Gaussian) posteriors with a moderate number of parameters. GPry does not need any pre-training, special hardware such as GPUs, and is intended as a drop-in replacement for traditional Monte Carlo methods for Bayesian inference. Our algorithm is based on generating a Gaussian Process surrogate model of the log-posterior, aided by a Support Vector Machine classifier that excludes extreme or non-finite values. An active learning scheme allows us to reduce the number of required posterior evaluations by two orders of magnitude compared to traditional Monte Carlo inference. Our algorithm allows for parallel evaluations of the posterior at optimal locations, further reducing wall-clock times. We significantly improve performance using properties of the posterior in our active learning scheme and for the definition of the GP prior. In particular we account for the expected dynamical range of the posterior in different dimensionalities. We test our model against a number of synthetic and cosmological examples. GPry outperforms traditional Monte Carlo methods when the evaluation time of the likelihood (or the calculation of theoretical observables) is of the order of seconds; for evaluation times of over a minute it can perform inference in days that would take months using traditional methods. GPry is distributed as an open source Python package (pip install gpry) and can also be found at https://github.com/jonaselgammal/GPry.

与单个IMU相比，多个刚性连接的惯性测量单元（IMU）传感器提供了更丰富的数据流。最先进的方法遵循IMU测量的概率模型，基于在贝叶斯框架下组合的错误的随机性质。但是，负担得起的低级IMU此外，由于其不受相应的概率模型所掩盖的缺陷而遭受了系统的错误。在本文中，我们提出了一种方法，即合并多个IMU（MIMU）传感器数据的最佳轴组成（BAC），以进行准确的3D置置估计，该数据通过从集合中动态选择最佳的IMU轴来考虑随机和系统误差所有可用的轴。我们在MIMU视觉惯性传感器上评估了我们的方法，并将方法的性能与MIMU数据融合的最新方法进行比较。我们表明，BAC的表现优于后者，并且在开放环路中的方向和位置估计都可以提高20％的精度，但需要适当的处理以保持获得的增益。

任何机器学习（ML）算法的性能受到其超参数的选择影响。由于培训和评估ML算法通常很昂贵，因此需要在实践中有效地计算高参数优化（HPO）方法。多数目标HPO的大多数现有方法都使用进化策略和基于元模型的优化。但是，很少有方法可以解释性能测量中的不确定性。本文提出了多目标超参数优化的结果，并在评估ML算法的情况下进行了不确定性。我们将树结构化parzen估计量（TPE）的采样策略与训练高斯过程回归（GPR）在异质噪声后获得的元模型相结合。关于三个分析测试功能和三个ML问题的实验结果表明，相对于超量指标，多目标TPE和GPR的改善。

可以通过玩游戏来训练代理商来回答困难的数学问题吗？我们考虑了整数可行性问题，这是决定线性方程和不平等系统是否具有具有整数值的解决方案的挑战。对于许多数学和计算机科学领域的应用，这是一个著名的NP完整问题。我们的论文描述了一个新颖的代数增强学习框架，该框架使代理商可以玩相当于整数可行性问题的游戏。我们解释了如何将整数可行性问题转换为具有固定保证金总和的一组阵列的游戏。游戏从初始状态（数组）开始，并采取法律举措使利润率保持不变，我们的目标是最终与零位置的零位置达到胜利状态。为了赢得比赛，玩家必须在初始状态和最终终端获胜状态之间找到一条路径。找到这样的获胜状态等同于解决整数可行性问题。关键代数成分是“基础轴向运输polyhedron的曲折理想的基础”。gr \'obner可以看作是游戏的一组连接移动（动作）。然后，我们提出了一种新型的RL方法，该方法训练代理以预测连续空间中的移动，以应对较大的动作空间。然后将连续的移动投射到一组法律移动上，以使该路径始终导致有效状态。作为概念的证明，我们在实验中证明了我们的代理商可以很好地发挥我们最简单的游戏版本，用于2向表。我们的工作突出了培训代理商通过当代机器学习方法来训练代理商玩游戏的潜力来解决非平凡的数学查询的潜力。

考虑到大量未标记的语音数据和高标签成本，无监督的学习方法对于更好的系统开发至关重要。最成功的方法之一是对比度的自我监督方法，这些方法需要负采样：采样替代样品与当前样品（锚）对比。但是，很难确保所有负样本属于与没有标签的锚类别不同的​​类别。本文在未标记的语音语料库上应用了一种非对抗性的自我监督学习方法来学习话语级的嵌入。我们使用没有标签的蒸馏（Dino），在计算机视觉中提出，并将其改编为语音域。与对比度方法不同，Dino不需要负采样。这些嵌入是根据说话者验证和情感识别评估的。在说话者验证中，无监督的恐龙与余弦评分嵌入了voxceleb1测试试验中的4.38％EER。这表现优于最佳的对比度自我监督方法，而EER中的相对相对40％。不需要扬声器标签的迭代伪标记训练管道将EER进一步提高到1.89％。在情感识别中，Iemocap，Crema-D和MSP播客的Micro-F1得分分别进行了60.87、79.21和56.98％的恐龙。结果暗示着恐龙嵌入到不同语音应用中的普遍性。