Search

The minimum sum-of-squares clustering problem (MSSC), also known as $k$-means clustering, refers to the problem of partitioning $n$ data points into $k$ clusters, with the objective of minimizing the total sum of squared Euclidean distances between each point and the center of its assigned cluster. We propose an efficient algorithm for solving large-scale MSSC instances, which combines column generation (CG) with dynamic constraint aggregation (DCA) to effectively reduce the number of constraints considered in the CG master problem. DCA was originally conceived to reduce degeneracy in set partitioning problems by utilizing an aggregated restricted master problem obtained from a partition of the set partitioning constraints into disjoint clusters. In this work, we explore the use of DCA within a CG algorithm for MSSC exact solution. Our method is fine-tuned by a series of ablation studies on DCA design choices, and is demonstrated to significantly outperform existing state-of-the-art exact approaches available in the literature.

Visual localization is crucial for Computer Vision and Augmented Reality (AR) applications, where determining the camera or device's position and orientation is essential to accurately interact with the physical environment. Traditional methods rely on detailed 3D maps constructed using Structure from Motion (SfM) or Simultaneous Localization and Mapping (SLAM), which is computationally expensive and impractical for dynamic or large-scale environments. We introduce MARLoc, a novel localization framework for AR applications that uses known relative transformations within image sequences to perform intra-sequence triangulation, generating 3D-2D correspondences for pose estimation and refinement. MARLoc eliminates the need for pre-built SfM maps, providing accurate and efficient localization suitable for dynamic outdoor environments. Evaluation with benchmark datasets and real-world experiments demonstrates MARLoc's state-of-the-art performance and robustness. By integrating MARLoc into an AR device, we highlight its capability to achieve precise localization in real-world outdoor scenarios, showcasing its practical effectiveness and potential to enhance visual localization in AR applications.

Due to the complexity of modern computer systems, novel and unexpected behaviors frequently occur. Such deviations are either normal occurrences, such as software updates and new user activities, or abnormalities, such as misconfigurations, latency issues, intrusions, and software bugs. Regardless, novel behaviors are of great interest to developers, and there is a genuine need for efficient and effective methods to detect them. Nowadays, researchers consider system calls to be the most fine-grained and accurate source of information to investigate the behavior of computer systems. Accordingly, this paper introduces a novelty detection methodology that relies on a probability distribution over sequences of system calls, which can be seen as a language model. Language models estimate the likelihood of sequences, and since novelties deviate from previously observed behaviors by definition, they would be unlikely under the model. Following the success of neural networks for language models, three architectures are evaluated in this work: the widespread LSTM, the state-of-the-art Transformer, and the lower-complexity Longformer. However, large neural networks typically require an enormous amount of data to be trained effectively, and to the best of our knowledge, no massive modern datasets of kernel traces are publicly available. This paper addresses this limitation by introducing a new open-source dataset of kernel traces comprising over 2 million web requests with seven distinct behaviors. The proposed methodology requires minimal expert hand-crafting and achieves an F-score and AuROC greater than 95% on most novelties while being data- and task-agnostic. The source code and trained models are publicly available on GitHub while the datasets are available on Zenodo., Comment: 12 pages, 7 figures, 3 tables

We consider the problem of predicting the future path of a pedestrian using its motion history and the motion history of the surrounding pedestrians, called social information. Since the seminal paper on Social-LSTM, deep-learning has become the main tool used to model the impact of social interactions on a pedestrian's motion. The demonstration that these models can learn social interactions relies on an ablative study of these models. The models are compared with and without their social interactions module on two standard metrics, the Average Displacement Error and Final Displacement Error. Yet, these complex models were recently outperformed by a simple constant-velocity approach. This questions if they actually allow to model social interactions as well as the validity of the proof. In this paper, we focus on the deep-learning models with a soft-attention mechanism for social interaction modeling and study whether they use social information at prediction time. We conduct two experiments across four state-of-the-art approaches on the ETH and UCY datasets, which were also used in previous work. First, the models are trained by replacing the social information with random noise and compared to model trained with actual social information. Second, we use a gating mechanism along with a $L_0$ penalty, allowing models to shut down their inner components. The models consistently learn to prune their soft-attention mechanism. For both experiments, neither the course of the convergence nor the prediction performance were altered. This demonstrates that the soft-attention mechanism and therefore the social information are ignored by the models., Comment: This work has been submitted to the IEEE for possible publication

Distance metric learning algorithms aim to appropriately measure similarities and distances between data points. In the context of clustering, metric learning is typically applied with the assist of side-information provided by experts, most commonly expressed in the form of cannot-link and must-link constraints. In this setting, distance metric learning algorithms move closer pairs of data points involved in must-link constraints, while pairs of points involved in cannot-link constraints are moved away from each other. For these algorithms to be effective, it is important to use a distance metric that matches the expert knowledge, beliefs, and expectations, and the transformations made to stick to the side-information should preserve geometrical properties of the dataset. Also, it is interesting to filter the constraints provided by the experts to keep only the most useful and reject those that can harm the clustering process. To address these issues, we propose to exploit the dual information associated with the pairwise constraints of the semi-supervised clustering problem. Experiments clearly show that distance metric learning algorithms benefit from integrating this dual information., Comment: 25 pages, 6 figures

Recurrent neural networks are effective models to process sequences. However, they are unable to learn long-term dependencies because of their inherent sequential nature. As a solution, Vaswani et al. introduced the Transformer, a model solely based on the attention mechanism that is able to relate any two positions of the input sequence, hence modelling arbitrary long dependencies. The Transformer has improved the state-of-the-art across numerous sequence modelling tasks. However, its effectiveness comes at the expense of a quadratic computational and memory complexity with respect to the sequence length, hindering its adoption. Fortunately, the deep learning community has always been interested in improving the models' efficiency, leading to a plethora of solutions such as parameter sharing, pruning, mixed-precision, and knowledge distillation. Recently, researchers have directly addressed the Transformer's limitation by designing lower-complexity alternatives such as the Longformer, Reformer, Linformer, and Performer. However, due to the wide range of solutions, it has become challenging for researchers and practitioners to determine which methods to apply in practice in order to meet the desired trade-off between capacity, computation, and memory. This survey addresses this issue by investigating popular approaches to make Transformers faster and lighter and by providing a comprehensive explanation of the methods' strengths, limitations, and underlying assumptions., Comment: ACM Computing Surveys; 40 pages, 18 figures, 4 tables

Robots and autonomous systems need to know where they are within a map to navigate effectively. Thus, simultaneous localization and mapping or SLAM is a common building block of robot navigation systems. When building a map via a SLAM system, robots need to re-recognize places to find loop closure and reduce the odometry drift. Image-based place recognition received a lot of attention in computer vision, and in this work, we investigate how such approaches can be used for 3D LiDAR data. Recent LiDAR sensors produce high-resolution 3D scans in combination with comparably stable intensity measurements. Through a cylindrical projection, we can turn this information into a panoramic image. As a result, we can apply techniques from visual place recognition to LiDAR intensity data. The question of how well this approach works in practice has not been answered so far. This paper provides an analysis of how such visual techniques can be with LiDAR data, and we provide an evaluation on different datasets. Our results suggest that this form of place recognition is possible and an effective means for determining loop closures., Comment: 7 pages, 6 figures

Drones have been getting more and more popular in many economy sectors. Both scientific and industrial communities aim at making the impact of drones even more disruptive by empowering collaborative autonomous behaviors -- also known as swarming behaviors -- within fleets of multiple drones. In swarming-powered 3D mapping missions, unmanned aerial vehicles typically collect the aerial pictures of the target area whereas the 3D reconstruction process is performed in a centralized manner. However, such approaches do not leverage computational and storage resources from the swarm members.We address the optimization of a swarm-powered distributed 3D mapping mission for a real-life humanitarian emergency response application through the exploitation of a swarm-powered ad hoc cloud. Producing the relevant 3D maps in a timely manner, even when the cloud connectivity is not available, is crucial to increase the chances of success of the operation. In this work, we present a mathematical programming heuristic based on decomposition and a variable neighborhood search heuristic to minimize the completion time of the 3D reconstruction process necessary in such missions. Our computational results reveal that the proposed heuristics either quickly reach optimality or improve the best known solutions for almost all tested realistic instances comprising up to 1000 images and fifteen drones.

Kernel traces are sequences of low-level events comprising a name and multiple arguments, including a timestamp, a process id, and a return value, depending on the event. Their analysis helps uncover intrusions, identify bugs, and find latency causes. However, their effectiveness is hindered by omitting the event arguments. To remedy this limitation, we introduce a general approach to learning a representation of the event names along with their arguments using both embedding and encoding. The proposed method is readily applicable to most neural networks and is task-agnostic. The benefit is quantified by conducting an ablation study on three groups of arguments: call-related, process-related, and time-related. Experiments were conducted on a novel web request dataset and validated on a second dataset collected on pre-production servers by Ciena, our partnering company. By leveraging additional information, we were able to increase the performance of two widely-used neural networks, an LSTM and a Transformer, by up to 11.3% on two unsupervised language modelling tasks. Such tasks may be used to detect anomalies, pre-train neural networks to improve their performance, and extract a contextual representation of the events., Comment: 11 pages, 11 figures, IEEE/ACM MSR 2021

The execution of similar units can be compared by their internal behaviors to determine the causes of their potential performance issues. For instance, by examining the internal behaviors of different fast or slow web requests more closely and by clustering and comparing their internal executions, one can determine what causes some requests to run slowly or behave in unexpected ways. In this paper, we propose a method of extracting the internal behavior of web requests as well as introduce a pipeline that detects performance issues in web requests and provides insights into their root causes. First, low-level and fine-grained information regarding each request is gathered by tracing both the user space and the kernel space. Second, further information is extracted and fed into an outlier detector. Finally, these outliers are then clustered by their behavior, and each group is analyzed separately. Experiments revealed that this pipeline is indeed able to detect slow web requests and provide additional insights into their true root causes. Notably, we were able to identify a real PHP cache contention using the proposed approach., Comment: 8 pages, 7 figures, IEEE ISSREW 2019

In order to process efficiently ever-higher dimensional data such as images, sentences, or audio recordings, one needs to find a proper way to reduce the dimensionality of such data. In this regard, SVD-based methods including PCA and Isomap have been extensively used. Recently, a neural network alternative called autoencoder has been proposed and is often preferred for its higher flexibility. This work aims to show that PCA is still a relevant technique for dimensionality reduction in the context of classification. To this purpose, we evaluated the performance of PCA compared to Isomap, a deep autoencoder, and a variational autoencoder. Experiments were conducted on three commonly used image datasets: MNIST, Fashion-MNIST, and CIFAR-10. The four different dimensionality reduction techniques were separately employed on each dataset to project data into a low-dimensional space. Then a k-NN classifier was trained on each projection with a cross-validated random search over the number of neighbours. Interestingly, our experiments revealed that k-NN achieved comparable accuracy on PCA and both autoencoders' projections provided a big enough dimension. However, PCA computation time was two orders of magnitude faster than its neural network counterparts., Comment: 4 pages, 4 figures, IEEE AIKE 2019

The popularity of drones is rapidly increasing across the different sectors of the economy. Aerial capabilities and relatively low costs make drones the perfect solution to improve the efficiency of those operations that are typically carried out by humans (e.g., building inspection, photo collection). The potential of drone applications can be pushed even further when they are operated in fleets and in a fully autonomous manner, acting de facto as a drone swarm. Besides automating field operations, a drone swarm can serve as an ad-hoc cloud infrastructure built on top of computing and storage resources available across the swarm members and other connected elements. Even in the absence of Internet connectivity, this cloud can serve the workloads generated by the swarm members themselves, as well as by the field agents operating within the area of interest. By considering the practical example of a swarm-powered 3D reconstruction application, we present a new optimization problem for the efficient generation and execution, on top of swarm-powered ad-hoc cloud infrastructure, of multi-node computing workloads subject to data geolocation and clustering constraints. The objective is the minimization of the overall computing times, including both networking delays caused by the inter-drone data transmission and computation delays. We prove that the problem is NP-hard and present two combinatorial formulations to model it. Computational results on the solution of the formulations show that one of them can be used to solve, within the configured time-limit, more than 50% of the considered real-world instances involving up to two hundred images and six drones.

Nowadays, SLAM (Simultaneous Localization and Mapping) is considered by the Robotics community to be a mature field. Currently, there are many open-source systems that are able to deliver fast and accurate estimation in typical real-world scenarios. Still, all these systems often provide an ad-hoc implementation that entailed to predefined sensor configurations. In this work, we tackle this issue, proposing a novel SLAM architecture specifically designed to address heterogeneous sensors' configuration and to standardize SLAM solutions. Thanks to its modularity and to specific design patterns, the presented architecture is easy to extend, enhancing code reuse and efficiency. Finally, adopting our solution, we conducted comparative experiments for a variety of sensor configurations, showing competitive results that confirm state-of-the-art performance.

Nowadays, Non-Linear Least-Squares embodies the foundation of many Robotics and Computer Vision systems. The research community deeply investigated this topic in the last years, and this resulted in the development of several open-source solvers to approach constantly increasing classes of problems. In this work, we propose a unified methodology to design and develop efficient Least-Squares Optimization algorithms, focusing on the structures and patterns of each specific domain. Furthermore, we present a novel open-source optimization system, that addresses transparently problems with a different structure and designed to be easy to extend. The system is written in modern C++ and can run efficiently on embedded systems. Source code: https://srrg.gitlab.io/srrg2-solver.html. We validated our approach by conducting comparative experiments on several problems using standard datasets. The results show that our system achieves state-of-the-art performances in all tested scenarios., Comment: 29 pages, 15 figures, source code at https://srrg.gitlab.io/srrg2-solver.html