Abstract

Abstract
High-performance computing is pivotal for processing large datasets and executing complex simulations, ensuring faster and more accurate results. Improving the performance of software and scientific workflows in such environments requires careful analysis of their computational behavior and energy consumption. Therefore, maximizing computational throughput in these environments, through adequate software configuration and resource allocation, is essential for improving performance. The work presented in this paper focuses on leveraging regression-based machine learning and decision trees to analyze and optimize resource allocation in high-performance computing environments based on application's performance and energy metrics. Applied to a bioinformatics case study, these models enable informed decision-making by selecting the appropriate computing resources to enhance the performance of a phylogenomics software. Our contribution is to better explore and understand the efficient resource management of supercomputers, namely Santos Dumont. We show that the predictions for application's execution time using the proposed method are accurate for various amounts of computing nodes, while energy consumption predictions are less precise. The application parameters most relevant for this work are identified and the relative importance of each application parameter to the accuracy of the prediction is analysed.

Abstract
This paper presents an indoor benchmarking study of state-of-the-art 3D LiDAR-based Simultaneous Localization and Mapping (SLAM) algorithms using the newly developed IILABS 3D-iilab Indoor LiDAR-based SLAM 3D dataset. Existing SLAM datasets often focus on outdoor environments, rely on a single type of LiDAR sensor, or lack additional sensor data such as wheel odometry in ground-based robotic platforms. Consequently, the existing datasets lack data diversity required to comprehensively evaluate performance under diverse indoor conditions. The IILABS 3D dataset fills this gap by providing a sensor-rich, indoor-exclusive dataset recorded in a controlled laboratory environment using a wheeled mobile robot platform. It includes four heterogeneous 3D LiDAR sensors-Velodyne VLP-16, Ouster OS1-64, RoboSense RS-Helios-5515, and Livox Mid-360-featuring both mechanical spinning and non-repetitive scanning patterns, as well as an IMU and wheel odometry for sensor fusion. The dataset also contains calibration sequences, challenging benchmark trajectories, and high-precision ground-truth poses captured with a motion capture system. Using this dataset, we benchmark nine representative LiDAR-based SLAM algorithms across multiple sequences, analyzing their performance in terms of accuracy and consistency under varying sensor configurations. The results provide a comprehensive performance comparison and valuable insights into the strengths and limitations of current SLAM algorithms in indoor environments.

Abstract
Balancing supply and demand in free-floating one-way carsharing systems is a critical operational challenge. This paper presents a novel approach that integrates a binary logit model into a mixed integer linear programming framework to optimize short-term pricing and fleet relocation. Demand modeling, based on a binary logit model, aggregates different trips under a unified utility model and improves estimation by incorporating information from similar trips. To speed up the estimation process, a categorizing approach is used, where variables such as location and time are classified into a few categories based on shared attributes. This is particularly beneficial for trips with limited observations as information gained from similar trips can be used for these trips effectively. The modeling framework adopts a dynamic structure where the binary logit model estimates demand using accumulated observations from past iterations at each decision point. This continuous learning environment allows for dynamic improvement in estimation and decision-making. At the core of the framework is a mathematical program that prescribes optimal levels of promotion and relocation. The framework then includes simulated market responses to the decisions, allowing for real-time adjustments to effectively balance supply and demand. Computational experiments demonstrate the effectiveness of the proposed approach and highlight its potential for real-world applications. The continuous learning environment, combining demand modeling and operational decisions, opens avenues for future research in transportation systems.

Abstract
Audio-visual correlation learning aims at capturing and understanding natural phenomena between audio and visual data. The rapid growth of Deep Learning propelled the development of proposals that process audio-visual data and can be observed in the number of proposals in the past years. Thus encouraging the development of a comprehensive survey. Besides analyzing the models used in this context, we also discuss some tasks of definition and paradigm applied in AI multimedia. In addition, we investigate objective functions frequently used and discuss how audio-visual data is exploited in the optimization process, i.e., the different methodologies for representing knowledge in the audio-visual domain. In fact, we focus on how human-understandable mechanisms, i.e., structured knowledge that reflects comprehensible knowledge, can guide the learning process. Most importantly, we provide a summarization of the recent progress of Audio-Visual Correlation Learning (AVCL) and discuss the future research directions.

Abstract