1. Nearest neighbor permutation entropy detects phase transitions in complex high-pressure systemsArthur A. B. Pessa, Leonardo G. J. M. Voltarelli, Lúcio Cardozo-Filho, Andres G. M. Tamara, Claudio Dariva, Papa M. Ndiaye, Matjaž Perc, Haroldo V. Ribeiro, 2025, izvirni znanstveni članek Opis: Understanding the high-pressure phase behavior of carbon dioxide-hydrocarbon mixtures is of considerable interest owing to their wide range of applications. Under certain conditions, these systems are not amenable to direct visual monitoring, and experimentalists often rely on spectrophotometric data to infer phase behavior. Consequently, developing computationally efficient and robust methods to leverage such data is crucial. Here, we combine nearest neighbor permutation entropy, computed directly from in situ near-infrared absorbance spectra acquired during depressurization trials of mixtures of carbon dioxide and a distilled petroleum fraction, with an anomaly detection approach to identify phase transitions. We show that changes in nearest neighbor entropy effectively signal transitions from initially homogeneous mixtures to two-phase equilibria, thereby enabling accurate out-of-sample online predictions of transition pressures. Our approach requires minimum data preprocessing, no specialized detection techniques or visual inspection of the spectra, and is sufficiently general to be adapted for studying phase behavior in other high-pressure systems monitored via spectrophotometry.
Ključne besede: nearest neighbors, permutation entropy, phase transition, complex system, social physics
Objavljeno v DKUM: 04.07.2025; Ogledov: 0; Prenosov: 2
 Celotno besedilo (3,19 MB)
Gradivo ima več datotek! Več... |
2. LLM in the loop: a framework for contextualizing counterfactual segment perturbations in point cloudsVeljka Kočić, Niko Lukač, Dzemail Rozajac, Stefan Schweng, Christoph Gollob, Arne Nothdurft, Karl Stampfer, Javier Del Ser, Andreas Holzinger, 2025, izvirni znanstveni članek Opis: Point Cloud Data analysis has seen a major leap forward with the introduction of PointNet algorithms, revolutionizing how we process 3D environments. Yet, despite these advancements, key challenges remain, particularly in optimizing segment perturbations to influence model outcomes in a controlled and meaningful way. Traditional methods struggle to generate realistic and contextually appropriate perturbations, limiting their effectiveness in critical applications like autonomous systems and urban planning. This paper takes a bold step by integrating Large Language Models into the counterfactual reasoning process, unlocking a new level of automation and intelligence in segment perturbation. Our approach begins with semantic segmentation, after which LLMs intelligently select optimal replacement segments based on features such as class label, color, area, and height. By leveraging the reasoning capabilities of LLMs, we generate perturbations that are not only computationally efficient but also semantically meaningful. The proposed framework undergoes rigorous evaluation, combining human inspection of LLM-generated suggestions with quantitative analysis of semantic classification model performance across different LLM variants. By bridging the gap between geometric transformations and high-level semantic reasoning, this research redefines how we approach perturbation generation in Point Cloud Data analysis. The results pave the way for more interpretable, adaptable, and intelligent AI-driven solutions, bringing us closer to realworld applications where explainability and robustness are paramount.
Ključne besede: explainable AI, point cloud data, counterfactual reasoning, LiDAR, 3D point cloud data, interpretability, human-centered AI, large language models, K-nearest neighbors
Objavljeno v DKUM: 19.05.2025; Ogledov: 0; Prenosov: 3
Celotno besedilo (7,24 MB) |
