67639 research outputs found
Sort by
The Looming Privacy Challenges posed by Commercial Satellite Imaging: Remedies and Research Directions
No full textAs commercial satellites proliferate and data-capture technologies become more diverse, accessible, and high-resolution, new privacy concerns emerge. Additionally, emerging trends such as direct-to-consumer image sales and AI-enhanced computer vision strongly impact the satellite imaging industry. While these technologies benefit applications such as natural disaster monitoring and precision agriculture, they also introduce novel privacy risks. This perspective paper provides several contributions: we start by analyzing privacy risks using the LINDDUN privacy framework, categorizing threats like linking, identification, and data disclosure. Moving on from these findings, we introduce and discuss a few mitigation strategies, including inpainting and access control. Finally, we highlight some research directions to address these emerging challenges. Overall, this work characterizes the major issues in the evolving privacy landscape of satellite imaging, provides actionable solutions to mitigate these privacy risks, and highlights future research directions, positioning itself as a reference for researchers, policymakers, and industry stakeholders
Operando unveiling the activity origin via preferential structural evolution in Ni-Fe (oxy)phosphides for efficient oxygen evolution
No full textNon-noble metal-based heteroatom compounds demonstrate excellent electrocatalytic activity for the oxygen evolution reaction (OER). However, the origin of this activity, driven by structure evolution effects, remains unclear due to the lack of effective in situ/operando techniques. Herein, we employ the operando quick-scan x-ray absorption fine structure (Q-XAFS) technique coupled with in situ controlled electrochemical potential to establish a structure-activity correlation of the OER catalyst. Using Ni-Fe bimetallic phosphides as a model catalyst, operando Q-XAFS experiments reveal that the structural transformation initiates at the preferential oxidation of Fe sites over Ni sites. The in situ–generated O-Fe-P structure serves as the origin of the enhanced electrocatalytic OER activity of the catalyst, a finding supported by theoretical calculations. This work provides crucial insights into understanding the reaction mechanism of the state-of-the-art Ni-Fe–based OER electrocatalysts, thus advancing the rational design of more efficient OER electrocatalysts.X.W.L. acknowledges the funding support for the Global STEM Professorship from the Innovation, Technology and Industry Bureau (“ITIB”) and Education Bureau (“EDB”) of Hong Kong
Acoustic cavitation phenomena: understanding cavitation inception and bubble dynamics in ultrasonic reactors
No full textCavitation induced by ultrasound and acoustic waves is a physical phenomenon widely used nowadays in many sectors, from sonochemistry and process intensification to biomedical applications such as histotripsy. Given their wide adoption, acoustic transducers and ultrasonic reactors operate at various frequencies and on different fluid mixtures. However, considering the large number of variables, the optimal combination of parameters for a specific application is often unknown.
This dissertation aims to clarify aspects of acoustic cavitation phenomena that are essential for optimizing the design of acoustic transducers and ultrasonic reactors. One key parameter required to calibrate and operate an acoustic system is the cavitation threshold. The thesis presents a study on cavitation inception induced by the propagation of 24 kHz acoustic waves in water. A rigorous and repeatable methodology, based on high-speed imaging and hydrophone measurements, is introduced to detect the onset of cavitation in acoustic systems. The tensile strength of the liquid is estimated by applying the analytical solution of the Rayleigh integral to the data extracted from high-speed sequences. From the results, the frequency arises as the most relevant parameter determining the inception of cavitation bubbles in a continuous medium. This introduces the need for a novel cavitation number. The framework was then extended, through an acoustic analogy, to a case of impulsive cavitation.
Cavitation in different fluids was also investigated to confirm the findings. Despite their different physical properties, all the liquid media exhibit a similar behavior in terms of cavitation threshold and expansion velocity of the nucleated bubble. This sheds light on the nucleation mechanism responsible for initiating cavitation in the lab-scale acoustic system.
The extent of cavitation leveraging sonochemistry was then explored through coumarin dosimetry and high-speed imaging. Key findings indicate that the chemical efficiency strongly depends on the dynamics and structure of the vapor field.
Finally, experiments were conducted with a calibrated Schlieren technique to study fluid dynamic effects. Thanks to calibration and tomographic reconstruction procedures, the pressure fields induced by acoustic and shock waves were quantitatively estimated. The methodologies developed in this study are applicable for analyzing cavitation phenomena characterized by other frequency ranges
Mutual Coupling-Aware Channel Estimation and Beamforming for RIS-Assisted Communications
No full textThis work studies the problems of channel estimation and beamforming for active reconfigurable intelligent surface (RIS)-assisted multiple-input multiple-output (MIMO) communication, incorporating the mutual coupling (MC) effect through an electromagnetically consistent model. We first demonstrate that MC can be incorporated into a compressed sensing (CS) formulation, albeit with an increase in the dimensionality of the sensing matrix. To overcome this increased complexity, we propose a two-stage strategy. Initially, a low-complexity MC-unaware CS estimation is performed to obtain a coarse channel estimate, which is then used to implement a dictionary reduction (DR) for the MC-aware estimation, effectively reducing the dimensionality of the sensing matrices. This method achieves estimation accuracy close to the direct MC-aware CS method with less overall computational complexity. Furthermore, we consider the joint optimization of RIS configuration, base station precoding, and user combining in a single-user MIMO system. We employ an alternating optimization strategy to optimize these three beamformers. The primary challenge lies in optimizing the RIS configuration, as the MC effect renders the problem non-convex and intractable. To address this, we propose a novel algorithm based on the successive convex approximation (SCA) and the Neumann series expansion. Within the SCA framework, we propose a surrogate function that rigorously satisfies both convexity and equal-gradient conditions to update the iteration direction. Numerical results validate our proposal, demonstrating that the proposed channel estimation and beamforming methods effectively manage the MC in RIS, achieving higher spectral efficiency compared to state-of-the-art approaches
Investigation of Di-tert-butyl peroxide combustion: time-resolved speciation, laminar flame speed, and model evaluation
No full textDi‑tert‑butyl peroxide (DTBP), a cetane improver, consists of two tert‑butoxy groups bonded by a weak peroxide bond. A thorough understanding of the combustion mechanisms of DTBP is essential for its effective use as a fuel additive. This study systematically investigates the pyrolysis and oxidation characteristics of DTBP through a combination of experimental measurements and kinetic modeling. Laser absorption spectroscopy was employed to achieve time-resolved quantification of key species, including CO, CO2, OH, and H2O, during DTBP pyrolysis and oxidation under conditions spanning 1265 - 2000 K and 1.1 - 1.6 bar. The laminar flame speed of DTBP was measured for the first time over a range of equivalence ratios (0.65 - 1.4), initial pressures (0.5 - 2 bar), and a fixed initial temperature of 373 ± 3 K. These experimental results provide essential constraints for optimizing and validating the DTBP kinetic model. The proposed model significantly improved the accuracy of predictions for ignition delay times and laminar flame speeds. Furthermore, the model demonstrated excellent performance in capturing the second-stage ignition delay, overcoming the limitations of previous models. However, the current model still exhibits noticeable deviations in predictions below 1300 K, particularly in the formation dynamics of key species such as CO. To improve the model's accuracy under these conditions, further high-fidelity quantum chemical calculations are needed to refine the rate constants of additional unimolecular decomposition and hydrogen abstraction pathways of DTBP. Overall, by incorporating the latest sub-mechanism of acetone oxidation and updated rate constants for DTBP decomposition reactions, this study provides valuable experimental data and kinetic insights to advance combustion kinetic models for oxygenated fuels and support the development of high-efficiency combustion technologies.The work at XJTU was supported by the Sichuan Gas Turbine Establishment of AECC program (Grant no KTZX-030120210007) and National Science and Technology Major Project (J2019-III-0004–0047). The work of KAUST authors was funded by the FLEET consortium. Congjie Hong would like to thank the support of the China Scholarship Council (CSC202306280221) and all the group members at Propulsion ThermoChemistry Lab
Vgent: Graph-based Retrieval-Reasoning-Augmented Generation For Long Video Understanding
No full textUnderstanding and reasoning over long videos pose significant challenges for large video language models (LVLMs) due to the difficulty in processing intensive video tokens beyond context window and retaining long-term sequential information. Retrieval-Augmented Generation (RAG) has demonstrated effectiveness in processing long context for Large Language Models (LLMs); however, applying RAG to long video faces challenges such as disrupted temporal dependencies and inclusion of irrelevant information that can hinder accurate reasoning. To address these limitations, we propose Vgent, a novel graph-based retrieval-reasoning-augmented generation framework to enhance LVLMs for long video understanding. Our approach introduces two key innovations: (i) It represents videos by structured graphs with semantic relationships across video clips preserved to improve retrieval effectiveness. (ii) It introduces an intermediate reasoning step to mitigate the reasoning limitation of LVLMs, which leverages structured verification to reduce retrieval noise and facilitate the explicit aggregation of relevant information across clips, resulting in more accurate and context-aware responses. We comprehensively evaluate our framework with various open-source LVLMs on three long-video understanding benchmarks. Our approach yielded an overall performance improvement of over base models on MLVU, and outperformed state-of-the-art video RAG methods by . Our code is publicly available at https://xiaoqian-shen.github.io/Vgent
PSDMQ: A parallel method for shortest distances multi-querying on encrypted graph
No full textSearchable symmetric encryption (SSE) enables a client to outsource a set of encrypted data in the cloud and retain the ability to perform retrieval without revealing client's private information. Graph as an import data structure and shortest distance querying as a significant fundamental operations on graph, has been applied a lot in real-life. Although efficient SSE shortest distance querying on graph are known, previous solutions are highly sequential. This is mainly due to the fact that, currently, the method for querying shortest distance on encrypted is usually designed based on 2HCL, which requires the search algorithm to access a sequence of memory locations, each of which is unpredictable and stored at the previous location in the sequence. Motivated by advances in multi-core architectures, we construct a new parallized method for batch shortest distance multi-querying(PSDMQ) on encrypted graph-structured data. Our approach is highly parallel thus can significantly improve efficiency. Both theoretical analysis and experiments are provided. Experiments on 3 real-world data sets demonstrates the effectiveness of our proposed scheme.This work is supported by the Strategic Priority Research Program of the Chinese Academy of Sciences, Grant No. XDB0690303
Millennial-scale fingerprint of macroalgae in Arctic marine sediments.
No full textMacroalgae are the most widely distributed marine vegetated habitats and contribute to marine carbon cycling and storage but with limited empirical documentation of long-term burial. To evaluate long-term burial of macroalgal-derived carbon in Arctic sediments, we analyzed eDNA from six dated sediment cores from off the coast of West Greenland (79°N-60°N). We applied metabarcoding of 18S rRNA genes to selected sediment layers covering the past ∼2600 years, assessed spatio-temporal patterns of macroalgal taxa, and evaluated climatic drivers of macroalgal change using proxies for past sea surface conditions. Macroalgal DNA was present in all cores and 86.5 % of samples. Orders of brown algal (Laminariales, Fucales) were more prevalent than red or green orders and we found no consistent changes in the identity of macroalgal taxa, despite significant changes in sea surface conditions (temperature, sea ice, salinity) over this period. However, locally, we observed a significant decline in richness of macroalgal taxa buried over the last c. 800 years (Qeqertarsuup Tunua, Disko Bay). This decline coincided with the onset of the Little Ice Age and greater fluctuations in sea surface conditions. Overall, we demonstrate that macroalgae can be preserved for millennia in marine sediments along Greenland's west coast, documenting that macroalgae contribute to long-term carbon burial in the Arctic although their quantitative significance is still unknown.We are grateful to the captains, crew and scientific party of the following cruises: Porsild 1998, Dana2006, Galathea3 expedition 2006–2007, Sanna 2013, Amundsen 2014. We are also grateful to the Galathea3 organization committee. Many thanks also to Wajitha J. Raja Mohamed Sait (KAUST) for help with DNA extractions and to Jacques Giraudeau (University of Bordeaux), Karen Luise Knudsen (AU), Antoon Kuijpers (GEUS), and Guillaume Massé (CNRS) for access to core material. The research of this project has been funded through the CARMA project from the Independent Research Fund Denmark project no. 8021-00222 B, with additional funds through the DFF projects G-Ice and GreenShelf (DFF grants no. 7014-00113B and 0135-00165B) to MSS and the European Union's Horizon 2020 research and innovation program under Grant Agreement No. 869383 (ECOTIP) and the European Union's Horizon Europe research and innovation program under Grant Agreement No. 101136480 (SEA-Quester) and Grant Agreement No. 101136875 (POMP), baseline funding provided by King Abdullah University of Science and Technology (KAUST) to CMD, and Independent Research Fund Denmark (0217-00244B), the Research Council of Norway (NFR 324520) and the VILLUM Foundation (YIP 10100) to CSA
An Unbiased and Robust Privacy-Preserving Fingerprinting Scheme for Relational Databases
No full textSharing relational databases is essential in today’s data-driven world for fostering collaboration, enhancing efficiency, and enabling real-time data access. However, privacy and copyright issues arise when sharing privacy-sensitive or valuable data. Additionally, high utility is required in shared data to enable accurate data mining and analysis. Entry-level differentially private fingerprinting schemes (DPFS) could address these concerns. In a DPFS, data can be securely shared without leaking original values while still supporting accurate analysis. Moreover, detectable fingerprints can deter unauthorized redistribution. However, existing DPFSs often lack utility—due to format changes and entry-wise bias—or robustness, as fingerprints can be removed undetected. In this paper, we propose an unbiased and robust differential privacy-based fingerprinting scheme (DPFS), which ensures that the fingerprinted copy remains an unbiased estimate of the original data. By incorporating differential privacy noise, our scheme effectively mitigates alteration, collusion, and hybrid attacks. Our DPFS satisfies ϵ-entry-level differential privacy, enabling clients to conduct unbiased analysis. To improve robustness, we design group-based fingerprint detection, which estimates the mean of injected noise per group with error tolerance. We provide a theoretical robustness analysis and propose a method for achieving optimal robustness. Experiments on four real-world databases show that our scheme consistently detects fingerprints and improves accuracy by up to 20% on machine learning tasks compared to existing DPFSs
CCDC 2347185: Experimental Crystal Structure Determination : tetrakis(N,N-diethyl-N-methylethanaminium) heptakis(thiocyanato)-erbium(iii)
No full textAn entry from the Cambridge Structural Database, the world’s repository for small molecule crystal structures. The entry contains experimental data from a crystal diffraction study. The deposited dataset for this entry is freely available from the CCDC and typically includes 3D coordinates, cell parameters, space group, experimental conditions and quality measures