arXiv subjects

Bu Sung Lee

Publications and source records attributed to Bu Sung Lee.

8 recordsLinked to original sources

arXiv2025

On Competitiveness of Dynamic Replication for Distributed Data Access

This paper studies an online cost optimization problem for distributed storage and access. The goal is to dynamically create and delete copies of data objects over time at geo-distributed servers to serve access requests and minimize the total storage and network cost. We revisit a recent algorithm in the literature and show that it does not have a competitive ratio of $2$ as claimed by constructing a counterexample. We further prove that no deterministic online algorithm can achieve a competitive ratio bounded by $2$ for the general cost optimization problem. We develop an online algorithm and prove that it achieves a competitive ratio of $\max\{2, \min\{γ, 3\}\}$, where $γ$ is the max/min storage cost ratio among all servers. Examples are given to confirm the tightness of competitive analysis. We also empirically evaluate algorithms using real object access traces.

cs.DS↗

arXiv2024

FakeBench: Probing Explainable Fake Image Detection via Large Multimodal Models

The ability to distinguish whether an image is generated by artificial intelligence (AI) is a crucial ingredient in human intelligence, usually accompanied by a complex and dialectical forensic and reasoning process. However, current fake image detection models and databases focus on binary classification without understandable explanations for the general populace. This weakens the credibility of authenticity judgment and may conceal potential model biases. Meanwhile, large multimodal models (LMMs) have exhibited immense visual-text capabilities on various tasks, bringing the potential for explainable fake image detection. Therefore, we pioneer the probe of LMMs for explainable fake image detection by presenting a multimodal database encompassing textual authenticity descriptions, the FakeBench. For construction, we first introduce a fine-grained taxonomy of generative visual forgery concerning human perception, based on which we collect forgery descriptions in human natural language with a human-in-the-loop strategy. FakeBench examines LMMs with four evaluation criteria: detection, reasoning, interpretation and fine-grained forgery analysis, to obtain deeper insights into image authenticity-relevant capabilities. Experiments on various LMMs confirm their merits and demerits in different aspects of fake image detection tasks. This research presents a paradigm shift towards transparency for the fake image detection area and reveals the need for greater emphasis on forensic elements in visual-language research and AI risk control. FakeBench will be available at https://github.com/Yixuan423/FakeBench.

cs.CV↗

arXiv2024

Cost-Driven Data Replication with Predictions

This paper studies an online replication problem for distributed data access. The goal is to dynamically create and delete data copies in a multi-server system as time passes to minimize the total storage and network cost of serving access requests. We study the problem in the emergent learning-augmented setting, assuming simple binary predictions about inter-request times at individual servers. We develop an online algorithm and prove that it is ($\frac{5+α}{3}$)-consistent (competitiveness under perfect predictions) and ($1 + \frac{1}α$)-robust (competitiveness under terrible predictions), where $α\in (0, 1]$ is a hyper-parameter representing the level of distrust in the predictions. We also study the impact of mispredictions on the competitive ratio of the proposed algorithm and adapt it to achieve a bounded robustness while retaining its consistency. We further establish a lower bound of $\frac{3}{2}$ on the consistency of any deterministic learning-augmented algorithm. Experimental evaluations are carried out to evaluate our algorithms using real data access traces.

cs.DS↗

arXiv2012

A Self-Organization Framework for Wireless Ad Hoc Networks as Small Worlds

Motivated by the benefits of small world networks, we propose a self-organization framework for wireless ad hoc networks. We investigate the use of directional beamforming for creating long-range short cuts between nodes. Using simulation results for randomized beamforming as a guideline, we identify crucial design issues for algorithm design. Our results show that, while significant path length reduction is achievable, this is accompanied by the problem of asymmetric paths between nodes. Subsequently, we propose a distributed algorithm for small world creation that achieves path length reduction while maintaining connectivity. We define a new centrality measure that estimates the structural importance of nodes based on traffic flow in the network, which is used to identify the optimum nodes for beamforming. We show, using simulations, that this leads to significant reduction in path length while maintaining connectivity.

cs.NI↗

arXiv2012

Achieving Small World Properties using Bio-Inspired Techniques in Wireless Networks

It is highly desirable and challenging for a wireless ad hoc network to have self-organization properties in order to achieve network wide characteristics. Studies have shown that Small World properties, primarily low average path length and high clustering coefficient, are desired properties for networks in general. However, due to the spatial nature of the wireless networks, achieving small world properties remains highly challenging. Studies also show that, wireless ad hoc networks with small world properties show a degree distribution that lies between geometric and power law. In this paper, we show that in a wireless ad hoc network with non-uniform node density with only local information, we can significantly reduce the average path length and retain the clustering coefficient. To achieve our goal, our algorithm first identifies logical regions using Lateral Inhibition technique, then identifies the nodes that beamform and finally the beam properties using Flocking. We use Lateral Inhibition and Flocking because they enable us to use local state information as opposed to other techniques. We support our work with simulation results and analysis, which show that a reduction of up to 40% can be achieved for a high-density network. We also show the effect of hopcount used to create regions on average path length, clustering coefficient and connectivity.

cs.NI↗

arXiv2011

Self-organization of Nodes using Bio-Inspired Techniques for Achieving Small World Properties

In an autonomous wireless sensor network, self-organization of the nodes is essential to achieve network wide characteristics. We believe that connectivity in wireless autonomous networks can be increased and overall average path length can be reduced by using beamforming and bio-inspired algorithms. Recent works on the use of beamforming in wireless networks mostly assume the knowledge of the network in aggregation to either heterogeneous or hybrid deployment. We propose that without the global knowledge or the introduction of any special feature, the average path length can be reduced with the help of inspirations from the nature and simple interactions between neighboring nodes. Our algorithm also reduces the number of disconnected components within the network. Our results show that reduction in the average path length and the number of disconnected components can be achieved using very simple local rules and without the full network knowledge.

cs.NI↗

arXiv2011

Self-Organization of Wireless Ad Hoc Networks as Small Worlds Using Long Range Directional Beams

We study how long range directional beams can be used for self-organization of a wireless network to exhibit small world properties. Using simulation results for randomized beamforming as a guideline, we identify crucial design issues for algorithm design. Subsequently, we propose an algorithm for deterministic creation of small worlds. We define a new centrality measure that estimates the structural importance of nodes based on traffic flow in the network, which is used to identify the optimum nodes for beamforming. This results in significant reduction in path length while maintaining connectivity.

cs.NI↗

arXiv2011

From Linked Data to Relevant Data -- Time is the Essence

The Semantic Web initiative puts emphasis not primarily on putting data on the Web, but rather on creating links in a way that both humans and machines can explore the Web of data. When such users access the Web, they leave a trail as Web servers maintain a history of requests. Web usage mining approaches have been studied since the beginning of the Web given the log's huge potential for purposes such as resource annotation, personalization, forecasting etc. However, the impact of any such efforts has not really gone beyond generating statistics detailing who, when, and how Web pages maintained by a Web server were visited.

cs.IR↗