arXiv subjects

Yon Dohn Chung

Publications and source records attributed to Yon Dohn Chung.

7 recordsLinked to original sources

arXiv2026

ABC: Numerical Data Collection under Local Differential Privacy without Prior Knowledge

Local Differential Privacy (LDP) provides strong privacy guarantees for collecting numerical data. A fundamental challenge, however, is that existing LDP mechanisms require a predefined data domain, which is often unknown in practice. This lack of prior knowledge creates a critical dilemma for the data collector: if the chosen domain is too narrow, values outside the range are clipped, leading to information loss. Conversely, if the domain is too wide, excessive noise is added during the privatization process, which degrades the quality of collected data. This highlights the need for methods that can dynamically estimate the data domain. In this work, we propose an adaptive LDP framework that addresses this problem. In our method, each user sends two pieces of information: their perturbed numerical data, and a privatized signal indicating if their original value was clipped by the current domain. By aggregating these signals, our proposed method, Adaptive Bounding of Clipping regions (ABC) method, iteratively adjusts the domain to fit the underlying data distribution without prior knowledge. Our theoretical analysis shows that the estimated data domain converges to an appropriate range. In the empirical evaluation, the results demonstrate that our framework significantly improves the quality of numerical data collection across various datasets and underlying LDP mechanisms. We also show that the estimated range successfully converges in practice and our approach is robust to its hyperparameters through comprehensive ablation studies.

cs.CR↗

arXiv2025

Federated Learning with Feedback Alignment

Federated Learning (FL) enables collaborative training across multiple clients while preserving data privacy, yet it struggles with data heterogeneity, where clients' data are not distributed independently and identically (non-IID). This causes local drift, hindering global model convergence. To address this, we introduce Federated Learning with Feedback Alignment (FLFA), a novel framework that integrates feedback alignment into FL. FLFA uses the global model's weights as a shared feedback matrix during local training's backward pass, aligning local updates with the global model efficiently. This approach mitigates local drift with minimal additional computational cost and no extra communication overhead. Our theoretical analysis supports FLFA's design by showing how it alleviates local drift and demonstrates robust convergence for both local and global models. Empirical evaluations, including accuracy comparisons and measurements of local drift, further illustrate that FLFA can enhance other FL methods demonstrating its effectiveness.

cs.LG↗

arXiv2025

N-output Mechanism: Estimating Statistical Information from Numerical Data under Local Differential Privacy

Local Differential Privacy (LDP) addresses significant privacy concerns in sensitive data collection. In this work, we focus on numerical data collection under LDP, targeting a significant gap in the literature: existing LDP mechanisms are optimized for either a very small ($|\Omega| \in \{2, 3\}$) or infinite output spaces. However, no generalized method for constructing an optimal mechanism for an arbitrary output size $N$ exists. To fill this gap, we propose the \textbf{N-output mechanism}, a generalized framework that maps numerical data to one of $N$ discrete outputs. We formulate the mechanism's design as an optimization problem to minimize estimation variance for any given $N \geq 2$ and develop both numerical and analytical solutions. This results in a mechanism that is highly accurate and adaptive, as its design is determined by solving an optimization problem for any chosen $N$. Furthermore, we extend our framework and existing mechanisms to the task of distribution estimation. Empirical evaluations show that the N-output mechanism achieves state-of-the-art accuracy for mean, variance, and distribution estimation with small communication costs.

cs.CR↗

arXiv2025

Decoupled Contrastive Learning for Federated Learning

Federated learning is a distributed machine learning paradigm that allows multiple participants to train a shared model by exchanging model updates instead of their raw data. However, its performance is degraded compared to centralized approaches due to data heterogeneity across clients. While contrastive learning has emerged as a promising approach to mitigate this, our theoretical analysis reveals a fundamental conflict: its asymptotic assumptions of an infinite number of negative samples are violated in finite-sample regime of federated learning. To address this issue, we introduce Decoupled Contrastive Learning for Federated Learning (DCFL), a novel framework that decouples the existing contrastive loss into two objectives. Decoupling the loss into its alignment and uniformity components enables the independent calibration of the attraction and repulsion forces without relying on the asymptotic assumptions. This strategy provides a contrastive learning method suitable for federated learning environments where each client has a small amount of data. Our experimental results show that DCFL achieves stronger alignment between positive samples and greater uniformity between negative samples compared to existing contrastive learning methods. Furthermore, experimental results on standard benchmarks, including CIFAR-10, CIFAR-100, and Tiny-ImageNet, demonstrate that DCFL consistently outperforms state-of-the-art federated learning methods.

cs.LG↗

arXiv2022

Multi-Dimensional Data Compression and Query Processing in Array Databases

In recent times, the production of multidimensional data in various domains and their storage in array databases has witnessed a sharp increase; this rapid growth in data volumes necessitates compression in array databases. However, existing compression schemes used in array databases are general-purpose and not designed specifically for the databases. They could degrade query performance with complex analytical tasks, which incur huge computing costs. Thus, a compression scheme that considers the workflow of array databases is required. This study presents a compression scheme, SEACOW, for storing and querying multidimensional array data. The scheme is specially designed to be efficient for both dimension-based and value-based exploration. It considers data access patterns for exploration queries and embeds a synopsis, which can be utilized as an index, in the compressed array. In addition, we implement an array storage system, namely MSDB, to perform experiments. We evaluate query performance on real scientific datasets and compared it with those of existing compression schemes. Finally, our experiments demonstrate that SEACOW provides high compression rates compared to existing compression schemes, and the synopsis improves analytical query processing performance.

cs.DB↗

arXiv2022

Multi-object Tracking with an Adaptive Generalized Labeled Multi-Bernoulli Filter

The challenges in multi-object tracking mainly stem from the random variations in the cardinality and states of objects during the tracking process. Further, the information on locations where the objects appear, their detection probabilities, and the statistics of the sensor's false alarms significantly influence the tracking accuracy of the filter. However, this information is usually assumed to be known and provided by the users. In this paper, we propose an adaptive generalized labeled multi-Bernoulli (GLMB) filter which can track multiple objects without prior knowledge of the aforementioned information. Experimental results show that the performance of the proposed filter is comparable to an ideal GLMB filter supplied with correct information of the tracking scenarios.

eess.SP↗

arXiv2021

Successive Point-of-Interest Recommendation with Local Differential Privacy

A point-of-interest (POI) recommendation system performs an important role in location-based services because it can help people to explore new locations and promote advertisers to launch advertisements at appropriate locations. The existing POI recommendation systems require raw check-in history of users, which might cause location privacy violations. Although there have been several matrix factorization (MF) based privacy-preserving recommendation systems, they can only focus on user-POI relationships without considering the human movements in check-in history. To tackle this problem, we design a successive POI recommendation framework with local differential privacy, named SPIREL. SPIREL uses two types of information derived from the check-in history as input for the factorization: a transition pattern between two POIs and the visit counts of POIs. We propose a novel objective function for learning the user-POI and POI-POI relationships simultaneously. We further integrate local differential privacy mechanisms in our proposed framework to prevent potential location privacy breaches. Experiments using four public datasets demonstrate that SPIREL achieves better POI recommendation quality while accomplishing stronger privacy protection.

cs.IR↗