arXiv subjects

Taehwan Lee

Publications and source records attributed to Taehwan Lee.

3 recordsLinked to original sources

arXiv2025

Benchmarking Federated Learning for Semantic Datasets: Federated Scene Graph Generation

Federated learning (FL) enables decentralized training while preserving data privacy, yet existing FL benchmarks address relatively simple classification tasks, where each sample is annotated with a one-hot label. However, little attention has been paid to demonstrating an FL benchmark that handles complicated semantics, where each sample encompasses diverse semantic information, such as relations between objects. Because the existing benchmarks are designed to distribute data in a narrow view of a single semantic, managing the complicated semantic heterogeneity across clients when formalizing FL benchmarks is non-trivial. In this paper, we propose a benchmark process to establish an FL benchmark with controllable semantic heterogeneity across clients: two key steps are (i) data clustering with semantics and (ii) data distributing via controllable semantic heterogeneity across clients. As a proof of concept, we construct a federated PSG benchmark, demonstrating the efficacy of the existing PSG methods in an FL setting with controllable semantic heterogeneity of scene graphs. We also present the effectiveness of our benchmark by applying robust federated learning algorithms to data heterogeneity to show increased performance. To our knowledge, this is the first benchmark framework that enables federated learning and its evaluation for multi-semantic vision tasks under the controlled semantic heterogeneity. Our code is available at https://github.com/Seung-B/FL-PSG.

cs.CV↗

arXiv2025

Understanding Flatness in Generative Models: Its Role and Benefits

Flat minima, known to enhance generalization and robustness in supervised learning, remain largely unexplored in generative models. In this work, we systematically investigate the role of loss surface flatness in generative models, both theoretically and empirically, with a particular focus on diffusion models. We establish a theoretical claim that flatter minima improve robustness against perturbations in target prior distributions, leading to benefits such as reduced exposure bias -- where errors in noise estimation accumulate over iterations -- and significantly improved resilience to model quantization, preserving generative performance even under strong quantization constraints. We further observe that Sharpness-Aware Minimization (SAM), which explicitly controls the degree of flatness, effectively enhances flatness in diffusion models even surpassing the indirectly promoting flatness methods -- Input Perturbation (IP) which enforces the Lipschitz condition, ensembling-based approach like Stochastic Weight Averaging (SWA) and Exponential Moving Average (EMA) -- are less effective. Through extensive experiments on CIFAR-10, LSUN Tower, and FFHQ, we demonstrate that flat minima in diffusion models indeed improve not only generative performance but also robustness.

cs.CV↗

arXiv2019

Strain control of the Neel vector in Mn-based antiferromagnets

Control of the Neel vector in antiferromagnetic materials is one of the challenges preventing their use as active device components. Several methods have been investigated such as exchange bias, electric current, and spin injection, but little is known about strain-mediated anisotropy. This study of the antiferromagnetic L10-type MnX alloys MnIr, MnRh, MnNi, MnPd, and MnPt shows that a small amount of strain effectively rotates the direction of the Neel vector by 90 degrees for all of the materials. For MnIr, MnRh, MnNi, and MnPd, the Neel vector rotates within the basal plane. For MnPt, the Neel vector rotates from out-of-plane to in-plane under tensile strain. The effectiveness of strain control is quantified by a metric of efficiency and by direct calculation of the magnetostriction coefficients. The values of the magnetostriction coefficients are comparable with those from ferromagnetic materials. These results indicate that strain is a mechanism that can be exploited for control of the Neel vectors in this family of antiferromagnets.

physics.app-ph↗