arXiv subjects

Gyeongmin Kim

Publications and source records attributed to Gyeongmin Kim.

11 recordsLinked to original sources

arXiv2026

Extracting Voice Styles from Frozen TTS Models via Gradient-Based Inverse Optimization

Some text-to-speech systems ship a synthesis model and preset style vectors but withhold the reference encoder that turns a recording into a style vector, so a user cannot obtain a style for a new voice. We recover that vector without the encoder by inverting the released pipeline with gradient descent: all weights stay frozen and only the style vector is optimized, against time-pooled WavLM statistics of one recording of the target. The objective discards the time axis, so no transcript is needed. Two constraints from the release make the result a drop-in style: every row of the timbre style stays on the unit sphere where the presets lie, and the small duration style, which a time-pooled loss cannot see, is fitted to the recording's speaking rate. On SupertonicTTS, over 147 speakers and 100 held-out sentences each, ECAPA-TDNN similarity rises from 0.129 to 0.419, every recovered style starts from a preset and ends closer to its target than that preset was, and the pooled word error rate stays below that of the presets. A verifier at its equal-error point accepts 52% of the recovered voices as the target speaker, against 1% of the presets.

eess.AS↗

arXiv2026

faster-enhancer.c: A Dependency-Free int8 Runtime for Streaming Speech Enhancement on Commodity CPUs

This is an implementation and measurement study of what it costs to run a streaming speech enhancer on a CPU. We port FastEnhancer-Medium at 48 kHz to faster-enhancer.c, a C runtime with six int8 GEMM tiers selected at initialization, leaving architecture and weights untouched. One Apple M2 core reaches 0.069 real-time factor, against 0.230 for the fp32 ONNX Runtime graph on the same machine, a 3.3x speedup. A Galaxy S23+ (Snapdragon 8 Gen 2) reaches 0.096. The speedup comes from specializing every layer of the runtime around one fixed model. Activation ranges are recomputed per frame, so no calibration set is needed; the k=3 convolutions use Winograd F(2,3); cross-stage state is fp16; the GRU and the dequantization epilogues are fused; and nothing is allocated after startup. Over 824 VoiceBank-DEMAND utterances the engine tracks fp32 to within -0.006 PESQ and -0.08 dB SNR. Speed alone does not settle deployment cost. The enhancer holds a fraction of a core for as long as the microphone is open, so its real-time factor is a duty cycle. A benchmark races through a file; an audio callback does not. Pacing to the 6.67 ms deadline costs 4.2x per frame, saves 49% of the energy, and leaves the cheapest core placement missing 96% of its deadlines. All SIMD tiers within an architecture family emit byte-identical output. The runtime is released as a dependency-free library.

eess.AS↗

arXiv2026

Engineering electrically-switchable quantum anomalous Hall states by spin-orbit coupling

Nonvolatile gate-driven switching of quantum anomalous Hall (QAH) states in graphene moiré systems provides a promising route toward topological electronics based on chiral edge states. However, deliberate use of this switching mechanism requires control over both the magnetic properties and metastability of QAH states. While previous demonstrations mostly relied on the intrinsic magnetic energy landscape of moiré devices, here we show that this landscape can be engineered through proximity coupling to WSe2. We find that proximitizing twisted monolayer-bilayer graphene by WSe2 reshapes the magnetization reversals responsible for nonvolatile electrical switching of QAH states. We attribute this effect to the proximity-induced spin-orbit coupling (SOC), which can lock spin and valley and modify the magnetization of the competing states involved in switching compared with non-proximitized graphene systems. Our findings establish proximity-induced SOC as a new way to engineer magnetic properties and switchable magnetic states in graphene-based systems. We further demonstrate that strong magnetic metastability in tMBG allows the magnetic states to be gate-tuned between QAH and metallic regimes, and between QAH states with Chern numbers |C| = 2 and 1 without resetting the magnetic state. This functionality points toward new device architectures based on QAH chiral edge states.

cond-mat.mes-hall↗

arXiv2026

Training Time Prediction for Mixed Precision-based Distributed Training

Accurate prediction of training time in distributed deep learning is crucial for resource allocation, cost estimation, and job scheduling. We observe that the floating-point precision setting is a key determinant of training time, leading to training time variations of ~2.4x over its minimum. However, existing studies on distributed training time prediction rely on static model computation graphs that do not capture precision variations, including mixed precision. According to our experiments, training time prediction without considering precision results in significant prediction errors - reaching up to 147.85% in mean absolute percentage error (MAPE). To address this issue, we propose a precision-aware distributed training time predictor that achieves robust accuracy across diverse precision settings, including mixed precision, with 9.8% MAPE.

cs.LG↗

arXiv2026

MAESIL: Masked Autoencoder for Enhanced Self-supervised Medical Image Learning

Training deep learning models for three-dimensional (3D) medical imaging, such as Computed Tomography (CT), is fundamentally challenged by the scarcity of labeled data. While pre-training on natural images is common, it results in a significant domain shift, limiting performance. Self-Supervised Learning (SSL) on unlabeled medical data has emerged as a powerful solution, but prominent frameworks often fail to exploit the inherent 3D nature of CT scans. These methods typically process 3D scans as a collection of independent 2D slices, an approach that fundamentally discards critical axial coherence and the 3D structural context. To address this limitation, we propose the autoencoder for enhanced self-supervised medical image learning(MAESIL), a novel self-supervised learning framework designed to capture 3D structural information efficiently. The core innovation is the 'superpatch', a 3D chunk-based input unit that balances 3D context preservation with computational efficiency. Our framework partitions the volume into superpatches and employs a 3D masked autoencoder strategy with a dual-masking strategy to learn comprehensive spatial representations. We validated our approach on three diverse large-scale public CT datasets. Our experimental results show that MAESIL demonstrates significant improvements over existing methods such as AE, VAE and VQ-VAE in key reconstruction metrics such as PSNR and SSIM. This establishes MAESIL as a robust and practical pre-training solution for 3D medical imaging tasks.

cs.CV↗

arXiv2026

3D-LLDM: Label-Guided 3D Latent Diffusion Model for Improving High-Resolution Synthetic MR Imaging in Hepatic Structure Segmentation

Deep learning and generative models are advancing rapidly, with synthetic data increasingly being integrated into training pipelines for downstream analysis tasks. However, in medical imaging, their adoption remains constrained by the scarcity of reliable annotated datasets. To address this limitation, we propose 3D-LLDM, a label-guided 3D latent diffusion model that generates high-quality synthetic magnetic resonance (MR) volumes with corresponding anatomical segmentation masks. Our approach uses hepatobiliary phase MR images enhanced with the Gd-EOB-DTPA contrast agent to derive structural masks for the liver, portal vein, hepatic vein, and hepatocellular carcinoma, which then guide volumetric synthesis through a ControlNet-based architecture. Trained on 720 real clinical hepatobiliary phase MR scans from Samsung Medical Center, 3D-LLDM achieves a Fréchet Inception Distance (FID) of 28.31, improving over GANs by 70.9% and over state-of-the-art diffusion baselines by 26.7%. When used for data augmentation, the synthetic volumes improve hepatocellular carcinoma segmentation by up to 11.153% Dice score across five CNN architectures.

cs.CV↗

arXiv2026

Flow Actor-Critic for Offline Reinforcement Learning

The dataset distributions in offline reinforcement learning (RL) often exhibit complex and multi-modal distributions, necessitating expressive policies to capture such distributions beyond widely-used Gaussian policies. To handle such complex and multi-modal datasets, in this paper, we propose Flow Actor-Critic, a new actor-critic method for offline RL, based on recent flow policies. The proposed method not only uses the flow model for actor as in previous flow policies but also exploits the expressive flow model for conservative critic acquisition to prevent Q-value explosion in out-of-data regions. To this end, we propose a new form of critic regularizer based on the flow behavior proxy model obtained as a byproduct of flow-based actor design. Leveraging the flow model in this joint way, we achieve new state-of-the-art performance for test datasets of offline RL including the D4RL and recent OGBench benchmarks.

cs.LG↗

arXiv2023

DynaCon: Dynamic Robot Planner with Contextual Awareness via LLMs

Mobile robots often rely on pre-existing maps for effective path planning and navigation. However, when these maps are unavailable, particularly in unfamiliar environments, a different approach become essential. This paper introduces DynaCon, a novel system designed to provide mobile robots with contextual awareness and dynamic adaptability during navigation, eliminating the reliance of traditional maps. DynaCon integrates real-time feedback with an object server, prompt engineering, and navigation modules. By harnessing the capabilities of Large Language Models (LLMs), DynaCon not only understands patterns within given numeric series but also excels at categorizing objects into matched spaces. This facilitates dynamic path planner imbued with contextual awareness. We validated the effectiveness of DynaCon through an experiment where a robot successfully navigated to its goal using reasoning. Source code and experiment videos for this work can be found at: https://sites.google.com/view/dynacon.

cs.RO↗

arXiv2023

BIRP: Bitcoin Information Retrieval Prediction Model Based on Multimodal Pattern Matching

Financial time series have historically been assumed to be a martingale process under the Random Walk hypothesis. Instead of making investment decisions using the raw prices alone, various multimodal pattern matching algorithms have been developed to help detect subtly hidden repeatable patterns within the financial market. Many of the chart-based pattern matching tools only retrieve similar past chart (PC) patterns given the current chart (CC) pattern, and leaves the entire interpretive and predictive analysis, thus ultimately the final investment decision, to the investors. In this paper, we propose an approach of ranking similar PC movements given the CC information and show that exploiting this as additional features improves the directional prediction capacity of our model. We apply our ranking and directional prediction modeling methodologies on Bitcoin due to its highly volatile prices that make it challenging to predict its future movements.

q-fin.ST↗

arXiv2022

QUAK: A Synthetic Quality Estimation Dataset for Korean-English Neural Machine Translation

With the recent advance in neural machine translation demonstrating its importance, research on quality estimation (QE) has been steadily progressing. QE aims to automatically predict the quality of machine translation (MT) output without reference sentences. Despite its high utility in the real world, there remain several limitations concerning manual QE data creation: inevitably incurred non-trivial costs due to the need for translation experts, and issues with data scaling and language expansion. To tackle these limitations, we present QUAK, a Korean-English synthetic QE dataset generated in a fully automatic manner. This consists of three sub-QUAK datasets QUAK-M, QUAK-P, and QUAK-H, produced through three strategies that are relatively free from language constraints. Since each strategy requires no human effort, which facilitates scalability, we scale our data up to 1.58M for QUAK-P, H and 6.58M for QUAK-M. As an experiment, we quantitatively analyze word-level QE results in various ways while performing statistical analysis. Moreover, we show that datasets scaled in an efficient way also contribute to performance improvements by observing meaningful performance gains in QUAK-M, P when adding data up to 1.58M.

cs.CL↗

arXiv2022

KoCHET: a Korean Cultural Heritage corpus for Entity-related Tasks

As digitized traditional cultural heritage documents have rapidly increased, resulting in an increased need for preservation and management, practical recognition of entities and typification of their classes has become essential. To achieve this, we propose KoCHET - a Korean cultural heritage corpus for the typical entity-related tasks, i.e., named entity recognition (NER), relation extraction (RE), and entity typing (ET). Advised by cultural heritage experts based on the data construction guidelines of government-affiliated organizations, KoCHET consists of respectively 112,362, 38,765, 113,198 examples for NER, RE, and ET tasks, covering all entity types related to Korean cultural heritage. Moreover, unlike the existing public corpora, modified redistribution can be allowed both domestic and foreign researchers. Our experimental results make the practical usability of KoCHET more valuable in terms of cultural heritage. We also provide practical insights of KoCHET in terms of statistical and linguistic analysis. Our corpus is freely available at https://github.com/Gyeongmin47/KoCHET.

cs.CL↗