arXiv subjects

Jongmin Kim

Publications and source records attributed to Jongmin Kim.

At least 19 recordsLinked to original sources

arXiv2026

KBS: A Korean Benchmark Suite for Foundation Model Evaluation

We present a Korean Benchmark Suite (KBS) comprising Ko-ARC, Ko-GSM8K, Ko-EQ-Bench, Ko-WinoGrande, Ko-LAMBADA, and Ko-IFEval, covering six capabilities across 9,396 items. The suite combines review, localization, direct Korean construction, and redesign for Korean, avoiding a translation-only approach to benchmark adaptation. We evaluate Korean-developed models, multilingual open models, and closed API models under protocols that account for model access, prompt format, and scoring differences. Results show that Korean benchmark scores depend on task construction, scoring rules, prompting, and model post-training. These findings suggest that Korean evaluation should report both final scores and the settings used to obtain them.

cs.CL↗

arXiv2026

Thunder-DeID+: Boundary-Aware Korean De-identification from Court Judgments to Broader Sensitive Documents

This study examines the transferability of a de-identification model developed for Korean court judgments to a broader range of sensitive documents, focusing on boundary precision, label granularity, and recall. The agglutinative structure of Korean causes Byte Pair Encoding (BPE) tokenizers to frequently merge Personally Identifiable Information (PII) with adjacent particles, which complicates accurate span recovery. Furthermore, current PII taxonomies lack sufficient granularity for legal texts. To address these issues, this study introduces Thunder-DeID+, which integrates a refined three-tier PII taxonomy with a boundary-aware tokenizer. The tokenizer enhances a pretrained BPE vocabulary by incorporating N-gram and boundary-scoring components to better align token and PII span boundaries. The proposed approach is evaluated using a tokenizer-independent word-segment-level protocol. On full-length judgments, the tokenizer consistently improves W. Strict Typed F1-score from 0.76-0.77 to 0.87-0.88 across four model sizes (150M, 395M, 750M, and 1B), mainly by increasing precision rather than model scale. With fine-tuning on 1,000 broader sensitive documents, the judgment-trained model further achieves 0.93 W. Strict Typed F1-score, indicating effective transfer beyond court judgments.

cs.CL↗

arXiv2026

Counterfactual Gate-Support Decomposition: Separating Selection from Redistribution in Training-Data Filtering

A training-data quality gate changes both which examples survive within regions of the input space and how training mass is distributed across those regions, so a gate-versus-random comparison does not isolate these mechanisms. We introduce Counterfactual Gate-Support Decomposition (CGSD), a four-arm, budget-matched design that crosses gate eligibility with support distribution and decomposes the observed gate effect into within-support selection, between-support redistribution, and their interaction. The practical constraint is finite-pool realizability: we characterize mass and token-assembly conditions under which the required arms can be constructed, together with a support-extinction case in which the ordinary design is undefined. On Korean-English and Korean-Japanese retrieval, the mean within-support selection component is the only one that exceeds its calibrated null band (+29.7 and +5.6 holdout items; 30/30 positive replicates), decisively in Korean-English and marginally in Korean-Japanese; redistribution and interaction remain unresolved at the available power. The same pattern holds for a second, unrelated gate family -- a model-confidence threshold rather than a deterministic verifier -- judged against its own null (+27.3 items, 9/9 positive), so the result is not a property of one gate's mechanism. On a mathematics gate, zero-eligibility supports leave the ordinary design undefined at every budget, while the extended design is constructible only inside narrow budget bands that an exhaustive scan locates and that the natural operating point misses -- the assembly condition separating a property of the pool from a limitation of the packer. CGSD therefore returns a decomposition, a constructible-but-underpowered design, or an explicit non-constructibility result, and says which; it prevents support redistribution from being silently interpreted as sample-level gate quality.

cs.LG↗

arXiv2026

Quantize by Drift: Label-Free Mixed-Precision Post-Training Quantization for Text Embedders

Mixed-precision post-training quantization needs a per-module sensitivity signal; for a text embedder the obvious one -- the retrieval quality a module costs when quantized -- needs relevance labels that deployments rarely have. We measure a label-free substitute: quantization-induced representation drift, obtained by quantizing one module, re-encoding the corpus, and recording how far the output embeddings moved from their full-precision positions. What is specific is the observable: the deployed output representation a dense retriever ranks with. Across five development embedders, configuration-level drift orders sampled mixed-precision plans against held-out retrieval quality at a macro Spearman of 0.911, the sensitivity transports across calibration corpora and retrieval domains in the usable regime, module drifts compose rank-consistently but not numerically, and relevance-derived sensitivity adds no consistent value. The method is one additive allocation under a hard packed-byte budget, with no labels and no search. On three embedders held untouched until method, baselines and hypotheses were frozen and sealed, the pre-registered directional hypothesis against the prior LieQ criterion holds (3/3 at the main budget, no collapse) and drift scores above a two-sided LieQ steelman in 2/3; but at the main budget drift is numerically lower than same-budget uniform precision on all three (-0.99, -0.85, -1.01 points), having reduced module and whole-model drift as designed. Output drift is thus a robust coarse sensitivity signal, not a universally optimal allocation objective: it avoids the catastrophic failures of the transferred signed-geometry adaptation and can remain usable at stressed budgets where uniform collapses, but fine-grained redistribution around a strong uniform operating point remains unresolved.

cs.IR↗

arXiv2026

CompVLA: A Variable Compliance Vision-Language-Action Model for Contact-rich Manipulation

Contact-rich manipulation, requiring robots to regulate not only motion but also how they yield to external forces, has emerged as the next frontier for Vision-Language-Action (VLA) models. However, existing VLAs output purely kinematic commands, degrading performance on real-world contact-rich tasks. In this paper, we introduce CompVLA, a unified VLA framework that jointly predicts motion and stiffness matrix from RGB and language inputs. Our approach augments the conventional architecture with a dedicated Compliance Expert, which outputs time-varying stiffness and virtual displacement profiles executed via geometric impedance control. We demonstrate that CompVLA achieves the highest average success rate across diverse contact-rich tasks, outperforming both vanilla and compliance-aware VLA baselines, with ablations confirming each component is essential.

cs.RO↗

arXiv2026

Keep-or-Drop? Adaptive Tokenizer for Compact Video Representation

Latent diffusion models have emerged as a dominant framework for high-fidelity image and video synthesis, operating in compact latent spaces with variational autoencoders (VAEs) to enhance computational efficiency without compromising visual quality. However, conventional VAEs are suboptimal for video data as they employ fixed compression ratios that cannot adapt to the varying complexity of spatio-temporal content. We present KATok (Keep-or-Drop? Adaptive Tokenizer for Compact Video Representation), a transformer-based VAE that incorporates an adaptive token selector which is jointly learned with latent tokens. By evaluating each token's content-richness as keep-or-drop probability, the token selector effectively discards uninformative tokens, naturally allowing data-dependent compression. Applying adaptive tokenization to diffusion models may cause spatial misalignment, as token dropping can disturb the original spatio-temporal structure. To alleviate this issue, we propose two position-prediction strategies: cascaded and joint generation, to ensure spatial consistency. We empirically show that our model achieves strong reconstruction and generation quality at a state-of-the-art compression ratio. Further analysis on video data reveals that this improvement is primarily achieved by reducing spatio-temporal redundancy and removing uninformative tokens, as supported by both quantitative and qualitative results.

cs.CV↗

arXiv2026

VIPIR: A Versatile GPU Framework for Integrating Private Information Retrieval Protocols

While private information retrieval (PIR) enables private database services by fully concealing access patterns, it simultaneously requires high computational throughput, large memory capacity, and substantial memory bandwidth. We introduce VIPIR, a versatile GPU framework that co-designs PIR protocols with GPU acceleration. We develop a unified analytic model showing that state-of-the-art PIR protocols fall into two categories with complementary limitations, and propose two protocols that flexibly combine techniques across these categories, overcoming the limitations of both classes. These protocols incorporate a GPU-friendly data compression method called expansion-based ring packing (ExpPack), which offers a high degree of parallelism and minimal communication cost. VIPIR applies further optimizations to core operations, including number-theoretic transforms (NTTs) and various matrix-matrix multiplications (GEMMs). Notably, we develop a tensor-core-based execution method for database multiplication by interpreting it as a mixed-integer-type GEMM. We also design memory-efficient scheduling methods that minimize intermediate buffers and enable multi-GPU scaling under memory capacity constraints. Overall, VIPIR achieves orders-of-magnitude higher throughput than prior PIR systems while reducing communication and memory overheads, making large-scale PIR practical.

cs.CR↗

arXiv2026

WHET: Welding Homomorphic Encryption to Accelerator Architectures

Fully homomorphic encryption (FHE) enables computations on encrypted data without decryption, offering strong data privacy at the expense of substantial computational and memory overheads. Prior efforts have steadily improved FHE performance through cryptographic and algorithmic enhancements or hardware acceleration, yet these two directions have progressed largely in isolation, hindering the full exploitation of available hardware capabilities. This work presents WHET, which introduces memory-centric, architecture-aware optimizations to better align cryptographic and algorithmic constructions with FHE accelerator architectures. We identify conventional FHE constructions as major sources of excessive working sets and heavy off-chip memory traffic. We propose accelerator-specific techniques, including fine-grained coefficient-to-slot transformation, plaintext compression, and intermediate modulus raising, to reduce the on-chip data footprint by minimizing temporary ciphertexts and plaintext loads. With these techniques applied, we observe additional opportunities to improve on-chip memory efficiency; hence, we introduce lightweight architectural refinements, including a special-purpose buffer and functional unit extensions. With these optimizations, WHET achieves 1.38-8.74$\times$ per-area performance improvements over state-of-the-art FHE accelerators and the first-ever sub-millisecond CKKS bootstrapping.

cs.CR↗

arXiv2026

Energy spectra and cascade in the spin turbulence of a driven spinor Bose-Einstein condensate

We investigate the spin-interaction energy spectrum of spin turbulence in a driven spinor Bose-Einstein condensate. Continuous spin driving of a spin-1 condensate produces a nonequilibrium steady state with spatially fluctuating magnetization. We observe a power-law scaling consistent with the $-7/3$ exponent predicted for spin-wave turbulence, which persists across our full range of drive strengths despite substantial changes in the spectral anisotropy. After switching off the drive, we track the free-decay evolution and find evidence consistent with a direct cascade of spin-interaction energy toward higher wavenumbers. These results establish an energy-spectral hallmark of spin turbulence and enable quantitative studies of cascade dynamics in spinor superfluids.

cond-mat.quant-gas↗

arXiv2026

Efficient ensemble randomization by tuning chaos in a nonlinear spin-1 system

We present an efficient scheme to randomize a spin-state ensemble in a nonlinear spin-1 system by tuning chaos with an external periodic drive. Without modulation, the system exhibits a mixed phase space featuring regular islands embedded in a chaotic sea, where global mixing is inhibited by energy conservation. Using numerical simulations, we demonstrate that weak modulation of a linear Zeeman field not only facilitates transport between different energy shells but also drives ensembles toward a Haar-random distribution over spin states. Under optimized conditions, complete randomization is achieved on a timescale set by the inverse nonlinear interaction energy. In the overdriven regime, randomization is unexpectedly suppressed at specific modulation amplitudes, accompanied by the formation of sticky regions in phase space. We attribute this behavior to the dynamical cancellation of the leading low-order harmonic component of the periodic drive. These results illustrate how time-periodic driving can be used to engineer chaotic systems and achieve controllable randomization in nonlinear spin systems.

quant-ph↗

arXiv2026

Slot-MLLM: Object-Centric Visual Tokenization for Multimodal LLM

Recently, multimodal large language models (MLLMs) have emerged as a key approach in achieving artificial general intelligence. In particular, vision-language MLLMs have been developed to generate not only text but also visual outputs from multimodal inputs. This advancement requires efficient image tokens that LLMs can process effectively both in input and output. However, existing image tokenization methods for MLLMs typically capture only global abstract concepts or uniformly segmented image patches, restricting MLLMs' capability to effectively understand or generate detailed visual content, particularly at the object level. To address this limitation, we propose an object-centric visual tokenizer based on Slot Attention specifically for MLLMs. In particular, based on the Q-Former encoder, diffusion decoder, and residual vector quantization, our proposed discretized slot tokens can encode local visual details while maintaining high-level semantics, and also align with textual data to be integrated seamlessly within a unified next-token prediction framework of LLMs. The resulting Slot-MLLM demonstrates significant performance improvements over baselines with previous visual tokenizers across various vision-language tasks that entail local detailed comprehension and generation. Notably, this work is the first demonstration of the feasibility of object-centric slot attention performed with MLLMs and in-the-wild natural images.

cs.CV↗

arXiv2026

GPIR: Enabling Practical Private Information Retrieval with GPUs

Private information retrieval (PIR) allows private database queries; however, it is hindered by intense server-side computation and memory traffic. Numerous modern lattice-based PIR protocols consist of three phases: ExpandQuery (expanding a query into encrypted indices), RowSel (encrypted row selection), and ColTor (recursive "column tournament" for final selection). ExpandQuery and ColTor primarily perform number-theoretic transforms (NTTs), whereas RowSel reduces to large-scale independent matrix-matrix multiplications (GEMMs). GPUs are well suited for these tasks when combined with multi-client batching, which is necessary for high throughput. However, batching fundamentally reshapes the performance bottlenecks: while it amortizes database access costs, it expands working sets beyond the L2 cache capacity, causing divergent memory access behavior and excessive DRAM traffic. We present GPIR, a GPU-accelerated PIR system that rethinks kernel design, data layout, and execution scheduling. We introduce a stage-aware hybrid execution model that dynamically switches between operation-level kernels, which execute each primitive operation separately, and stage-level kernels, which fuse all operations within a stage into a single kernel to maximize on-chip data reuse. For RowSel, we resolve the mismatch between NTT-driven layouts and tiled GEMMs using a transposed-layout design with fine-grained pipelining. We further extend GPIR to multi-GPU systems, scaling throughput and database capacity with negligible communication overhead. GPIR achieves up to 297.2x higher throughput than PIRonGPU, the state-of-the-art GPU implementation.

cs.CR↗

arXiv2026

Theodosian: A Deep Dive into Memory-Hierarchy-Centric FHE Acceleration

Fully homomorphic encryption (FHE) enables secure computation on encrypted data, mitigating privacy concerns in cloud and edge environments. However, due to its high compute and memory demands, extensive acceleration research has been pursued across diverse hardware platforms, especially GPUs. In this paper, we perform a microarchitectural analysis of CKKS, a popular FHE scheme, on modern GPUs. Focusing on the memory hierarchy, we demonstrate that dominant kernels remain bound by the on-chip L2 cache despite its high bandwidth, exposing a persistent inner memory wall beyond the conventional off-chip DRAM bottleneck. Further, we reveal that the overall CKKS throughput is constrained by low per-kernel hardware utilization, caused by insufficient intra-kernel parallelism. Motivated by these findings, we introduce Theodosian, a set of complementary, memory-aware optimizations that improve cache efficiency and reduce runtime overheads. Theodosian achieves 1.45--1.83x performance improvements over a highly optimized baseline, Cheddar, across representative CKKS workloads. On an RTX 5090, we reduce the bootstrapping latency for 32,768 complex numbers from 22.1ms to 15.2ms, and further to 12.8ms with additional algorithmic optimizations, establishing a new state-of-the-art GPU performance to the best of our knowledge.

cs.CR↗

arXiv2026

Enhancement of damping in a turbulent atomic Bose-Einstein condensate

Turbulence enhances momentum transport in classical fluids, effectively increasing their viscosity. We investigate an analogous effect in a superfluid by measuring the damping of collective oscillations in an atomic Bose-Einstein condensate (BEC) containing stationary spin-superflow turbulence. Using continuous spin driving to maintain turbulence in a spin-1 $^{23}$Na BEC, we excite its quadrupole mode and measure the damping rate over a range of temperatures. The damping consistently exceeds the Landau-damping rate expected for an equilibrium, non-turbulent BEC. The enhancement likely originates from two complementary processes: direct energy transfer from the mode to turbulent condensate fluctuations and turbulence-induced modification of the thermal cloud that amplifies Landau damping. These results establish collective-mode damping as a sensitive probe of momentum transport in superfluid turbulence.

cond-mat.quant-gas↗

arXiv2026

Rethinking LLM Inference Bottlenecks: Insights from Latent Attention and Mixture-of-Experts

Computational workloads composing traditional transformer models are starkly bifurcated. Multi-Head Attention (MHA) and Grouped-Query Attention are memory-bound due to low arithmetic intensity, while FeedForward Networks are compute-bound. This dichotomy has long motivated research into specialized hardware to mitigate the attention bottleneck. This paper argues that recent architectural advances in transformer models -- Multi-head Latent Attention (MLA) and Mixture of Experts (MoE) -- introduce new dominant bottlenecks, shifting the challenge away from memory-intensive attention. We make two key observations. First, the arithmetic intensity of MLA is over two orders of magnitude higher than that of MHA, moving it toward a compute-bound regime well-matched to modern accelerators such as GPUs. Second, distributing MoE experts across a pool of accelerators allows batching to tune their arithmetic intensity to that of dense layers, producing a more balanced computational profile. Consequently, the focus of hardware and system optimization should shift from attention acceleration to high-bandwidth interconnects and balancing expert workloads across accelerators.

cs.AR↗

arXiv2025

A Delayed Acceptance Auxiliary Variable MCMC for Spatial Models with Intractable Likelihood Function

A large class of spatial models contains intractable normalizing functions, such as spatial lattice models, interaction spatial point processes, and social network models. Bayesian inference for such models is challenging since the resulting posterior distribution is doubly intractable. Although auxiliary variable MCMC (AVM) algorithms are known to be the most practical, they are computationally expensive due to the repeated auxiliary variable simulations. To address this, we propose delayed-acceptance AVM (DA-AVM) methods, which can reduce the number of auxiliary variable simulations. The first stage of the kernel uses a cheap surrogate to decide whether to accept or reject the proposed parameter value. The second stage guarantees detailed balance with respect to the posterior. The auxiliary variable simulation is performed only on the parameters accepted in the first stage. We construct various surrogates specifically tailored for doubly intractable problems, including subsampling strategy, Gaussian process emulation, and frequentist estimator-based approximation. We validate our method through simulated and real data applications, demonstrating its practicality for complex spatial models.

stat.ME↗

arXiv2025

IVE: An Accelerator for Single-Server Private Information Retrieval Using Versatile Processing Elements

Private information retrieval (PIR) is an essential cryptographic protocol for privacy-preserving applications, enabling a client to retrieve a record from a server's database without revealing which record was requested. Single-server PIR based on homomorphic encryption has particularly gained immense attention for its ease of deployment and reduced trust assumptions. However, single-server PIR remains impractical due to its high computational and memory bandwidth demands. Specifically, reading the entirety of large databases from storage, such as SSDs, severely limits its performance. To address this, we propose IVE, an accelerator for single-server PIR with a systematic extension that enables practical retrieval from large databases using DRAM. Recent advances in DRAM capacity allow PIR for large databases to be served entirely from DRAM, removing its dependence on storage bandwidth. Although the memory bandwidth bottleneck still remains, multi-client batching effectively amortizes database access costs across concurrent requests to improve throughput. However, client-specific data remains a bottleneck, whose bandwidth requirements ultimately limits performance. IVE overcomes this by employing a large on-chip scratchpad with an operation scheduling algorithm that maximizes data reuse, further boosting throughput. Additionally, we introduce sysNTTU, a versatile functional unit that enhances area efficiency without sacrificing performance. We also propose a heterogeneous memory system architecture, which enables a linear scaling of database sizes without a throughput degradation. Consequently, IVE achieves up to 1,275x higher throughput compared to prior PIR hardware solutions.

cs.CR↗

arXiv2025

Beyond Line-Level Filtering for the Pretraining Corpora of LLMs

While traditional line-level filtering techniques, such as line-level deduplication and trailing-punctuation filters, are commonly used, these basic methods can sometimes discard valuable content, negatively affecting downstream performance. In this paper, we introduce two methods-pattern-aware line-level deduplication (PLD) and pattern-aware trailing punctuation filtering (PTF)-by enhancing the conventional filtering techniques. Our approach not only considers line-level signals but also takes into account their sequential distribution across documents, enabling us to retain structurally important content that might otherwise be removed. We evaluate these proposed methods by training small language models (1 B parameters) in both English and Korean. The results demonstrate that our methods consistently improve performance on multiple-choice benchmarks and significantly enhance generative question-answering accuracy on both SQuAD v1 and KorQuAD v1.

cs.CL↗