arXiv subjects

Chanho Park

Publications and source records attributed to Chanho Park.

At least 19 recordsLinked to original sources

arXiv2026

Retrieval Heads Meet Vision: Uncovering How VLMs Locate and Extract Visual Information

Vision-language models (VLMs) can locate an image region referred to by a text prompt and route the corresponding visual evidence to the output, yet the internal mechanism behind this behavior is not understood. Inspired by retrieval heads in large language models, we ask whether VLMs contain an analogous mechanism for visual retrieval. We answer affirmatively by introducing Visual Retrieval Heads (VRHs), a small subset of attention heads (about 1.7-2.6%) that are causally responsible for grounding text descriptions to image regions. To find them, we recast existing head-scoring methods under a unified design space over query tokens, key aggregation, and cross-sample aggregation. We then show that scoring attention from output prediction tokens with a sum over the ground-truth referent region most reliably identifies causal heads. Across eleven VLMs and five referring-expression benchmarks, masking only the top 20 VRHs reduces grounding accuracy by up to 80 percentage points, while masking the same number of random heads has little effect. Beyond replicating the causal-sparse-universal triad established for text retrieval heads, VRHs exhibit several properties not previously reported: they generalize across visual reference tasks, remaining causal on attribute, spatial, counting, and visual-math benchmarks despite being discovered through bounding-box prediction; they are functionally specific, preserving output format while corrupting localization; and they are architecturally shared, transferring causally across VLMs that share an LLM backbone but differ in vision encoder, projector, and instruction tuning.

cs.CV↗

arXiv2026

Biodegradable, Millimeter-Scale Light-Emitting Sensors for Distributed Environmental Monitoring-Functional Pixie Dust

Methods for large-area, precise monitoring across natural environments are of growing interest due to pressing needs for sustainable management of rapidly increasing anthropogenic activities. Established approaches involve sparse spatial sampling and/or sequential measurements, while emerging techniques exploit miniaturized electronics or passive optical methods. Various constraints in scalability, costs, robustness, operational range and other factors create a need for alternatives. Here, we introduce a concept that overcomes many of these limitations through the combined use of chemically induced light emission and chemically responsive optical filter elements in millimeter-scale systems that we refer to as functional pixie dust (fPD) sensors, designed specifically for monitoring natural water systems during nighttime to eliminate background optical interference and to enhance remote analysis. These floating devices act as Lagrangian tracers to follow surface flows and to simultaneously measure the concentrations of key chemical species along their trajectories. Optimized designs exploit environmentally compatible constituent materials that are also degradable through natural processes to benign end products, thereby eliminating the need for recovery. Spatially and spectrally resolved ratiometric measurement schemes ensure robust operation and ability to address practical requirements in range, operational lifetime, time response and sensitivity. Demonstrations include distributed measurements of pH, Hg2+, and NO2-, each of relevance to industrial discharge, toxic metal contamination, and nitrogen-rich runoff, adapted for static concentration gradients, flow-driven transport conditions, and outdoor aquatic settings. The results establish a framework for environmental sensing using degradable, self-powered microsystems capable of scalable deployment and remote readout.

physics.app-ph↗

arXiv2026

Early Cue Precision Shapes Visual Shortcut Learning in Controlled Cue-Manipulation Benchmarks

Visual classifiers can achieve high matched-distribution accuracy while relying on low-level cues that fail under conflict or suppression. We test whether this failure is shaped by early cue precision: the reliability with which a low-level cue predicts the label during early learning or downstream probe fitting. Across synthetic shape-texture tasks, sequential digit training, a 10-class frozen-representation audit, and a CIFAR-10 natural-image-based texture-overlay benchmark, we manipulate object-texture match probability and evaluate matched-ID accuracy, conflict accuracy, texture-choice rate, and suppression behavior. Degraded-but-predictive input does not substitute for cue decorrelation. In 10-class digit probes, conflict accuracy drops from 0.589 under chance-like cue precision to 0.005 under target-perfect texture. In CIFAR-10 frozen probes, conflict accuracy drops from 0.569 to 0.114, while texture choice rises from 0.049 to 0.855; this ordering persists across texture-overlay strengths alpha in {0.15,0.25,0.35,0.50}. End-to-end CIFAR-10 training shows that low early cue precision improves pre-target conflict behavior, but shortcut-rich fine-tuning can rapidly overwrite this benefit. Cue decorrelation must therefore be maintained during downstream adaptation rather than treated as a one-time inoculation.

cs.CV↗

arXiv2026

PrismQuant: Rate-Distortion-Optimal Vector Quantization for Gaussian-Mixture Sources

For a Gaussian source under mean-squared error (MSE), classical transform coding is rate--distortion (RD) optimal: the Karhunen--Loeve transform (KLT) diagonalizes the covariance, reverse waterfilling allocates the bits, and scalar quantization closes the loop. This elegant story breaks down for multimodal sources, where no single covariance can capture heterogeneous local geometries, and the RD function loses its closed form. We revisit this problem through Gaussian-mixture sources and develop a constructive RD theory for them. Our key finding is that the mixture structure incurs only a component label cost. Conditioned on the active mixture component, each branch is Gaussian; the challenge is allocating bits across heterogeneous branches. We prove that the genie-aided conditional RD function is governed by a single global reverse-waterfilling level shared across all components and eigenmodes. Building on this result, we introduce PrismQuant, which transmits the component label losslessly and encodes the residual using the component-matched KLT, followed by scalar quantization, achieving a rate of H(C)/n bits per source dimension of the converse, with a vanishing asymptotic gap. We further develop a practical implementation based on EM-driven Gaussian-mixture learning, component-adaptive KLTs, and entropy-constrained scalar quantization (ECSQ). Experiments on synthetic Gaussian mixtures show that PrismQuant closely approaches the theoretical RD bound, while experiments on real-world channel-state-information (CSI) data demonstrate competitive or superior performance compared with transformer-based learned codecs at more than one order of magnitude smaller model size.

cs.IT↗

arXiv2026

CSI Compression for Massive MIMO-OFDM: Mismatch-Aware Rate-Distortion Trade-offs

We study channel state information (CSI) compression for wideband frequency division duplex massive multiple-input multiple-output (MIMO) when the base station (BS) reconstructs CSI using an imperfect covariance model. Under matched second-order statistics, remote rate--distortion theory yields transform coding with reverse water-filling (RWF) over covariance eigenmodes. With decoder-side covariance mismatch, however, this allocation is no longer end-to-end optimal. We derive an achievable mismatched Gaussian rate--distortion characterization based on a Gaussian test channel and a mismatched minimum mean square error (MMSE) reconstruction rule. In a shared-eigenvector regime (common eigenbasis, mismatched eigenvalues), the problem decouples across modes and leads to a robust reverse water-filling (RRWF) allocation computable via bisection and per-mode root finding. Simulations using wideband massive MIMO covariance models show that RRWF consistently improves reconstruction distortion and end-to-end mean square error relative to conventional RWF under mismatch.

cs.IT↗

arXiv2026

Token Warping Helps MLLMs Look from Nearby Viewpoints

Can warping tokens, rather than pixels, help multimodal large language models (MLLMs) understand how a scene appears from a nearby viewpoint? While MLLMs perform well on visual reasoning, they remain fragile to viewpoint changes, as pixel-wise warping is highly sensitive to small depth errors and often introduces geometric distortions. Drawing on theories of mental imagery that posit part-level structural representations as the basis for human perspective transformation, we examine whether image tokens in ViT-based MLLMs serve as an effective substrate for viewpoint changes. We compare forward and backward warping, finding that backward token warping, which defines a dense grid on the target view and retrieves a corresponding source-view token for each grid point, achieves greater stability and better preserves semantic coherence under viewpoint shifts. Experiments on our proposed ViewBench benchmark demonstrate that token-level warping enables MLLMs to reason reliably from nearby viewpoints, consistently outperforming all baselines including pixel-wise warping approaches, spatially fine-tuned MLLMs, and a generative warping method.

cs.CV↗

arXiv2026

Fundamental Limits of CSI Compression in FDD Massive MIMO

Channel state information (CSI) feedback in frequency-division duplex (FDD) massive multiple-input multiple-output (MIMO) systems is fundamentally limited by the high dimensionality of wideband channels. In this paper, we model the stacked wideband CSI vector as a Gaussian-mixture source with a latent geometry state that represents different propagation environments. Each component corresponds to a locally stationary regime characterized by a correlated proper complex Gaussian distribution with its own covariance matrix. This representation captures the multimodal nature of practical CSI datasets while preserving the analytical tractability of Gaussian models. Motivated by this structure, we propose Gaussian-mixture transform coding (GMTC), a practical CSI feedback architecture that combines state inference with state-adaptive TC. The mixture parameters are learned offline from channel samples and stored as a shared statistical dictionary at both the user equipment (UE) and the base station. For each CSI realization, the UE identifies the most likely geometry state, encodes the corresponding label using a lossless source code, and compresses the CSI using the Karhunen-Loeve transform matched to that state. We further characterize the fundamental limits of CSI compression under this model by deriving analytical converse and achievability bounds on the rate-distortion (RD) function. A key structural result is that the optimal bit allocation across all mixture components is governed by a single global reverse-waterfilling level. Simulations on the COST2100 dataset show that GMTC significantly improves the RD tradeoff relative to neural transform coding approaches while requiring substantially smaller model memory and lower inference complexity. These results indicate that near-optimal CSI compression can be achieved through state-adaptive TC without relying on large neural encoders.

cs.IT↗

arXiv2026

BoxSplitGen: A Generative Model for 3D Part Bounding Boxes in Varying Granularity

Human creativity follows a perceptual process, moving from abstract ideas to finer details during creation. While 3D generative models have advanced dramatically, models specifically designed to assist human imagination in 3D creation -- particularly for detailing abstractions from coarse to fine -- have not been explored. We propose a framework that enables intuitive and interactive 3D shape generation by iteratively splitting bounding boxes to refine the set of bounding boxes. The main technical components of our framework are two generative models: the box-splitting generative model and the box-to-shape generative model. The first model, named BoxSplitGen, generates a collection of 3D part bounding boxes with varying granularity by iteratively splitting coarse bounding boxes. It utilizes part bounding boxes created through agglomerative merging and learns the reverse of the merging process -- the splitting sequences. The model consists of two main components: the first learns the categorical distribution of the box to be split, and the second learns the distribution of the two new boxes, given the set of boxes and the indication of which box to split. The second model, the box-to-shape generative model, is trained by leveraging the 3D shape priors learned by an existing 3D diffusion model while adapting the model to incorporate bounding box conditioning. In our experiments, we demonstrate that the box-splitting generative model outperforms token prediction models and the inpainting approach with an unconditional diffusion model. Also, we show that our box-to-shape model, based on a state-of-the-art 3D diffusion model, provides superior results compared to a previous model.

cs.CV↗

arXiv2026

Covariant Chu-Kovasznay Decomposition: Resolving Thermodynamic Ambiguity in Compressible Flows

We establish the Covariant Chu--Kovasznay Decomposition (CCKD), a geometric framework that resolves thermodynamic ambiguity in compressible mode content by formulating the decomposition on the effective acoustic spacetime. Enforcing orthogonality in the covariant Chu energy norm, we show that shock--turbulence interaction, often treated as a scattering source, is, in the idealized linear, inviscid setting, a near-unitary (Chu-isometric) scattering map constrained by conservation of covariant Chu-energy flux. In the canonical Shu-Osher problem, CCKD characterizes the shock as a thermo-acoustic lens, mathematically demonstrating that the transfer of entropy fluctuations into sound follows a geometric blue-shift ($k_{\mathrm{out}}=\Lambda k_{\mathrm{in}}$) analogous to gravitational blue-shift. Thus, while the mean flow produces entropy across the shock, the fluctuation mapping is information-preserving on the retained subspace; practical information loss arises from noise, truncation, and model mismatch, not shock physics.

physics.flu-dyn↗

arXiv2026

Hierarchical Subcode Ensemble Decoding of Polar Codes

Subcode-ensemble decoders improve iterative decoding by running multiple decoders in parallel over carefully chosen subcodes, increasing the likelihood that at least one decoder avoids the dominant trapping structures. Achieving strong diversity gains, however, requires constructing many subcodes that satisfy a linear covering property-yet existing approaches lack a systematic way to scale the ensemble size while preserving this property. This paper introduces hierarchical subcode ensemble decoding (HSCED), a new ensemble decoding framework that expands the number of constituent decoders while still guaranteeing linear covering. The key idea is to recursively generate subcode parity constraints in a hierarchical structure so that coverage is maintained at every level, enabling large ensembles with controlled complexity. To demonstrate its effectiveness, we apply HSCED to belief propagation (BP) decoding of polar codes, where dense parity-check matrices induce severe stopping-set effects that limit conventional BP. Simulations confirm that HSCED delivers significant block-error-rate improvements over standard BP and conventional subcode-ensemble decoding under the same decoding-latency constraint.

cs.IT↗

arXiv2026

Covariant Helmholtz-Hodge Decomposition: Resolving Spurious Vorticity via Acoustic Geometry

The separation of acoustic and vortical fluctuations in compressible turbulence becomes ambiguous in thermodynamically inhomogeneous media, where refraction by entropy gradients and shocks can be misclassified as solenoidal content by Euclidean post-processing. We introduce a covariant Helmholtz--Hodge decomposition (CHHD) with respect to an effective acoustic metric, which identifies the irrotational (potential) component with the exact part of the metric-dual velocity one-form. Thermal refraction and shock-induced bending are absorbed into the induced curvature, ensuring that such geometric variations are not misidentified as physical vorticity. For canonical entropy-spot refraction and normal-shock discontinuities, Euclidean Helmholtz--Hodge and momentum-potential post-processing produce significant leakage in the refracting/discontinuous region, whereas the covariant splitting remains at the numerical noise floor (typically $\lesssim 10^{-12}$) throughout the domain, demonstrating robustness even at the sonic horizon, where the Euclidean metric singularity typically causes catastrophic error amplification. This geometric framework for velocity fields resolves the ambiguity of irrotational motion in inhomogeneous media and establishes a necessary foundation for future generalizations to full thermodynamic state vectors.

physics.flu-dyn↗

arXiv2025

Perspective-Aware Reasoning in Vision-Language Models via Mental Imagery Simulation

We present a framework for perspective-aware reasoning in vision-language models (VLMs) through mental imagery simulation. Perspective-taking, the ability to perceive an environment or situation from an alternative viewpoint, is a key benchmark for human-level visual understanding, essential for environmental interaction and collaboration with autonomous agents. Despite advancements in spatial reasoning within VLMs, recent research has shown that modern VLMs significantly lack perspective-aware reasoning capabilities and exhibit a strong bias toward egocentric interpretations. To bridge the gap between VLMs and human perception, we focus on the role of mental imagery, where humans perceive the world through abstracted representations that facilitate perspective shifts. Motivated by this, we propose a framework for perspective-aware reasoning, named Abstract Perspective Change (APC), that effectively leverages vision foundation models, such as object detection, segmentation, and orientation estimation, to construct scene abstractions and enable perspective transformations. Our experiments on synthetic and real-image benchmarks, compared with various VLMs, demonstrate significant improvements in perspective-aware reasoning with our framework, further outperforming fine-tuned spatial reasoning models and novel-view-synthesis-based approaches.

cs.CV↗

arXiv2024

SHARE: Shared Memory-Aware Open-Domain Long-Term Dialogue Dataset Constructed from Movie Script

Shared memories between two individuals strengthen their bond and are crucial for facilitating their ongoing conversations. This study aims to make long-term dialogue more engaging by leveraging these shared memories. To this end, we introduce a new long-term dialogue dataset named SHARE, constructed from movie scripts, which are a rich source of shared memories among various relationships. Our dialogue dataset contains the summaries of persona information and events of two individuals, as explicitly revealed in their conversation, along with implicitly extractable shared memories. We also introduce EPISODE, a long-term dialogue framework based on SHARE that utilizes shared experiences between individuals. Through experiments using SHARE, we demonstrate that shared memories between two individuals make long-term dialogues more engaging and sustainable, and that EPISODE effectively manages shared memories during dialogue. Our dataset and code are available at https://github.com/e1kim/SHARE.

cs.CL↗

arXiv2024

Automatic Speech Recognition System-Independent Word Error Rate Estimation

Word error rate (WER) is a metric used to evaluate the quality of transcriptions produced by Automatic Speech Recognition (ASR) systems. In many applications, it is of interest to estimate WER given a pair of a speech utterance and a transcript. Previous work on WER estimation focused on building models that are trained with a specific ASR system in mind (referred to as ASR system-dependent). These are also domain-dependent and inflexible in real-world applications. In this paper, a hypothesis generation method for ASR System-Independent WER estimation (SIWE) is proposed. In contrast to prior work, the WER estimators are trained using data that simulates ASR system output. Hypotheses are generated using phonetically similar or linguistically more likely alternative words. In WER estimation experiments, the proposed method reaches a similar performance to ASR system-dependent WER estimators on in-domain data and achieves state-of-the-art performance on out-of-domain data. On the out-of-domain data, the SIWE model outperformed the baseline estimators in root mean square error and Pearson correlation coefficient by relative 17.58% and 18.21%, respectively, on Switchboard and CALLHOME. The performance was further improved when the WER of the training set was close to the WER of the evaluation dataset.

cs.CL↗

arXiv2024

SignSGD with Federated Voting

Distributed learning is commonly used for accelerating model training by harnessing the computational capabilities of multiple-edge devices. However, in practical applications, the communication delay emerges as a bottleneck due to the substantial information exchange required between workers and a central parameter server. SignSGD with majority voting (signSGD-MV) is an effective distributed learning algorithm that can significantly reduce communication costs by one-bit quantization. However, due to heterogeneous computational capabilities, it fails to converge when the mini-batch sizes differ among workers. To overcome this, we propose a novel signSGD optimizer with \textit{federated voting} (signSGD-FV). The idea of federated voting is to exploit learnable weights to perform weighted majority voting. The server learns the weights assigned to the edge devices in an online fashion based on their computational capabilities. Subsequently, these weights are employed to decode the signs of the aggregated local gradients in such a way to minimize the sign decoding error probability. We provide a unified convergence rate analysis framework applicable to scenarios where the estimated weights are known to the parameter server either perfectly or imperfectly. We demonstrate that the proposed signSGD-FV algorithm has a theoretical convergence guarantee even when edge devices use heterogeneous mini-batch sizes. Experimental results show that signSGD-FV outperforms signSGD-MV, exhibiting a faster convergence rate, especially in heterogeneous mini-batch sizes.

cs.LG↗

arXiv2024

SignSGD with Federated Defense: Harnessing Adversarial Attacks through Gradient Sign Decoding

Distributed learning is an effective approach to accelerate model training using multiple workers. However, substantial communication delays emerge between workers and a parameter server due to massive costs associated with communicating gradients. SignSGD with majority voting (signSGD-MV) is a simple yet effective optimizer that reduces communication costs through one-bit quantization, yet the convergence rates considerably decrease as adversarial workers increase. In this paper, we show that the convergence rate is invariant as the number of adversarial workers increases, provided that the number of adversarial workers is smaller than that of benign workers. The key idea showing this counter-intuitive result is our novel signSGD with federated defense (signSGD-FD). Unlike the traditional approaches, signSGD-FD exploits the gradient information sent by adversarial workers with the proper weights, which are obtained through gradient sign decoding. Experimental results demonstrate signSGD-FD achieves superior convergence rates over traditional algorithms in various adversarial attack scenarios.

cs.LG↗

arXiv2023

Posterior Distillation Sampling

We introduce Posterior Distillation Sampling (PDS), a novel optimization method for parametric image editing based on diffusion models. Existing optimization-based methods, which leverage the powerful 2D prior of diffusion models to handle various parametric images, have mainly focused on generation. Unlike generation, editing requires a balance between conforming to the target attribute and preserving the identity of the source content. Recent 2D image editing methods have achieved this balance by leveraging the stochastic latent encoded in the generative process of diffusion models. To extend the editing capabilities of diffusion models shown in pixel space to parameter space, we reformulate the 2D image editing method into an optimization form named PDS. PDS matches the stochastic latents of the source and the target, enabling the sampling of targets in diverse parameter spaces that align with a desired attribute while maintaining the source's identity. We demonstrate that this optimization resembles running a generative process with the target attribute, but aligning this process with the trajectory of the source's generative process. Extensive editing results in Neural Radiance Fields and Scalable Vector Graphics representations demonstrate that PDS is capable of sampling targets to fulfill the aforementioned balance across various parameter spaces.

cs.CV↗

arXiv2023

Fast Word Error Rate Estimation Using Self-Supervised Representations for Speech and Text

Word error rate (WER) estimation aims to evaluate the quality of an automatic speech recognition (ASR) system's output without requiring ground-truth labels. This task has gained increasing attention as advanced ASR systems are trained on large amounts of data. In this context, the computational efficiency of a WER estimator becomes essential in practice. However, previous works have not prioritised this aspect. In this paper, a Fast estimator for WER (Fe-WER) is introduced, utilizing average pooling over self-supervised learning representations for speech and text. Our results demonstrate that Fe-WER outperformed a baseline relatively by 14.10% in root mean square error and 1.22% in Pearson correlation coefficient on Ted-Lium3. Moreover, a comparative analysis of the distributions of target WER and WER estimates was conducted, including an examination of the average values per speaker. Lastly, the inference speed was approximately 3.4 times faster in the real-time factor.

eess.AS↗