arXiv subjects

Dongsoo Han

Publications and source records attributed to Dongsoo Han.

6 recordsLinked to original sources

arXiv2026

PedestrianDiffusion: Multimodal Generative Denoising and Dense State Estimation for Inertial Navigation

The accuracy of consumer-grade inertial navigation is bottlenecked by the stochastic noise of Micro-Electro-Mechanical Systems (MEMS). Traditional deterministic neural architectures often succumb to ``estimation jittering,'' sacrificing high-frequency kinematic fidelity for numerical stability. We propose PedestrianDiffusion, a multimodal spectral-domain generative framework reformulating dense 6D state estimation as a continuous conditional denoising process. By operating in the frequency domain, our formulation bounds the spectral covariance, acting as a mathematical preconditioner to stabilize the reverse diffusion trajectory. Furthermore, we introduce a zero-shot semantic conditioning mechanism leveraging vision-language embeddings as categorical priors to generalize across heterogeneous sensor noise profiles. To address the computational intractability of generative tracking, we deploy a single-step deterministic probability flow ODE solver ($T=1$). This yields high-capacity asynchronous batch trajectory refinement, establishing the viability of generative architectures for asynchronous batch trajectory refinement on edge hardware. Extensive evaluations on the OxIOD, RIDI, RoNIN, and TLIO benchmarks demonstrate that PedestrianDiffusion achieves state-of-the-art performance, exhibiting unprecedented robustness to impulse perturbations and coupled 6D kinematic drift. This work provides a rigorous algorithmic blueprint for next-generation Neural Inertial Measurement Units (N-IMUs).

cs.LG↗

arXiv2026

AI-Augmented Peer Review and Scientific Productivity: A Cross-Country Panel and SEM Analysis

This study empirically investigates the impact of AI-augmented peer review systems on scientific productivity using panel data from OECD countries. While prior research has highlighted inefficiencies in traditional peer review, little empirical work has quantified the systemic impact of AI integration at the national level. We construct a novel AI Review Capability Index (AIRC) and examine its effects on research productivity, reproducibility, and innovation output. Using fixed-effects regression and structural equation modeling (SEM), we show that AI-assisted evaluation significantly enhances productivity and reduces variance in research quality. Results indicate that a one standard deviation increase in AIRC is associated with an 18-25% increase in scientific productivity, mediated through improvements in review efficiency and reproducibility. This paper provides the first cross-country empirical validation of AI-augmented scientific evaluation systems and contributes to the emerging literature on AI as a structural driver of knowledge production.

cs.CY↗

arXiv2026

AI Civilization and the Transformation of Work

The emergence of artificial intelligence and robotics is catalyzing a profound transformation in the nature of human labor, fueling a contentious debate about the future of employment. While prominent studies predict substantial job displacement due to automation, historical precedents from past technological revolutions suggest that innovation tends to expand, rather than shrink, the scope of economic activity and employment in the long run. This paper advances the thesis that the transition to an AI-civilization will fundamentally restructure the mechanisms of employment creation. We argue for a paradigm shift from a centralized model, where a limited number of organizations create jobs for the mass to a decentralized ecosystem where individuals are empowered to generate their own employment opportunities. This shift is enabled by AI-driven productivity augmentation, which dramatically lowers the barriers to creating economic value. Drawing on an analysis of economic history, contemporary data on labor market dynamics, and the growth of digital platforms, this paper posits that human-AI co-evolution will significantly increase individual productivity and open new frontiers of economic activity. We explore the implications of this structural transformation for education and workforce development, concluding that the focus must shift from rote knowledge accumulation to cultivating skills in human AI collaboration, creative problem-solving, and the design of novel economic domains. This paper contributes to the literature by offering a forward-looking framework that emphasizes the decentralizing potential of AI on labor markets, moving beyond the traditional displacement-versus-creation dichotomy.

cs.CY↗

arXiv2025

ConvXformer: Differentially Private Hybrid ConvNeXt-Transformer for Inertial Navigation

Data-driven inertial sequence learning has revolutionized navigation in GPS-denied environments, offering superior odometric resolution compared to traditional Bayesian methods. However, deep learning-based inertial tracking systems remain vulnerable to privacy breaches that can expose sensitive training data. \hl{Existing differential privacy solutions often compromise model performance by introducing excessive noise, particularly in high-frequency inertial measurements.} In this article, we propose ConvXformer, a hybrid architecture that fuses ConvNeXt blocks with Transformer encoders in a hierarchical structure for robust inertial navigation. We propose an efficient differential privacy mechanism incorporating adaptive gradient clipping and gradient-aligned noise injection (GANI) to protect sensitive information while ensuring model performance. Our framework leverages truncated singular value decomposition for gradient processing, enabling precise control over the privacy-utility trade-off. Comprehensive performance evaluations on benchmark datasets (OxIOD, RIDI, RoNIN) demonstrate that ConvXformer surpasses state-of-the-art methods, achieving more than 40% improvement in positioning accuracy while ensuring $(\epsilon,\delta)$-differential privacy guarantees. To validate real-world performance, we introduce the Mech-IO dataset, collected from the mechanical engineering building at KAIST, where intense magnetic fields from industrial equipment induce significant sensor perturbations. This demonstrated robustness under severe environmental distortions makes our framework well-suited for secure and intelligent navigation in cyber-physical systems.

cs.LG↗

arXiv2025

QSCL-EWIL: Quantum Stochastic Contrast Learning for Enhanced WiFi-Based Indoor Localization

WiFi-based indoor localization is essential for asset tracking, healthcare monitoring, and smart buildings. However, existing systems face challenges such as data variability, environmental noise, and difficulty detecting floor and building levels, compounded by limited labeled data and high received signal strength (RSS) collection costs. This paper introduces quantum stochastic contrast learning (QSCL), a novel framework grounded in rigorous theoretical foundations. We present four theorems and one lemma that establish the probabilistic augmentation, diversity enhancement, relationship preservation, and resilience of QSCL under quantum noise, supported by formal proofs. Leveraging these foundations, QSCL utilizes quantum computing (QC) to generate strong data augmentations with stochastic perturbations, enhancing data diversity, while classical weak augmentations provide subtle variations for robust feature learning. We propose a spatial temporal adaptive attention (STAA) encoder that integrates convolutional layers with adaptive attention mechanisms to capture spatial and temporal dependencies in sequential data. Furthermore, a bidirectional contrastive loss function is introduced to capture forward and reverse relationships between augmented views, ensuring robust representations. Comprehensive evaluations on the UJIIndoorLoc and UTSIndoorLoc datasets validate QSCL, demonstrating superior performance with reduced labeled data and resilience to quantum noise such as bit-flip, dephasing, and measurement noise. The proposed framework significantly improves localization accuracy, floor and building detection, and generalizability in challenging indoor environments.

quant-ph↗

arXiv2020

Multiview Variational Deep Learning with Application to Scalable Indoor Localization

Radio channel state information (CSI) measured with many receivers is a good resource for localizing a transmit device with machine learning with a discriminative model. However, CSI localization is nontrivial when the radio map is complicated, such as in building corridors. This paper introduces a view-selective deep learning (VSDL) system for indoor localization using CSI of WiFi. The multiview training with CSI obtained from multiple groups of access points (APs) generates latent features on supervised variational deep network. This information is then applied to an additional network for dominant view classification to enhance the regression accuracy of localization. As non-informative latent features from multiple views are rejected, we can achieve a localization accuracy of 0.77 m, which outperforms by 30 $\%$ the best known accuracy in practical applications in a real building environment. To the best of our knowledge, this is the first approach to apply variational inference and to construct a scalable system for radio localization. Furthermore, our work investigates a methodology for supervised learning with multiview data where informative and non-informative views coexist.

eess.SP↗