arXiv subjects

Jangwon Kim

Publications and source records attributed to Jangwon Kim.

9 recordsLinked to original sources

arXiv2026

CheXmix: Unified Generative Pretraining for Vision Language Models in Medical Imaging

Recent medical multimodal foundation models are built as multimodal LLMs (MLLMs) by connecting a CLIP-pretrained vision encoder to an LLM using LLaVA-style finetuning. This two-stage, decoupled approach introduces a projection layer that can distort visual features. This is especially concerning in medical imaging where subtle cues are essential for accurate diagnoses. In contrast, early-fusion generative approaches such as Chameleon eliminate the projection bottleneck by processing image and text tokens within a single unified sequence, enabling joint representation learning that leverages the inductive priors of language models. We present CheXmix, a unified early-fusion generative model trained on a large corpus of chest X-rays paired with radiology reports. We expand on Chameleon's autoregressive framework by introducing a two-stage multimodal generative pretraining strategy that combines the representational strengths of masked autoencoders with MLLMs. The resulting models are highly flexible, supporting both discriminative and generative tasks at both coarse and fine-grained scales. Our approach outperforms well-established generative models across all masking ratios by 6.0% and surpasses CheXagent by 8.6% on AUROC at high image masking ratios on the CheXpert classification task. We further inpaint images over 51.0% better than text-only generative models and outperform CheXagent by 45% on the GREEN metric for radiology report generation. These results demonstrate that CheXmix captures fine-grained information across a broad spectrum of chest X-ray tasks. Our code is at: https://github.com/StanfordMIMI/CheXmix.

cs.CV↗

arXiv2026

Delayed homomorphic reinforcement learning for environments with delayed feedback

Reinforcement learning in real-world systems often involves delayed feedback, which breaks the Markov assumption and impedes both learning and control. Canonical augmentation-based approaches cause state-space explosion, which imposes a severe sample-complexity burden. Despite recent progress, state-of-the-art augmentation-based baselines either mainly alleviate the burden on the critic or rely on non-unified treatments for the actor and critic. In this study, we propose delayed homomorphic reinforcement learning (DHRL), a framework grounded in MDP homomorphisms that defines a belief-equivalence relation over the augmented state space to collapse control-redundant augmented states. In principle, this yields exact abstraction under deterministic dynamics and approximate abstraction under stochastic dynamics, enabling both the actor and critic to benefit from a structured abstraction mechanism. In finite domains, exact abstraction preserves optimality and recovers the delay-free sample-complexity order, whereas approximate abstraction admits a value-loss bound on the resulting policy. For continuous domains, we introduce deep delayed homomorphic policy gradient (D$^2$HPG), a deep actor-critic instantiation of the DHRL framework. Experiments on continuous-control tasks in MuJoCo show that D$^2$HPG outperforms strong augmentation-based baselines.

cs.LG↗

arXiv2025

Optimizing Long-Form Clinical Text Generation with Claim-Based Rewards

Automating clinical documentation with large language models requires precise alignment with priorities such as completeness and factual grounding. We present an evaluation-integrated reinforcement learning framework for long-form clinical text generation that couples Group Relative Policy Optimization (GRPO) with DocLens, a claim-level evaluator that provides deterministic, dialogue-grounded rewards. Our method directly optimizes factual grounding and completeness without training a separate reward model or relying on human-authored references. Empirically, the approach improves clinical note quality and reduces training cost via a simple reward-gating strategy. An independent GPT-5 qualitative evaluation further supports these gains, showing higher preference for GRPO outputs in factuality, completeness, and brevity, with fewer omissions and hallucinations. Because the benchmarks are relatively clean and the base model already well aligned, these improvements likely represent a conservative lower bound. The framework is scalable to real-world settings and can incorporate custom objectives such as guideline adherence or billing preferences.

cs.CL↗

arXiv2025

RadAgents: Multimodal Agentic Reasoning for Chest X-ray Interpretation with Radiologist-like Workflows

Agentic systems offer a potential path to solve complex clinical tasks through collaboration among specialized agents, augmented by tool use and external knowledge bases. Nevertheless, for chest X-ray (CXR) interpretation, prevailing methods remain limited: (i) reasoning is frequently neither clinically interpretable nor aligned with guidelines, reflecting mere aggregation of tool outputs; (ii) multimodal evidence is insufficiently fused, yielding text-only rationales that are not visually grounded; and (iii) systems rarely detect or resolve cross-tool inconsistencies and provide no principled verification mechanisms. To bridge the above gaps, we present RadAgents, a multi-agent framework that couples clinical priors with task-aware multimodal reasoning and encodes a radiologist-style workflow into a modular, auditable pipeline. In addition, we integrate grounding and multimodal retrieval-augmentation to verify and resolve context conflicts, resulting in outputs that are more reliable, transparent, and consistent with clinical practice.

cs.MA↗

arXiv2025

Reinforcement Learning via Conservative Agent for Environments with Random Delays

Real-world reinforcement learning applications are often hindered by delayed feedback from environments, which violates the Markov assumption and introduces significant challenges. Although numerous delay-compensating methods have been proposed for environments with constant delays, environments with random delays remain largely unexplored due to their inherent variability and unpredictability. In this study, we propose a simple yet robust agent for decision-making under random delays, termed the conservative agent, which reformulates the random-delay environment into its constant-delay equivalent. This transformation enables any state-of-the-art constant-delay method to be directly extended to the random-delay environments without modifying the algorithmic structure or sacrificing performance. We evaluate the conservative agent-based algorithm on continuous control tasks, and empirical results demonstrate that it significantly outperforms existing baseline algorithms in terms of asymptotic performance and sample efficiency.

cs.LG↗

arXiv2025

RedactOR: An LLM-Powered Framework for Automatic Clinical Data De-Identification

Ensuring clinical data privacy while preserving utility is critical for AI-driven healthcare and data analytics. Existing de-identification (De-ID) methods, including rule-based techniques, deep learning models, and large language models (LLMs), often suffer from recall errors, limited generalization, and inefficiencies, limiting their real-world applicability. We propose a fully automated, multi-modal framework, RedactOR for de-identifying structured and unstructured electronic health records, including clinical audio records. Our framework employs cost-efficient De-ID strategies, including intelligent routing, hybrid rule and LLM based approaches, and a two-step audio redaction approach. We present a retrieval-based entity relexicalization approach to ensure consistent substitutions of protected entities, thereby enhancing data coherence for downstream applications. We discuss key design desiderata, de-identification and relexicalization methodology, and modular architecture of RedactOR and its integration with the Oracle Health Clinical AI system. Evaluated on the i2b2 2014 De-ID dataset using standard metrics with strict recall, our approach achieves competitive performance while optimizing token usage to reduce LLM costs. Finally, we discuss key lessons and insights from deployment in real-world AI- driven healthcare data pipelines.

cs.AI↗

arXiv2025

DNA Nanotechnology for Superradiance

Superradiance, first proposed by Dicke in 1954, is a highly efficient quantum light source that differs from conventional spontaneous emission. Unlike typical spontaneous emission, where intensity scales linearly with the number of electric dipoles, superradiance exhibits an intensity that scales quadratically with the number of electric dipoles. Similarly, the decay rate also increases proportionally to the dipole numbers. This collective emission is especially powerful when it manifests as superfluorescence, a physical regime of superradiance where spontaneously emerging coherence is achieved by arranging excited dipoles in the same orientation within a volume much smaller than their emission wavelength. Numerous experimental strategies have been employed to generate superradiance, with one common approach being the use of stochastically formed aggregates of quantum dots and organic dyes. However, the inherent randomness in such systems prevents precise control over the number, spatial distribution, and relative orientation of the emitters. This often leads to non-uniform coupling strengths and parasitic dephasing effects, which make it difficult to predict the resulting quantum emission and limit its use in engineered devices. A deterministic platform that provides precise control over these parameters is therefore essential for realizing the full potential of superradiant systems. Here, we (i) specifically outline the advantages of DNA nanotechnology in tackling this challenge, (ii) discuss the reasons why superradiance has not yet been realized even with the state-of-the art DNA nanotechnology, and (iii) propose potential solutions for overcoming the current limitations.

quant-ph↗

arXiv2024

A dataset and benchmark for hospital course summarization with adapted large language models

Brief hospital course (BHC) summaries are clinical documents that summarize a patient's hospital stay. While large language models (LLMs) depict remarkable capabilities in automating real-world tasks, their capabilities for healthcare applications such as synthesizing BHCs from clinical notes have not been shown. We introduce a novel pre-processed dataset, the MIMIC-IV-BHC, encapsulating clinical note and brief hospital course (BHC) pairs to adapt LLMs for BHC synthesis. Furthermore, we introduce a benchmark of the summarization performance of two general-purpose LLMs and three healthcare-adapted LLMs. Using clinical notes as input, we apply prompting-based (using in-context learning) and fine-tuning-based adaptation strategies to three open-source LLMs (Clinical-T5-Large, Llama2-13B, FLAN-UL2) and two proprietary LLMs (GPT-3.5, GPT-4). We evaluate these LLMs across multiple context-length inputs using natural language similarity metrics. We further conduct a clinical study with five clinicians, comparing clinician-written and LLM-generated BHCs across 30 samples, focusing on their potential to enhance clinical decision-making through improved summary quality. We observe that the Llama2-13B fine-tuned LLM outperforms other domain-adapted models given quantitative evaluation metrics of BLEU and BERT-Score. GPT-4 with in-context learning shows more robustness to increasing context lengths of clinical note inputs than fine-tuned Llama2-13B. Despite comparable quantitative metrics, the reader study depicts a significant preference for summaries generated by GPT-4 with in-context learning compared to both Llama2-13B fine-tuned summaries and the original summaries, highlighting the need for qualitative clinical evaluation.

cs.CL↗

arXiv2020

Strongly adhesive dry transfer technique for van der Waals heterostructure

That one can stack van der Waals materials with atomically sharp interfaces has provided a new material platform of constructing heterostructures. The technical challenge of mechanical stacking is picking up the exfoliated atomically thin materials after mechanical exfoliation without chemical and mechanical degradation. Chemically inert hexagonal boron nitride (hBN) has been widely used for encapsulating and picking up vdW materials. However, due to the relatively weak adhesion of hBN, assembling vdW heterostructures based on hBN has been limited. We report a new dry transfer technique. We used two vdW semiconductors (ZnPS3 and CrPS4) to pick up and encapsulate layers for vdW heterostructures, which otherwise are known to be hard to fabricate. By combining with optimized polycaprolactone (PCL) providing strong adhesion, we demonstrated various vertical heterostructure devices, including quasi-2D superconducting NbSe2 Josephson junctions with atomically clean interface. The versatility of the PCL-based vdW stacking method provides a new route for assembling complex 2D vdW materials without interfacial degradation.

physics.app-ph↗