arXiv subjects

Sungjin Park

Publications and source records attributed to Sungjin Park.

At least 19 recordsLinked to original sources

arXiv2026

Universality of the cokernels of random $p$-adic matrices with inhomogeneously balanced columns

In this paper, we prove universality of the distribution of the cokernels of random $p$-adic matrices with inhomogeneously balanced columns. More precisely, let $u \ge 0$ be an integer, and for each positive integer $n$, let $A(n)$ be a random $n \times (n+u)$ matrix over $\mathbb{Z}_p$ whose $i$-th column is $α_n(i)$-balanced. We prove that if $\sum_{i=1}^{n+u} \exp(-εα_n(i)n) \to 0$ as $n \to \infty$ for every $ε>0$, then the cokernels of $A(n)$ converge in distribution, as $n \to \infty$, to the same limiting law as the cokernels of Haar-random $n \times (n+u)$ matrices over $\mathbb{Z}_p$. This extends a universality theorem of Nguyen and Wood to random $p$-adic matrices with inhomogeneously balanced columns.

math.NT↗

arXiv2026

Enhancing LLMs' Clinical Reasoning with Real-World Data from a Nationwide Sepsis Registry

Although large language models (LLMs) have demonstrated impressive reasoning capabilities across general domains, their effectiveness in real-world clinical practice remains limited. This is likely due to their insufficient exposure to real-world clinical data during training, as such data is typically not included due to privacy concerns. To address this, we propose enhancing the clinical reasoning capabilities of LLMs by leveraging real-world clinical data. We constructed reasoning-intensive questions from a nationwide sepsis registry and fine-tuned Phi-4 on these questions using reinforcement learning, resulting in C-Reason. C-Reason exhibited strong clinical reasoning capabilities on the in-domain test set, as evidenced by both quantitative metrics and expert evaluations. Furthermore, its enhanced reasoning capabilities generalized to a sepsis dataset involving different tasks and patient cohorts, an open-ended consultations on antibiotics use task, and other diseases. Future research should focus on training LLMs with large-scale, multi-disease clinical datasets to develop more powerful, general-purpose clinical reasoning models.

cs.AI↗

arXiv2026

From Ambiguity to Accuracy: The Transformative Effect of Coreference Resolution on Retrieval-Augmented Generation systems

Retrieval-Augmented Generation (RAG) has emerged as a crucial framework in natural language processing (NLP), improving factual consistency and reducing hallucinations by integrating external document retrieval with large language models (LLMs). However, the effectiveness of RAG is often hindered by coreferential complexity in retrieved documents, introducing ambiguity that disrupts in-context learning. In this study, we systematically investigate how entity coreference affects both document retrieval and generative performance in RAG-based systems, focusing on retrieval relevance, contextual understanding, and overall response quality. We demonstrate that coreference resolution enhances retrieval effectiveness and improves question-answering (QA) performance. Through comparative analysis of different pooling strategies in retrieval tasks, we find that mean pooling demonstrates superior context capturing ability after applying coreference resolution. In QA tasks, we discover that smaller models benefit more from the disambiguation process, likely due to their limited inherent capacity for handling referential ambiguity. With these findings, this study aims to provide a deeper understanding of the challenges posed by coreferential complexity in RAG, providing guidance for improving retrieval and generation in knowledge-intensive AI applications.

cs.CL↗

arXiv2026

Generating Multi-Table Time Series EHR from Latent Space with Minimal Preprocessing

Electronic Health Records (EHR) are time-series relational databases that record patient interactions and medical events over time, serving as a critical resource for healthcare research and applications. However, privacy concerns and regulatory restrictions limit the sharing and utilization of such sensitive data, necessitating the generation of synthetic EHR datasets. Unlike previous EHR synthesis methods, which typically generate medical records consisting of expert-chosen features (e.g. a few vital signs or structured codes only), we introduce RawMed, the first framework to synthesize multi-table, time-series EHR data that closely resembles raw EHRs. Using text-based representation and compression techniques, RawMed captures complex structures and temporal dynamics with minimal preprocessing. We also propose a new evaluation framework for multi-table time-series synthetic EHRs, assessing distributional similarity, inter-table relationships, temporal dynamics, and privacy. Validated on two open-source EHR datasets, RawMed outperforms baseline models in fidelity and utility. The code is available at https://github.com/eunbyeol-cho/RawMed.

cs.LG↗

arXiv2025

A High-Quality Thermoelectric Material Database with Self-Consistent ZT Filtering

This study presents a curated thermoelectric material database, teMatDb, constructed by digitizing literature-reported data. It includes temperature-dependent thermoelectric properties (TEPs), Seebeck coefficient, electrical resistivity, thermal conductivity, and figure of merit (ZT), along with metadata on materials and their corresponding publications. A self-consistent ZT (Sc-ZT) filter set was developed to measure ZT errors by comparing reported ZT's from figures with ZT's recalculated from digitized TEPs. Using this Sc-ZT protocol, we generated tMatDb272, comprising 14,717 temperature-property pairs from 272 high-quality TEP sets across 262 publications. The method identifies various types of ZT errors, such as resolution error, publication bias, ZT overestimation, interpolation and extrapolation error, and digitization noise, and excludes inconsistent samples from the dataset. teMatDb272 and the Sc-ZT filtering framework offer a robust dataset for data-driven and machine-learning-based materials design, device modeling, and future thermoelectric research.

cond-mat.mtrl-sci↗

arXiv2025

Asymmetric Electronic Band Alignment and Potentially Enhanced Thermoelectric Properties in Phase-Separated Mg2X (X=Si,Ge,Sn) Alloys

The Mg2X (X=Si, Ge, Sn) based alloy is an eco-friendly thermoelectric material for mid-temperature applications. The Mg2Si1-xSnx and Mg2Ge1-xSnx alloys can be phase-separated into Si(Ge)- and Sn-rich phases during material synthesis, leading to a nanocomposite with locally varying electronic band structure. First-principles calculations reveal that the valence band offset is eight-times larger than the conduction band offset at the interface between Si- and Sn-rich phases for x=0.6, showing type-I and asymmetric band alignment (0.092 eV versus 0.013 eV). Using Boltzmann transport theory and thermionic emission calculations, we show that the large valence band energy discontinuity could allow for energy filtering effects to take place that can potentially increase the power factor substantially in the p-type material system if designed appropriately.

cond-mat.mtrl-sci↗

arXiv2024

Ensembling Large Language Models with Process Reward-Guided Tree Search for Better Complex Reasoning

Despite recent advances in large language models, open-source models often struggle to consistently perform well on complex reasoning tasks. Existing ensemble methods, whether applied at the token or output levels, fail to address these challenges. In response, we present Language model Ensemble with Monte Carlo Tree Search (LE-MCTS), a novel framework for process-level ensembling of language models. LE-MCTS formulates step-by-step reasoning with an ensemble of language models as a Markov decision process. In this framework, states represent intermediate reasoning paths, while actions consist of generating the next reasoning step using one of the language models selected from a predefined pool. Guided by a process-based reward model, LE-MCTS performs a tree search over the reasoning steps generated by different language models, identifying the most accurate reasoning chain. Experimental results on five mathematical reasoning benchmarks demonstrate that our approach outperforms both single language model decoding algorithms and language model ensemble methods. Notably, LE-MCTS improves performance by 3.6% and 4.3% on the MATH and MQA datasets, respectively, highlighting its effectiveness in solving complex reasoning problems.

cs.CL↗

arXiv2024

Multimodal Transformer With a Low-Computational-Cost Guarantee

Transformer-based models have significantly improved performance across a range of multimodal understanding tasks, such as visual question answering and action recognition. However, multimodal Transformers significantly suffer from a quadratic complexity of the multi-head attention with the input sequence length, especially as the number of modalities increases. To address this, we introduce Low-Cost Multimodal Transformer (LoCoMT), a novel multimodal attention mechanism that aims to reduce computational cost during training and inference with minimal performance loss. Specifically, by assigning different multimodal attention patterns to each attention head, LoCoMT can flexibly control multimodal signals and theoretically ensures a reduced computational cost compared to existing multimodal Transformer variants. Experimental results on two multimodal datasets, namely Audioset and MedVidCL demonstrate that LoCoMT not only reduces GFLOPs but also matches or even outperforms established models.

cs.LG↗

arXiv2023

FactKG: Fact Verification via Reasoning on Knowledge Graphs

In real world applications, knowledge graphs (KG) are widely used in various domains (e.g. medical applications and dialogue agents). However, for fact verification, KGs have not been adequately utilized as a knowledge source. KGs can be a valuable knowledge source in fact verification due to their reliability and broad applicability. A KG consists of nodes and edges which makes it clear how concepts are linked together, allowing machines to reason over chains of topics. However, there are many challenges in understanding how these machine-readable concepts map to information in text. To enable the community to better use KGs, we introduce a new dataset, FactKG: Fact Verification via Reasoning on Knowledge Graphs. It consists of 108k natural language claims with five types of reasoning: One-hop, Conjunction, Existence, Multi-hop, and Negation. Furthermore, FactKG contains various linguistic patterns, including colloquial style claims as well as written style claims to increase practicality. Lastly, we develop a baseline approach and analyze FactKG over these reasoning types. We believe FactKG can advance both reliability and practicality in KG-based fact verification.

cs.CL↗

arXiv2022

Do Language Models Understand Measurements?

Recent success of pre-trained language models (PLMs) has stimulated interest in their ability to understand and work with numbers. Yet, the numerical reasoning over measurements has not been formally studied despite their importance. In this study, we show that PLMs lack the capability required for reasoning over measurements. Furthermore, we find that a language model trained on a measurement-rich corpus shows better performance on understanding measurements. We propose a simple embedding strategy to better distinguish between numbers and units, which leads to a significant improvement in the probing tasks.

cs.CL↗

arXiv2022

Unconditional Image-Text Pair Generation with Multimodal Cross Quantizer

Although deep generative models have gained a lot of attention, most of the existing works are designed for unimodal generation. In this paper, we explore a new method for unconditional image-text pair generation. We design Multimodal Cross-Quantization VAE (MXQ-VAE), a novel vector quantizer for joint image-text representations, with which we discover that a joint image-text representation space is effective for semantically consistent image-text pair generation. To learn a multimodal semantic correlation in a quantized space, we combine VQ-VAE with a Transformer encoder and apply an input masking strategy. Specifically, MXQ-VAE accepts a masked image-text pair as input and learns a quantized joint representation space, so that the input can be converted to a unified code sequence, then we perform unconditional image-text pair generation with the code sequence. Extensive experiments show the correlation between the quantized joint space and the multimodal generation capability on synthetic and real-world datasets. In addition, we demonstrate the superiority of our approach in these two aspects over several baselines. The source code is publicly available at: https://github.com/ttumyche/MXQ-VAE.

cs.CV↗

arXiv2022

Graph-Text Multi-Modal Pre-training for Medical Representation Learning

As the volume of Electronic Health Records (EHR) sharply grows, there has been emerging interest in learning the representation of EHR for healthcare applications. Representation learning of EHR requires appropriate modeling of the two dominant modalities in EHR: structured data and unstructured text. In this paper, we present MedGTX, a pre-trained model for multi-modal representation learning of the structured and textual EHR data. MedGTX uses a novel graph encoder to exploit the graphical nature of structured EHR data, and a text encoder to handle unstructured text, and a cross-modal encoder to learn a joint representation space. We pre-train our model through four proxy tasks on MIMIC-III, an open-source EHR data, and evaluate our model on two clinical benchmarks and three novel downstream tasks which tackle real-world problems in EHR data. The results consistently show the effectiveness of pre-training the model for joint representation of both structured and unstructured information from EHR. Given the promising performance of MedGTX, we believe this work opens a new door to jointly understanding the two fundamental modalities of EHR data.

cs.CL↗

arXiv2021

FreeTalky: Don't Be Afraid! Conversations Made Easier by a Humanoid Robot using Persona-based Dialogue

We propose a deep learning-based foreign language learning platform, named FreeTalky, for people who experience anxiety dealing with foreign languages, by employing a humanoid robot NAO and various deep learning models. A persona-based dialogue system that is embedded in NAO provides an interesting and consistent multi-turn dialogue for users. Also, an grammar error correction system promotes improvement in grammar skills of the users. Thus, our system enables personalized learning based on persona dialogue and facilitates grammar learning of a user using grammar error feedback. Furthermore, we verified whether FreeTalky provides practical help in alleviating xenoglossophobia by replacing the real human in the conversation with a NAO robot, through human evaluation.

cs.CL↗

arXiv2021

Real-time Denoising and Dereverberation with Tiny Recurrent U-Net

Modern deep learning-based models have seen outstanding performance improvement with speech enhancement tasks. The number of parameters of state-of-the-art models, however, is often too large to be deployed on devices for real-world applications. To this end, we propose Tiny Recurrent U-Net (TRU-Net), a lightweight online inference model that matches the performance of current state-of-the-art models. The size of the quantized version of TRU-Net is 362 kilobytes, which is small enough to be deployed on edge devices. In addition, we combine the small-sized model with a new masking method called phase-aware $β$-sigmoid mask, which enables simultaneous denoising and dereverberation. Results of both objective and subjective evaluations have shown that our model can achieve competitive performance with the current state-of-the-art models on benchmark datasets using fewer parameters by orders of magnitude.

cs.SD↗

arXiv2020

Multi-View Attention Network for Visual Dialog

Visual dialog is a challenging vision-language task in which a series of questions visually grounded by a given image are answered. To resolve the visual dialog task, a high-level understanding of various multimodal inputs (e.g., question, dialog history, and image) is required. Specifically, it is necessary for an agent to 1) determine the semantic intent of question and 2) align question-relevant textual and visual contents among heterogeneous modality inputs. In this paper, we propose Multi-View Attention Network (MVAN), which leverages multiple views about heterogeneous inputs based on attention mechanisms. MVAN effectively captures the question-relevant information from the dialog history with two complementary modules (i.e., Topic Aggregation and Context Matching), and builds multimodal representations through sequential alignment processes (i.e., Modality Alignment). Experimental results on VisDial v1.0 dataset show the effectiveness of our proposed model, which outperforms the previous state-of-the-art methods with respect to all evaluation metrics.

cs.AI↗

arXiv2020

Native point defects and low $p$-doping efficiency in $Mg_2 (Si,Sn)$ solid solutions: A hybrid-density functional study

We perform hybrid-density functional calculations to investigate the charged defect formation energy of native point defects in $Mg_2 Si$, $Mg_2 Sn$, and their solid solutions. The band gap correction by hybrid-density functional is found to be critical to determine the charged defect density in these materials. For $Mg_2 Si$, $Mg$ interstitials are dominant and provide unintentional $n$-type conductivity. Additionally, as the $Mg$ vacancies can dominate in $Mg$-poor $Mg_2 Sn$, $p$-type conductivity is possible for $Mg_2 Sn$. However, the existence of low formation energy defects such as $Mg_{Sn}^{1+}$ and $I_{Mg}^{2+}$ in $Mg_2 Sn$ and their diffusion can cause severe charge compensation of hole carriers resulting in low $p$-type doping efficiency and thermal degradation. Our results indicate that, in addition to the extrinsic doping strategy, alloying of $Mg_2 Si$ with $Mg_2 Sn$ under $Mg$-poor conditions would be necessary to enhance the $p$-type conductivity with less charge compensation.

cond-mat.mtrl-sci↗

arXiv2020

Off-centered Pb interstitials in PbTe

In this work, we calculate the defect properties of low-symmetry Pb interstitials in PbTe using first-principles density-functional theory calculations. We break the symmetry imposed on on-centered interstitial defects and show that the lowest ground state of Pb interstitial defects is off-centered along the [111] directions. Due to the four multi-stable structures with low defect formation energies, the defect density of Pb interstitials is expected to be ~5.6 times larger than previous predictions when PbTe is synthesized at 900 K. In contrast to the on-centered Pbinterstitial, the off-centered Pb interstitials in PbTe can exhibit long-range lattice relaxation toward [111] direction beyond distance of 1 nm, indicating the potential formation of weak local dipoles. This result provides an alternative explanation for the emphanitic anharmonicity of PbTe.

cond-mat.mtrl-sci↗

arXiv2020

Semi-supervised Disentanglement with Independent Vector Variational Autoencoders

We aim to separate the generative factors of data into two latent vectors in a variational autoencoder. One vector captures class factors relevant to target classification tasks, while the other vector captures style factors relevant to the remaining information. To learn the discrete class features, we introduce supervision using a small amount of labeled data, which can simply yet effectively reduce the effort required for hyperparameter tuning performed in existing unsupervised methods. Furthermore, we introduce a learning objective to encourage statistical independence between the vectors. We show that (i) this vector independence term exists within the result obtained on decomposing the evidence lower bound with multiple latent vectors, and (ii) encouraging such independence along with reducing the total correlation within the vectors enhances disentanglement performance. Experiments conducted on several image datasets demonstrate that the disentanglement achieved via our method can improve classification performance and generation controllability.

cs.LG↗