arXiv subjects

Donghyeon Lee

Publications and source records attributed to Donghyeon Lee.

15 recordsLinked to original sources

arXiv2026

Signatures of Massive Neutrinos in the Cosmic Web via Persistent Homology

We present the second paper in our program characterizing the impact of massive neutrinos on the multiscale cosmic web using global topology and persistent homology. Building on the methodology established in Paper I, based on discrete Morse theory, we analyze a subset of the Quijote simulations to compute persistent diagrams, Betti curves, and additional topological statistics for both dark matter and halo density fields, across redshifts z=0,1,2. A central result of our study is the first clear demonstration that apex points in persistent diagrams are especially sensitive to neutrino mass, with enhanced sensitivity for specific pairs of saddle points at high redshift. In addition, Betti curves from dark matter density fields broaden and flatten with increasing neutrino masses, exhibiting two characteristic density thresholds where Betti numbers remain invariant. These mass-dependent signatures are detectable at the few-percent level, even for $M_ν \sim 0.1$ eV, providing a robust, physically grounded probe of massive neutrinos in the cosmic web. While traditional two-point statistics encode only pairwise correlations and cannot fully break parameter degeneracies, persistent homology captures higher-order, multiscale information that can lift these degeneracies. Moreover, its high sensitivity to the sum of neutrino masses makes it a promising complement to conventional analyses. Our results thus establish a solid foundation for forward-modeling or emulator-based approaches using persistent homology and environment-based statistics to constrain neutrino mass - potentially enabling direct detection - and additional cosmological parameters, with immediate relevance for ongoing and upcoming galaxy surveys, including DESI, Euclid, and Rubin-LSST.

astro-ph.CO↗

arXiv2025

Development of an Agentic AI Model for NGS Downstream Analysis Targeting Researchers with Limited Biological Background

Next-Generation Sequencing (NGS) has become a cornerstone of genomic research, yet the complexity of downstream analysis-ranging from differential expression gene (DEG) identification to biological interpretations-remains a significant barrier for researchers lacking specialized computational and biological expertise. While recent studies have introduced AI agents for RNA-seq analysis, most focus on general workflows without offering tailored interpretations or guidance for novices. To address this gap, we developed an Agentic AI model designed to automate NGS downstream analysis, provide literature-backed interpretations, and autonomously recommend advanced analytical methods. Built on the Llama 3 70B Large Language Model (LLM) and a Retrieval-Augmented Generation (RAG) framework, the model is deployed as an interactive Streamlit web application. The system integrates standard bioinformatics tools (Biopython, GSEApy, gProfiler) to execute core analyses, including DEG identification, clustering, and pathway enrichment. Uniquely, the agent utilizes RAG to query PubMed via Entrez, synthesizing biological insights and validating hypotheses with current literature. In a case study using cancer-related dataset, the model successfully identified significant DEGs, visualized clinical correlations, and derived evidence-based insights (e.g., linking BRAF mutations to prognosis), subsequently executing advanced survival modeling upon user selection. This framework democratizes bioinformatics by enabling researchers with limited backgrounds to seamlessly transition from basic data processing to advanced hypothesis testing and validation.

q-bio.GN↗

arXiv2025

Leveraging Language Models and RAG for Efficient Knowledge Discovery in Clinical Environments

Large language models (LLMs) are increasingly recognized as valuable tools across the medical environment, supporting clinical, research, and administrative workflows. However, strict privacy and network security regulations in hospital settings require that sensitive data be processed within fully local infrastructures. Within this context, we developed and evaluated a retrieval-augmented generation (RAG) system designed to recommend research collaborators based on PubMed publications authored by members of a medical institution. The system utilizes PubMedBERT for domain-specific embedding generation and a locally deployed LLaMA3 model for generative synthesis. This study demonstrates the feasibility and utility of integrating domain-specialized encoders with lightweight LLMs to support biomedical knowledge discovery under local deployment constraints.

cs.CL↗

arXiv2025

CoTox: Chain-of-Thought-Based Molecular Toxicity Reasoning and Prediction

Drug toxicity remains a major challenge in pharmaceutical development. Recent machine learning models have improved in silico toxicity prediction, but their reliance on annotated data and lack of interpretability limit their applicability. This limits their ability to capture organ-specific toxicities driven by complex biological mechanisms. Large language models (LLMs) offer a promising alternative through step-by-step reasoning and integration of textual data, yet prior approaches lack biological context and transparent rationale. To address this issue, we propose CoTox, a novel framework that integrates LLM with chain-of-thought (CoT) reasoning for multi-toxicity prediction. CoTox combines chemical structure data, biological pathways, and gene ontology (GO) terms to generate interpretable toxicity predictions through step-by-step reasoning. Using GPT-4o, we show that CoTox outperforms both traditional machine learning and deep learning model. We further examine its performance across various LLMs to identify where CoTox is most effective. Additionally, we find that representing chemical structures with IUPAC names, which are easier for LLMs to understand than SMILES, enhances the model's reasoning ability and improves predictive performance. To demonstrate its practical utility in drug development, we simulate the treatment of relevant cell types with drug and incorporated the resulting biological context into the CoTox framework. This approach allow CoTox to generate toxicity predictions aligned with physiological responses, as shown in case study. This result highlights the potential of LLM-based frameworks to improve interpretability and support early-stage drug safety assessment. The code and prompt used in this work are available at https://github.com/dmis-lab/CoTox.

cs.LG↗

arXiv2025

DICE-BENCH: Evaluating the Tool-Use Capabilities of Large Language Models in Multi-Round, Multi-Party Dialogues

Existing function-calling benchmarks focus on single-turn interactions. However, they overlook the complexity of real-world scenarios. To quantify how existing benchmarks address practical applications, we introduce DICE-SCORE, a metric that evaluates the dispersion of tool-related information such as function name and parameter values throughout the dialogue. Analyzing existing benchmarks through DICE-SCORE reveals notably low scores, highlighting the need for more realistic scenarios. To address this gap, we present DICE-BENCH, a framework that constructs practical function-calling datasets by synthesizing conversations through a tool graph that maintains dependencies across rounds and a multi-agent system with distinct personas to enhance dialogue naturalness. The final dataset comprises 1,607 high-DICE-SCORE instances. Our experiments on 19 LLMs with DICE-BENCH show that significant advances are still required before such models can be deployed effectively in real-world settings. Our code and data are all publicly available: https://snuhcc.github.io/DICE-Bench/.

cs.CL↗

arXiv2025

ChroKnowledge: Unveiling Chronological Knowledge of Language Models in Multiple Domains

Large language models (LLMs) have brought significant changes to many aspects of our lives. However, assessing and ensuring their chronological knowledge remains challenging. Existing approaches fall short in addressing the temporal adaptability of knowledge, often relying on a fixed time-point view. To overcome this, we introduce ChroKnowBench, a benchmark dataset designed to evaluate chronologically accumulated knowledge across three key aspects: multiple domains, time dependency, temporal state. Our benchmark distinguishes between knowledge that evolves (e.g., personal history, scientific discoveries, amended laws) and knowledge that remain constant (e.g., mathematical truths, commonsense facts). Building on this benchmark, we present ChroKnowledge (Chronological Categorization of Knowledge), a novel sampling-based framework for evaluating LLMs' non-parametric chronological knowledge. Our evaluation led to the following observations: (1) The ability of eliciting temporal knowledge varies depending on the data format that model was trained on. (2) LLMs partially recall knowledge or show a cut-off at temporal boundaries rather than recalling all aspects of knowledge correctly. Thus, we apply our ChroKnowPrompt, an in-depth prompting to elicit chronological knowledge by traversing step-by-step through the surrounding time spans. We observe that it successfully recalls objects across both open-source and proprietary LLMs, demonstrating versatility, though it faces challenges with dynamic datasets and unstructured formats.

cs.CL↗

arXiv2025

ETHIC: Evaluating Large Language Models on Long-Context Tasks with High Information Coverage

Recent advancements in large language models (LLM) capable of processing extremely long texts highlight the need for a dedicated evaluation benchmark to assess their long-context capabilities. However, existing methods, like the needle-in-a-haystack test, do not effectively assess whether these models fully utilize contextual information, raising concerns about the reliability of current evaluation techniques. To thoroughly examine the effectiveness of existing benchmarks, we introduce a new metric called information coverage (IC), which quantifies the proportion of the input context necessary for answering queries. Our findings indicate that current benchmarks exhibit low IC; although the input context may be extensive, the actual usable context is often limited. To address this, we present ETHIC, a novel benchmark designed to assess LLMs' ability to leverage the entire context. Our benchmark comprises 1,986 test instances spanning four long-context tasks with high IC scores in the domains of books, debates, medicine, and law. Our evaluations reveal significant performance drops in contemporary LLMs, highlighting a critical challenge in managing long contexts. Our benchmark is available at https://github.com/dmis-lab/ETHIC.

cs.CL↗

arXiv2024

From Ceilings to Walls: Universal Dynamic Perching of Small Aerial Robots on Surfaces with Variable Orientations

This work demonstrates universal dynamic perching capabilities for quadrotors of various sizes and on surfaces with different orientations. By employing a non-dimensionalization framework and deep reinforcement learning, we systematically assessed how robot size and surface orientation affect landing capabilities. We hypothesized that maintaining geometric proportions across different robot scales ensures consistent perching behavior, which was validated in both simulation and experimental tests. Additionally, we investigated the effects of joint stiffness and damping in the landing gear on perching behaviors and performance. While joint stiffness had minimal impact, joint damping ratios influenced landing success under vertical approaching conditions. The study also identified a critical velocity threshold necessary for successful perching, determined by the robot's maneuverability and leg geometry. Overall, this research advances robotic perching capabilities, offering insights into the role of mechanical design and scaling effects, and lays the groundwork for future drone autonomy and operational efficiency in unstructured environments.

cs.RO↗

arXiv2023

Interpretable pap smear cell representation for cervical cancer screening

Screening is critical for prevention and early detection of cervical cancer but it is time-consuming and laborious. Supervised deep convolutional neural networks have been developed to automate pap smear screening and the results are promising. However, the interest in using only normal samples to train deep neural networks has increased owing to class imbalance problems and high-labeling costs that are both prevalent in healthcare. In this study, we introduce a method to learn explainable deep cervical cell representations for pap smear cytology images based on one class classification using variational autoencoders. Findings demonstrate that a score can be calculated for cell abnormality without training models with abnormal samples and localize abnormality to interpret our results with a novel metric based on absolute difference in cross entropy in agglomerative clustering. The best model that discriminates squamous cell carcinoma (SCC) from normals gives 0.908 +- 0.003 area under operating characteristic curve (AUC) and one that discriminates high-grade epithelial lesion (HSIL) 0.920 +- 0.002 AUC. Compared to other clustering methods, our method enhances the V-measure and yields higher homogeneity scores, which more effectively isolate different abnormality regions, aiding in the interpretation of our results. Evaluation using in-house and additional open dataset show that our model can discriminate abnormality without the need of additional training of deep models.

cs.CV↗

arXiv2023

Energy-Flexible Network (EF-Net) for Dual-Energy CT Image Reconstruction

In dual-energy computed tomography (DECT), the X-ray tube energy pair often changes depending on the target organ or patient obesity. In practice, it makes difficult to apply deep learning (DL) based algorithms for image reconstruction since most of the existing DL-based algorithms are trained to be used for dedicated X-ray tube energies. In this paper, we propose 1) an energy flexibility training (EFT) method, which makes a network applicable for data measured at various X-ray tube energies between two trained energies, and 2) an effective dual-domain convolutional neural network for image reconstruction. The proposed network is derived from the regularized version of the primal-dual hybrid gradient algorithm, so its architecture has an unfolded iterative dual-domain structure. For validation, we generated datasets from a lab-made polychromatic X-ray simulator. The proposed method showed promising results not only at the trained X-ray tube energies but also at untrained X-ray tube energies outperforming an iterative algorithm and other DL-based algorithms.

physics.med-ph↗

arXiv2020

STAR: Spatio-Temporal Prediction of Air Quality Using A Multimodal Approach

With the increase of global economic activities and high energy demand, many countries have raised concerns about air pollution. However, air quality prediction is a challenging issue due to the complex interaction of many factors. In this paper, we propose a multimodal approach for spatio-temporal air quality prediction. Our model learns the multimodal fusion of critical factors to predict future air quality levels. Based on the analyses of data, we also assessed the impacts of critical factors on air quality prediction. We conducted experiments on two real-world air pollution datasets. For Seoul dataset, our method achieved 11% and 8.2% improvement of the mean absolute error in long-term predictions of PM2.5 and PM10, respectively, compared to baselines. Our method also reduced the mean absolute error of PM2.5 predictions by 20% compared to the previous state-of-the-art results on China 1-year dataset.

eess.SP↗

arXiv2019

Deep Reinforcement Learning for Chatbots Using Clustered Actions and Human-Likeness Rewards

Training chatbots using the reinforcement learning paradigm is challenging due to high-dimensional states, infinite action spaces and the difficulty in specifying the reward function. We address such problems using clustered actions instead of infinite actions, and a simple but promising reward function based on human-likeness scores derived from human-human dialogue data. We train Deep Reinforcement Learning (DRL) agents using chitchat data in raw text---without any manual annotations. Experimental results using different splits of training data report the following. First, that our agents learn reasonable policies in the environments they get familiarised with, but their performance drops substantially when they are exposed to a test set of unseen dialogues. Second, that the choice of sentence embedding size between 100 and 300 dimensions is not significantly different on test data. Third, that our proposed human-likeness rewards are reasonable for training chatbots as long as they use lengthy dialogue histories of >=10 sentences.

cs.AI↗

arXiv2019

Ensemble-Based Deep Reinforcement Learning for Chatbots

Trainable chatbots that exhibit fluent and human-like conversations remain a big challenge in artificial intelligence. Deep Reinforcement Learning (DRL) is promising for addressing this challenge, but its successful application remains an open question. This article describes a novel ensemble-based approach applied to value-based DRL chatbots, which use finite action sets as a form of meaning representation. In our approach, while dialogue actions are derived from sentence clustering, the training datasets in our ensemble are derived from dialogue clustering. The latter aim to induce specialised agents that learn to interact in a particular style. In order to facilitate neural chatbot training using our proposed approach, we assume dialogue data in raw text only -- without any manually-labelled data. Experimental results using chitchat data reveal that (1) near human-like dialogue policies can be induced, (2) generalisation to unseen data is a difficult problem, and (3) training an ensemble of chatbot agents is essential for improved performance over using a single agent. In addition to evaluations using held-out data, our results are further supported by a human evaluation that rated dialogues in terms of fluency, engagingness and consistency -- which revealed that our proposed dialogue rewards strongly correlate with human judgements.

cs.AI↗

arXiv2018

A Study on Dialogue Reward Prediction for Open-Ended Conversational Agents

The amount of dialogue history to include in a conversational agent is often underestimated and/or set in an empirical and thus possibly naive way. This suggests that principled investigations into optimal context windows are urgently needed given that the amount of dialogue history and corresponding representations can play an important role in the overall performance of a conversational system. This paper studies the amount of history required by conversational agents for reliably predicting dialogue rewards. The task of dialogue reward prediction is chosen for investigating the effects of varying amounts of dialogue history and their impact on system performance. Experimental results using a dataset of 18K human-human dialogues report that lengthy dialogue histories of at least 10 sentences are preferred (25 sentences being the best in our experiments) over short ones, and that lengthy histories are useful for training dialogue reward predictors with strong positive correlations between target dialogue rewards and predicted ones.

cs.CL↗

arXiv2017

Arachneum: Blockchain meets Distributed Web

Appearance of Bitcoin raise up evolution in currency. Blockchain database also raise up possibilities to share important data between untrusted peers. In this paper, we propose Arachneum, a decentralized, distributed, and model-view-controller-based web service using blockchain and new privileges model. Therefore, we can enjoy the freedom against censorship of government or organizations, keep transparency of our web services, fork our web services, and create/read/update/delete(CRUD) all model-view-controller(MVC) components dynamically.

cs.CR↗