Jisu Kim

Kwangho Kim, Jisu Kim

In many modern machine learning pipelines, abundant pretrained representations serve as noisy proxy covariates, while task-specific labels remain scarce. We study semi-supervised regression in this setting, and propose a simple two stage estimator that learns kernel eigenfeatures from all proxy covariates and fits a ridge predictor on labeled data. We derive finite sample bounds showing that fast labeled sample rates are recovered when proxy perturbation is controlled and unlabeled proxy covariates are sufficiently abundant. We also show that distribution regression is a direct special case, with analogous guarantees when the finite bag size is large enough. Experiments show consistent gains over supervised and semi-supervised baselines, especially in low label regimes.

Pum Jun Kim, Yoojin Jang, Jisu Kim et al.

We propose a robust and reliable evaluation metric for generative models by introducing topological and statistical treatments for rigorous support estimation. Existing metrics, such as Inception Score (IS), Frechet Inception Distance (FID), and the variants of Precision and Recall (P&R), heavily rely on supports that are estimated from sample features. However, the reliability of their estimation has not been seriously discussed (and overlooked) even though the quality of the evaluation entirely depends on it. In this paper, we propose Topological Precision and Recall (TopP&R, pronounced 'topper'), which provides a systematic approach to estimating supports, retaining only topologically and statistically important features with a certain level of confidence. This not only makes TopP&R strong for noisy features, but also provides statistical consistency. Our theoretical and experimental results show that TopP&R is robust to outliers and non-independent and identically distributed (Non-IID) perturbations, while accurately capturing the true trend of change in samples. To the best of our knowledge, this is the first evaluation metric focused on the robust estimation of the support and provides its statistical consistency under noise.

Hengrui Luo, Jisu Kim, Alice Patania et al.

Topology can extract the structural information in a dataset efficiently. In this paper, we attempt to incorporate topological information into a multiple output Gaussian process model for transfer learning purposes. To achieve this goal, we extend the framework of circular coordinates into a novel framework of mixed valued coordinates to take linear trends in the time series into consideration. One of the major challenges to learn from multiple time series effectively via a multiple output Gaussian process model is constructing a functional kernel. We propose to use topologically induced clustering to construct a cluster based kernel in a multiple output Gaussian process model. This kernel not only incorporates the topological structural information, but also allows us to put forward a unified framework using topological information in time and motion series.

Omkar Ranadive, Jisu Kim, Serin Lee et al.

Wildfires are a disastrous phenomenon which cause damage to land, loss of property, air pollution, and even loss of human life. Due to the warmer and drier conditions created by climate change, more severe and uncontrollable wildfires are expected to occur in the coming years. This could lead to a global wildfire crisis and have dire consequences on our planet. Hence, it has become imperative to use technology to help prevent the spread of wildfires. One way to prevent the spread of wildfires before they become too large is to perform early detection i.e, detecting the smoke before the actual fire starts. In this paper, we present our Wildfire Detection and Alert System which use machine learning to detect wildfire smoke with a high degree of accuracy and can send immediate alerts to users. Our technology is currently being used in the USA to monitor data coming in from hundreds of cameras daily. We show that our system has a high true detection rate and a low false detection rate. Our performance evaluation study also shows that on an average our system detects wildfire smoke faster than an actual person.

Kwangho Kim, Jisu Kim, Edward H. Kennedy

Causal effects are often characterized with population summaries. These might provide an incomplete picture when there are heterogeneous treatment effects across subgroups. Since the subgroup structure is typically unknown, it is more challenging to identify and evaluate subgroup effects than population effects. We propose a new solution to this problem: \emph{Causal k-Means Clustering}, which harnesses the widely-used k-means clustering algorithm to uncover the unknown subgroup structure. Our problem differs significantly from the conventional clustering setup since the variables to be clustered are unknown counterfactual functions. We present a plug-in estimator which is simple and readily implementable using off-the-shelf algorithms, and study its rate of convergence. We also develop a new bias-corrected estimator based on nonparametric efficiency theory and double machine learning, and show that this estimator achieves fast root-n rates and asymptotic normality in large nonparametric models. Our proposed methods are especially useful for modern outcome-wide studies with multiple treatment levels. Further, our framework is extensible to clustering with generic pseudo-outcomes, such as partially observed outcomes or otherwise unknown functions. Finally, we explore finite sample properties via simulation, and illustrate the proposed methods using a study of mobile-supported self-management for chronic low back pain.

JooYoung Jang, Youngseo Cha, Jisu Kim et al.

Recently, both the frequency and intensity of wildfires have increased worldwide, primarily due to climate change. In this paper, we propose a novel protocol for wildfire detection, leveraging semi-supervised Domain Adaptation for object detection, accompanied by a corresponding dataset designed for use by both academics and industries. Our dataset encompasses 30 times more diverse labeled scenes for the current largest benchmark wildfire dataset, HPWREN, and introduces a new labeling policy for wildfire detection. Inspired by CoordConv, we propose a robust baseline, Location-Aware Object Detection for Semi-Supervised Domain Adaptation (LADA), utilizing a teacher-student based framework capable of extracting translational variance features characteristic of wildfires. With only using 1% target domain labeled data, our framework significantly outperforms our source-only baseline by a notable margin of 3.8% in mean Average Precision on the HPWREN wildfire dataset. Our dataset is available at https://github.com/BloomBerry/LADA.

Jisu Kim, Juhwan Lee

The rapid advancement of Large Language Models (LLMs) has improved text understanding and generation but poses challenges in computational resources. This study proposes a curriculum learning-inspired, data-centric training strategy that begins with simpler tasks and progresses to more complex ones, using criteria such as prompt length, attention scores, and loss values to structure the training data. Experiments with Mistral-7B (Jiang et al., 2023) and Gemma-7B (Team et al., 2024) models demonstrate that curriculum learning slightly improves performance compared to traditional random data shuffling. Notably, we observed that sorting data based on our proposed attention criteria generally led to better performance. This approach offers a sustainable method to enhance LLM performance without increasing model size or dataset volume, addressing scalability challenges in LLM training.

Kwangho Kim, Jisu Kim, Larry A. Wasserman et al.

Understanding treatment effect heterogeneity is vital for scientific and policy research. However, identifying and evaluating heterogeneous treatment effects pose significant challenges due to the typically unknown subgroup structure. Recently, a novel approach, causal k-means clustering, has emerged to assess heterogeneity of treatment effect by applying the k-means algorithm to unknown counterfactual regression functions. In this paper, we expand upon this framework by integrating hierarchical and density-based clustering algorithms. We propose plug-in estimators that are simple and readily implementable using off-the-shelf algorithms. Unlike k-means clustering, which requires the margin condition, our proposed estimators do not rely on strong structural assumptions on the outcome process. We go on to study their rate of convergence, and show that under the minimal regularity conditions, the additional cost of causal clustering is essentially the estimation error of the outcome regression functions. Our findings significantly extend the capabilities of the causal clustering framework, thereby contributing to the progression of methodologies for identifying homogeneous subgroups in treatment response, consequently facilitating more nuanced and targeted interventions. The proposed methods also open up new avenues for clustering with generic pseudo-outcomes. We explore finite sample properties via simulation, and illustrate the proposed methods in voting and employment projection datasets.

Jisu Kim, Youngwoo Shin, Uiji Hwang et al.

Idioms have long posed a challenge due to their unique linguistic properties, which set them apart from other common expressions. While recent studies have leveraged large language models (LLMs) to handle idioms across various tasks, e.g., idiom-containing sentence generation and idiomatic machine translation, little is known about the underlying mechanisms of idiom processing in LLMs, particularly in multilingual settings. To this end, we introduce MIDAS, a new large-scale dataset of idioms in six languages, each paired with its corresponding meaning. Leveraging this resource, we conduct a comprehensive evaluation of LLMs' idiom processing ability, identifying key factors that influence their performance. Our findings suggest that LLMs rely not only on memorization, but also adopt a hybrid approach that integrates contextual cues and reasoning, especially when processing compositional idioms. This implies that idiom understanding in LLMs emerges from an interplay between internal knowledge retrieval and reasoning-based inference.

Keonvin Park, Jisu Kim, Jaemin Seo

This paper introduces PINT (Physics-Informed Neural Time Series Models), a framework that integrates physical constraints into neural time series models to improve their ability to capture complex dynamics. We apply PINT to the ERA5 WeatherBench dataset, focusing on long-term forecasting of 2m-temperature data. PINT incorporates the Simple Harmonic Oscillator Equation as a physics-informed prior, embedding its periodic dynamics into RNN, LSTM, and GRU architectures. This equation's analytical solutions (sine and cosine functions) facilitate rigorous evaluation of the benefits of incorporating physics-informed constraints. By benchmarking against a linear regression baseline derived from its exact solutions, we quantify the impact of embedding physical principles in data-driven models. Unlike traditional time series models that rely on future observations, PINT is designed for practical forecasting. Using only the first 90 days of observed data, it iteratively predicts the next two years, addressing challenges posed by limited real-time updates. Experiments on the WeatherBench dataset demonstrate PINT's ability to generalize, capture periodic trends, and align with physical principles. This study highlights the potential of physics-informed neural models in bridging machine learning and interpretable climate applications. Our models and datasets are publicly available on GitHub: https://github.com/KV-Park.

Juhwan Lee, Jisu Kim

This study addresses the hallucination problem in large language models (LLMs). We adopted Retrieval-Augmented Generation(RAG) (Lewis et al., 2020), a technique that involves embedding relevant information in the prompt to obtain accurate answers. However, RAG also faced inherent issues in retrieving correct information. To address this, we employed the Dense Passage Retrieval(DPR) (Karpukhin et al., 2020) model for fetching domain-specific documents related to user queries. Despite this, the DPR model still lacked accuracy in document retrieval. We enhanced the DPR model by incorporating control tokens, achieving significantly superior performance over the standard DPR model, with a 13% improvement in Top-1 accuracy and a 4% improvement in Top-20 accuracy.

Jisu Kim, Jinhee Park, Changhyun Jeon et al.

Traditional query expansion techniques for addressing vocabulary mismatch problems in information retrieval are context-sensitive and may lead to performance degradation. As an alternative, document expansion research has gained attention, but existing methods such as Doc2Query have limitations including excessive preprocessing costs, increased index size, and reliability issues with generated content. To mitigate these problems and seek more structured and efficient alternatives, this study proposes a method that divides documents into chunk units and generates textual data for each chunk to simultaneously improve retrieval efficiency and accuracy. The proposed "Chunk Knowledge Generation Model" adopts a T5-based multi-task learning structure that simultaneously generates titles and candidate questions from each document chunk while extracting keywords from user queries. This approach maximizes computational efficiency by generating and extracting three types of semantic information in parallel through a single encoding and two decoding processes. The generated data is utilized as additional information in the retrieval system. GPT-based evaluation on 305 query-document pairs showed that retrieval using the proposed model achieved 95.41% accuracy at Top@10, demonstrating superior performance compared to document chunk-level retrieval. This study contributes by proposing an approach that simultaneously generates titles and candidate questions from document chunks for application in retrieval pipelines, and provides empirical evidence applicable to large-scale information retrieval systems by demonstrating improved retrieval accuracy through qualitative evaluation.

Changhyun Jeon, Jinhee Park, Jungwoo Choi et al.

We propose a small-scale language model (SLM) based agent architecture, Planner-Caller-Generator (P-C-G), optimized for Korean tool use. P-C-G separates planning, calling, and generation by role: the Planner produces an initial batch plan with limited on-demand replanning; the Caller returns a normalized call object after joint schema-value validation; and the Generator integrates tool outputs to produce the final answer. We apply a Korean-first value policy to reduce execution failures caused by frequent Korean-to-English code switching in Korean settings. Evaluation assumes Korean queries and Korean tool/parameter specifications; it covers single-chain, multi-chain, missing-parameters, and missing-functions scenarios, and is conducted via an LLM-as-a-Judge protocol averaged over five runs under a unified I/O interface. Results show that P-C-G delivers competitive tool-use accuracy and end-to-end quality while reducing tokens and maintaining acceptable latency, indicating that role-specialized SLMs are a cost-effective alternative for Korean tool-use agents.

Jisu Kim, Alex Mattingly, Eung-Joo Lee et al.

Head detection and tracking are essential for downstream tasks, but current methods often require large computational budgets, which increase latencies and ties up resources (e.g., processors, memory, and bandwidth). To address this, we propose a framework to enhance tiny head detection and tracking by optimizing the balance between performance and efficiency. Our framework integrates (1) a cross-domain detection loss, (2) a multi-scale module, and (3) a small receptive field detection mechanism. These innovations enhance detection by bridging the gap between large and small detectors, capturing high-frequency details at multiple scales during training, and using filters with small receptive fields to detect tiny heads. Evaluations on the CroHD and CrowdHuman datasets show improved Multiple Object Tracking Accuracy (MOTA) and mean Average Precision (mAP), demonstrating the effectiveness of our approach in crowded scenes.

Yoonjoo Lee, Kihoon Son, Tae Soo Kim et al.

As Large Language Models (LLMs) are nondeterministic, the same input can generate different outputs, some of which may be incorrect or hallucinated. If run again, the LLM may correct itself and produce the correct answer. Unfortunately, most LLM-powered systems resort to single results which, correct or not, users accept. Having the LLM produce multiple outputs may help identify disagreements or alternatives. However, it is not obvious how the user will interpret conflicts or inconsistencies. To this end, we investigate how users perceive the AI model and comprehend the generated information when they receive multiple, potentially inconsistent, outputs. Through a preliminary study, we identified five types of output inconsistencies. Based on these categories, we conducted a study (N=252) in which participants were given one or more LLM-generated passages to an information-seeking question. We found that inconsistency within multiple LLM-generated outputs lowered the participants' perceived AI capacity, while also increasing their comprehension of the given information. Specifically, we observed that this positive effect of inconsistencies was most significant for participants who read two passages, compared to those who read three. Based on these findings, we present design implications that, instead of regarding LLM output inconsistencies as a drawback, we can reveal the potential inconsistencies to transparently indicate the limitations of these models and promote critical LLM usage.

Eunhye Lee, Jeongmu Kim, Jisu Kim et al.

Recent image inpainting methods have shown promising results due to the power of deep learning, which can explore external information available from the large training dataset. However, many state-of-the-art inpainting networks are still limited in exploiting internal information available in the given input image at test time. To mitigate this problem, we present a novel and efficient self-supervised fine-tuning algorithm that can adapt the parameters of fully pre-trained inpainting networks without using ground-truth target images. We update the parameters of the pre-trained state-of-the-art inpainting networks by utilizing existing self-similar patches (i.e., self-exemplars) within the given input image without changing the network architecture and improve the inpainting quality by a large margin. Qualitative and quantitative experimental results demonstrate the superiority of the proposed algorithm, and we achieve state-of-the-art inpainting results on publicly available benchmark datasets.

Jisu Kim, Alina Sîrbu, Giulio Rossetti et al.

The cultural integration of immigrants conditions their overall socio-economic integration as well as natives' attitudes towards globalisation in general and immigration in particular. At the same time, excessive integration -- or acculturation -- can be detrimental in that it implies forfeiting one's ties to the home country and eventually translates into a loss of diversity (from the viewpoint of host countries) and of global connections (from the viewpoint of both host and home countries). Cultural integration can be described using two dimensions: the preservation of links to the home country and culture, which we call home attachment, and the creation of new links together with the adoption of cultural traits from the new residence country, which we call destination attachment. In this paper we introduce a means to quantify these two aspects based on Twitter data. We build home and destination attachment indexes and analyse their possible determinants (e.g., language proximity, distance between countries), also in relation to Hofstede's cultural dimension scores. The results stress the importance of host language proficiency to explain destination attachment, but also the link between language and home attachment. In particular, the common language between home and destination countries corresponds to increased home attachment, as does low proficiency in the host language. Common geographical borders also seem to increase both home and destination attachment. Regarding cultural dimensions, larger differences among home and destination country in terms of Individualism, Masculinity and Uncertainty appear to correspond to larger destination attachment and lower home attachment.

Eunhye Lee, Jeongmu Kim, Jisu Kim et al.

Recent image inpainting methods show promising results due to the power of deep learning, which can explore external information available from a large training dataset. However, many state-of-the-art inpainting networks are still limited in exploiting internal information available in the given input image at test time. To mitigate this problem, we present a novel and efficient self-supervised fine-tuning algorithm that can adapt the parameters of fully pre-trained inpainting networks without using ground-truth target images. We update the parameters of the pre-trained state-of-the-art inpainting networks by utilizing existing self-similar patches within the given input image without changing network architecture and improve the inpainting quality by a large margin. Qualitative and quantitative experimental results demonstrate the superiority of the proposed algorithm, and we achieve state-of-the-art inpainting results on publicly available numerous benchmark datasets.

Hengrui Luo, Alice Patania, Jisu Kim et al.

Topological Data Analysis (TDA) provides novel approaches that allow us to analyze the geometrical shapes and topological structures of a dataset. As one important application, TDA can be used for data visualization and dimension reduction. We follow the framework of circular coordinate representation, which allows us to perform dimension reduction and visualization for high-dimensional datasets on a torus using persistent cohomology. In this paper, we propose a method to adapt the circular coordinate framework to take into account the roughness of circular coordinates in change-point and high-dimensional applications. We use a generalized penalty function instead of an $L_{2}$ penalty in the traditional circular coordinate algorithm. We provide simulation experiments and real data analysis to support our claim that circular coordinates with generalized penalty will detect the change in high-dimensional datasets under different sampling schemes while preserving the topological structures.

Shanxin Yuan, Radu Timofte, Ales Leonardis et al.

This paper reviews the Challenge on Image Demoireing that was part of the New Trends in Image Restoration and Enhancement (NTIRE) workshop, held in conjunction with CVPR 2020. Demoireing is a difficult task of removing moire patterns from an image to reveal an underlying clean image. The challenge was divided into two tracks. Track 1 targeted the single image demoireing problem, which seeks to remove moire patterns from a single image. Track 2 focused on the burst demoireing problem, where a set of degraded moire images of the same scene were provided as input, with the goal of producing a single demoired image as output. The methods were ranked in terms of their fidelity, measured using the peak signal-to-noise ratio (PSNR) between the ground truth clean images and the restored images produced by the participants' methods. The tracks had 142 and 99 registered participants, respectively, with a total of 14 and 6 submissions in the final testing stage. The entries span the current state-of-the-art in image and burst image demoireing problems.

Kwangho Kim, Jisu Kim, Manzil Zaheer et al.

We propose PLLay, a novel topological layer for general deep learning models based on persistence landscapes, in which we can efficiently exploit the underlying topological features of the input data structure. In this work, we show differentiability with respect to layer inputs, for a general persistent homology with arbitrary filtration. Thus, our proposed layer can be placed anywhere in the network and feed critical information on the topological features of input data into subsequent layers to improve the learnability of the networks toward a given task. A task-optimal structure of PLLay is learned during training via backpropagation, without requiring any input featurization or data preprocessing. We provide a novel adaptation for the DTM function-based filtration, and show that the proposed layer is robust against noise and outliers through a stability analysis. We demonstrate the effectiveness of our approach by classification experiments on various datasets.

Jisu Kim, Young J. Kim

In this paper, we propose a novel penetration metric, called deformable penetration depth PDd, to define a measure of inter-penetration between two linearly deforming tetrahedra using the object norm. First of all, we show that a distance metric for a tetrahedron deforming between two configurations can be found in closed form based on object norm. Then, we show that the PDd between an intersecting pair of static and deforming tetrahedra can be found by solving a quadratic programming (QP) problem in terms of the distance metric with non-penetration constraints. We also show that the PDd between two, intersected, deforming tetrahedra can be found by solving a similar QP problem under some assumption on penetrating directions, and it can be also accelerated by an order of magnitude using pre-calculated penetration direction. We have implemented our algorithm on a standard PC platform using an off-the-shelf QP optimizer, and experimentally show that both the static/deformable and deformable/deformable tetrahedra cases can be solvable in from a few to tens of milliseconds. Finally, we demonstrate that our penetration metric is three-times smaller (or tighter) than the classical, rigid penetration depth metric in our experiments.

Milad Kiaee, Adam B Kashlak, Jisu Kim et al.

Obstructive sleep apnea is a serious condition causing a litany of health problems especially in the pediatric population. However, this chronic condition can be treated if diagnosis is possible. The gold standard for diagnosis is an overnight sleep study, which is often unobtainable by many potentially suffering from this condition. Hence, we attempt to develop a fast non-invasive diagnostic tool by training a classifier on 2D and 3D facial images of a patient to recognize facial features associated with obstructive sleep apnea. In this comparative study, we consider both persistent homology and geometric shape analysis from the field of computational topology as well as convolutional neural networks, a powerful method from deep learning whose success in image and specifically facial recognition has already been demonstrated by computer scientists.

Kwangho Kim, Jisu Kim, Alessandro Rinaldo

We develop a novel algorithm for feature extraction in time series data by leveraging tools from topological data analysis. Our algorithm provides a simple, efficient way to successfully harness topological features of the attractor of the underlying dynamical system for an observed time series. The proposed methodology relies on the persistent landscapes and silhouette of the Rips complex obtained after a de-noising step based on principal components applied to a time-delayed embedding of a noisy, discrete time series sample. We analyze the stability properties of the proposed approach and show that the resulting TDA-based features are robust to sampling noise. Experiments on synthetic and real-world data demonstrate the effectiveness of our approach. We expect our method to provide new insights on feature extraction from granular, noisy time series data.

Jisu Kim, Hyungbo Shim, Jin Heon Seo

We study the design problems of state observers and tracking controllers for a class of hybrid systems whose state jumps. The idea is to utilize the well-known method of gluing the jump set (a part of domain where the jumps take place) onto its image, which converts the hybrid system into a continuous-time system whose state does not jump. Sufficient conditions for this idea to be implemented are listed and discussed with a few concrete examples. In particular, we present a structural condition for an observer design, and, for tracking control, we introduce a feedback to compensate residual discontinuity in the vector field after gluing. The benefits of the proposed approach include that the observer design does not require detection of the state jumps, and that the tracking control does not require the plant state jumps when the reference jumps.

Kwangho Kim, Jisu Kim, Edward H. Kennedy

Comparing counterfactual distributions can provide more nuanced and valuable measures for causal effects, going beyond typical summary statistics such as averages. In this work, we consider characterizing causal effects via distributional distances, focusing on two kinds of target parameters. The first is the counterfactual outcome density. We propose a doubly robust-style estimator for the counterfactual density and study its rates of convergence and limiting distributions. We analyze asymptotic upper bounds on the $L_q$ and the integrated $L_q$ risks of the proposed estimator, and propose a bootstrap-based confidence band. The second is a novel distributional causal effect defined by the $L_1$ distance between different counterfactual distributions. We study three approaches for estimating the proposed distributional effect: smoothing the counterfactual density, smoothing the $L_1$ distance, and imposing a margin condition. For each approach, we analyze asymptotic properties and error bounds of the proposed estimator, and discuss potential advantages and disadvantages. We go on to present a bootstrap approach for obtaining confidence intervals, and propose a test of no distributional effect. We conclude with a numerical illustration and a real-world example.

Jisu Kim, Yen-Chi Chen, Sivaraman Balakrishnan et al.

A cluster tree provides a highly-interpretable summary of a density function by representing the hierarchy of its high-density clusters. It is estimated using the empirical tree, which is the cluster tree constructed from a density estimator. This paper addresses the basic question of quantifying our uncertainty by assessing the statistical significance of topological features of an empirical cluster tree. We first study a variety of metrics that can be used to compare different trees, analyze their properties and assess their suitability for inference. We then propose methods to construct and summarize confidence sets for the unknown true cluster tree. We introduce a partial ordering on cluster trees which we use to prune some of the statistically insignificant features of the empirical tree, yielding interpretable and parsimonious cluster trees. Finally, we illustrate the proposed methods on a variety of synthetic examples and furthermore demonstrate their utility in the analysis of a Graft-versus-Host Disease (GvHD) data set.