arXiv subjects

Dongmin Kim

Publications and source records attributed to Dongmin Kim.

27 records · Page 2Linked to original sources

arXiv2024

When Model Meets New Normals: Test-time Adaptation for Unsupervised Time-series Anomaly Detection

Time-series anomaly detection deals with the problem of detecting anomalous timesteps by learning normality from the sequence of observations. However, the concept of normality evolves over time, leading to a "new normal problem", where the distribution of normality can be changed due to the distribution shifts between training and test data. This paper highlights the prevalence of the new normal problem in unsupervised time-series anomaly detection studies. To tackle this issue, we propose a simple yet effective test-time adaptation strategy based on trend estimation and a self-supervised approach to learning new normalities during inference. Extensive experiments on real-world benchmarks demonstrate that incorporating the proposed strategy into the anomaly detector consistently improves the model's performance compared to the baselines, leading to robustness to the distribution shifts.

cs.LG↗

arXiv2023

Deep Imbalanced Time-series Forecasting via Local Discrepancy Density

Time-series forecasting models often encounter abrupt changes in a given period of time which generally occur due to unexpected or unknown events. Despite their scarce occurrences in the training set, abrupt changes incur loss that significantly contributes to the total loss. Therefore, they act as noisy training samples and prevent the model from learning generalizable patterns, namely the normal states. Based on our findings, we propose a reweighting framework that down-weights the losses incurred by abrupt changes and up-weights those by normal states. For the reweighting framework, we first define a measurement termed Local Discrepancy (LD) which measures the degree of abruptness of a change in a given period of time. Since a training set is mostly composed of normal states, we then consider how frequently the temporal changes appear in the training set based on LD. Our reweighting framework is applicable to existing time-series forecasting models regardless of the architectures. Through extensive experiments on 12 time-series forecasting models over eight datasets with various in-output sequence lengths, we demonstrate that applying our reweighting framework reduces MSE by 10.1% on average and by up to 18.6% in the state-of-the-art model.

cs.LG↗

arXiv2023

WaveBound: Dynamic Error Bounds for Stable Time Series Forecasting

Time series forecasting has become a critical task due to its high practicality in real-world applications such as traffic, energy consumption, economics and finance, and disease analysis. Recent deep-learning-based approaches have shown remarkable success in time series forecasting. Nonetheless, due to the dynamics of time series data, deep networks still suffer from unstable training and overfitting. Inconsistent patterns appearing in real-world data lead the model to be biased to a particular pattern, thus limiting the generalization. In this work, we introduce the dynamic error bounds on training loss to address the overfitting issue in time series forecasting. Consequently, we propose a regularization method called WaveBound which estimates the adequate error bounds of training loss for each time step and feature at each iteration. By allowing the model to focus less on unpredictable data, WaveBound stabilizes the training process, thus significantly improving generalization. With the extensive experiments, we show that WaveBound consistently improves upon the existing models in large margins, including the state-of-the-art model.

cs.LG↗

arXiv2022

Residual Correction in Real-Time Traffic Forecasting

Predicting traffic conditions is tremendously challenging since every road is highly dependent on each other, both spatially and temporally. Recently, to capture this spatial and temporal dependency, specially designed architectures such as graph convolutional networks and temporal convolutional networks have been introduced. While there has been remarkable progress in traffic forecasting, we found that deep-learning-based traffic forecasting models still fail in certain patterns, mainly in event situations (e.g., rapid speed drops). Although it is commonly accepted that these failures are due to unpredictable noise, we found that these failures can be corrected by considering previous failures. Specifically, we observe autocorrelated errors in these failures, which indicates that some predictable information remains. In this study, to capture the correlation of errors, we introduce ResCAL, a residual estimation module for traffic forecasting, as a widely applicable add-on module to existing traffic forecasting models. Our ResCAL calibrates the prediction of the existing models in real time by estimating future errors using previous errors and graph signals. Extensive experiments on METR-LA and PEMS-BAY demonstrate that our ResCAL can correctly capture the correlation of errors and correct the failures of various traffic forecasting models in event situations.

cs.LG↗

arXiv2020

Large-scale Hybrid Approach for Predicting User Satisfaction with Conversational Agents

Measuring user satisfaction level is a challenging task, and a critical component in developing large-scale conversational agent systems serving the needs of real users. An widely used approach to tackle this is to collect human annotation data and use them for evaluation or modeling. Human annotation based approaches are easier to control, but hard to scale. A novel alternative approach is to collect user's direct feedback via a feedback elicitation system embedded to the conversational agent system, and use the collected user feedback to train a machine-learned model for generalization. User feedback is the best proxy for user satisfaction, but is not available for some ineligible intents and certain situations. Thus, these two types of approaches are complementary to each other. In this work, we tackle the user satisfaction assessment problem with a hybrid approach that fuses explicit user feedback, user satisfaction predictions inferred by two machine-learned models, one trained on user feedback data and the other human annotation data. The hybrid approach is based on a waterfall policy, and the experimental results with Amazon Alexa's large-scale datasets show significant improvements in inferring user satisfaction. A detailed hybrid architecture, an in-depth analysis on user feedback data, and an algorithm that generates data sets to properly simulate the live traffic are presented in this paper.

cs.HC↗

arXiv2019

Axion Dark Matter Research with IBS/CAPP

The axion, a consequence of the PQ mechanism, has been considered as the most elegant solution to the strong-CP problem and a compelling candidate for cold dark matter. The Center for Axion and Precision Physics Research (CAPP) of the Institute for Basic Science (IBS) was established on 16 October 2013 with a main objective to launch state of the art axion experiments in South Korea. Relying on the haloscope technique, our strategy is to run several experiments in parallel to explore a wide range of axion masses with sensitivities better than the QCD axion models. We utilize not only the advanced technologies, such as high-field large-volume superconducting (SC) magnets, ultra low temperature dilution refrigerators, and nearly quantum-limited noise amplifiers, but also some unique features solely developed at the Center, including high-quality SC resonant cavities surviving high magnetic fields and efficient cavity geometries to reach high-frequency regions. Our goal is to probe axion dark matter in the frequency range of 1-10 GHz in the first phase and then ultimately up to 25 GHz, even in a scenario where axions constitute only 10% of the local dark matter halo. In this report, the current status and future prospects of the experiments and R&D activities at IBS/CAPP are described.

physics.ins-det↗

arXiv2019

Saliency difference based objective evaluation method for a superimposed screen of the HUD with various background

The head-up display (HUD) is an emerging device which can project information on a transparent screen. The HUD has been used in airplanes and vehicles, and it is usually placed in front of the operator's view. In the case of the vehicle, the driver can see not only various information on the HUD but also the backgrounds (driving environment) through the HUD. However, the projected information on the HUD may interfere with the colors in the background because the HUD is transparent. For example, a red message on the HUD will be less noticeable when there is an overlap between it and the red brake light from the front vehicle. As the first step to solve this issue, how to evaluate the mutual interference between the information on the HUD and backgrounds is important. Therefore, this paper proposes a method to evaluate the mutual interference based on saliency. It can be evaluated by comparing the HUD part cut from a saliency map of a measured image with the HUD image.

cs.HC↗

arXiv2018

Hybrid Two-Qubit Gate using Circuit QED System with Triple-Leg Stripline Resonator

We theoretically propose a circuit QED system implemented with triple-leg stripline resonator (TSR). Unlikely from linear stripline resonator, the fundamental intra-cavity microwave modes of the TSR are two-fold degenerate. When a superconducting qubit is placed near one of the TSR legs, one fundamental mode is directly coupled to the qubit, while the other one remains uncoupled. Our system closely resembles an optical cavity QED system, where an atom in a cavity couples only to the incident photon with a specific polarization by placing a polarization beamsplitter in front of the optical cavity. Using our circuit QED system, we have theoretically studied a two-qubit quantum gate operation in a hybrid qubit composed of flying microwave qubit and superconducting qubit. We have demonstrated that for the hybrid qubit, the quantum controlled phase flip (CPF) gate can be reliably implemented for the experimentally available set of parameters.

quant-ph↗

arXiv2011

Sparse Inverse Covariance Estimation via an Adaptive Gradient-Based Method

We study the problem of estimating from data, a sparse approximation to the inverse covariance matrix. Estimating a sparsity constrained inverse covariance matrix is a key component in Gaussian graphical model learning, but one that is numerically very challenging. We address this challenge by developing a new adaptive gradient-based method that carefully combines gradient information with an adaptive step-scaling strategy, which results in a scalable, highly competitive method. Our algorithm, like its predecessors, maximizes an $\ell_1$-norm penalized log-likelihood and has the same per iteration arithmetic complexity as the best methods in its class. Our experiments reveal that our approach outperforms state-of-the-art competitors, often significantly so, for large problems.

stat.ML↗