Dongjoon Lee

Hyeryn Park

Hyuntae Kim

Chohee Kim

Rhoan Lee

Hanbyul Kang

Tongmin Kim

Mingyu Choi

Seongyeon Kim

The Actionable Intelligence Lab at Korea University, under the direction of Prof. Changhee Lee, focuses on the development of advanced machine learning and artificial intelligence theories and methodologies. The lab aims to improve decision-making processes through self- and semi-supervised learning, causal inference, and automated machine learning, addressing real-world problems in healthcare, manufacturing, and other critical domains.

고려대학교 Actionable Intelligence Lab은 Changhee Lee 교수의 지도하에 운영되며, 첨단 기계 학습 및 인공 지능 이론과 방법론 개발에 중점을 둔다. 자율 및 준지도 학습, 인과 추론, 자동화된 기계 학습을 통해 의료, 제조 등 주요 분야에서 실질적인 문제 해결과 더 나은 의사 결정을 목표로 한다.

TimeSeg: An Information-Theoretic Segment-Wise Explainer for Time-Series Predictions

Stochastic Encodings for Active Feature Acquisition

Toward a Better Discrimination and Calibration via Survival Outcome-Aware Contrastive Learning

Enhancing Anomaly Detection via Generating Diversified and Hard-to-distinguish Synthetic Anomalies

Discovering Features with Synergistic Interactions in Multiple Views

Conditional Information Bottleneck Approach for Time Series Imputation

Risk-Averse Active Sensing for Timely Outcome Prediction under Cost Pressure

Neural Stochastic Differential Games for Time-series Analysis

T-Phenotype: Discovering Phenotypes of Predictive Temporal Patterns in Disease Progression

Self-Supervision Enhanced Feature Selection with Correlated Gates

SurvITE: Learning Heterogeneous Treatment Effects from Time-to-Event Data

A Variational Information Bottleneck Approach to Multi-Omics Data Integration

Temporal Phenotyping using Deep Predictive Clustering of Disease Progression

Temporal Quilting for Survival Analysis

DeepHit: A Deep Learning Approach to Survival Analysis with Competing Risks

Deep Language Model-based Early Recognition of Out-of-Hospital Cardiac Arrest from Real-Time Emergency Calls

FA-DeepMSM: A Few-shot Adapted Interpretable Multimodal Survival Model for Improved Prognostic Prediction in Glioblastoma

Development of a Long-term Survival Prediction Model for Patients Undergoing Invasive Coronary Angiography using Ensemble-based Machine Learning in Time-to-event Analysis

Risk Prediction of Major Adverse Cardiovascular Events Within One Year After Percutaneous Coronary Intervention in Patients with Acute Coronary Syndrome: Machine Learning Based Time-to-Event Analysis

Enhancing Multi-Level Cross-Modal Interaction with False Negative-Aware Contrastive Learning for Text-Video Retrieval

Predicting Discharge Destination and Length of Stay after Open Reduction Internal Fixation of Distal Femur Fractures

A Pediatric Emergency Prediction Model using Natural Language Process in the Pediatric Emergency Department

A Novel Machine Learning Model to Predict Revision ACL Reconstruction Failure in the MARS Cohort

BioBridge: Unified Bio-Embedding with Bridging Modality in Code-Switched EMR

Predicting Sleep Based on Physical Activity, Light Exposure, and Heart Rate Variability Data Using Wearable Devices

DSG-KD: Knowledge Distillation from Domain-Specific to General Language Models

Feature Selection with Group-Sparse Stochastic Gates

Development of a Predictive Model for Survival over Time in Patients with Out-of-Hospital Cardiac Arrest using Ensemble-based Machine Learning

Mobility Management Paradigm Shift: from Reactive to Proactive Handover using AI/ML

A Predictive Algorithm for Perioperative Complications and Readmission after Ankle Arthrodesis

Development and External Validation of a Risk Calculator for Prediction of Major Complications and Readmission after Anterior Cervical Discectomy and Fusion

Development and Clinical Utility of Machine Learning Algorithms for Dynamic Longitudinal Real-Time Estimation of Progression Risks in Active Surveillance of Early Prostate Cancer

Machine Learning-Driven Identification of Novel Patient Factors for Prediction of Major Complications after Posterior Cervical Spinal Fusion

A Risk Calculator for Prediction of C5 Nerve Root Palsy after Instrumented Cervical Fusion

Development of a Machine Learning Algorithm for Prediction of Complications and Unplanned Readmission Following Primary Anatomic Total Shoulder Replacements

Prediction Of Good Sleep With Physical Activity and Light Exposure: A Preliminary Study

Development of a Machine Learning Algorithm for Prediction of Complications and Unplanned Readmission following Reverse Total Shoulder Arthroplasty

Outcome-Oriented Deep Temporal Phenotyping of Disease Progression

A Novel, Potentially Universal Machine-Learning Algorithm to Predict Complications in Total Knee Arthroplasty

Prediction of Major Complications and Readmission after Lumbar Spinal Fusion: A Machine Learning-Driven Approach

Development of a Novel, Potentially Universal Machine Learning Algorithm for Prediction of Complications After Total Hip Arthroplasty

Application of a Novel Machine Learning Framework for Predicting Cancer-Specific Mortality: Analysis of 171,942 Men with Non-Metastatic Prostate Cancer from the Surveillance, Epidemiology, and End Results Dataset

Dynamic-DeepHit: A Deep Learning Approach for Dynamic Survival Analysis with Competing Risks based on Longitudinal Data

Deep Survival Analysis from Whole Slide Images: A Multiple Instance Learning Approach

Time-series Anomaly Detection based on Imputation in the Latent Space

A Comprehensive Evaluation of Time-Series Forecasting Methods for Proactive Handover

Clinical Time Series Imputation using Conditional Information Bottleneck

Survey: Strategies for Loss-Based Discrete-Time Hazard and Survival Function Estimation

Estimation of Individual Treatment Effect in Latent Confounder Models via Adversarial Learning

Efficient Interference Control Technology for Vehicular Moving Network (VMN)

Joint User Scheduling and Adaptive Intercell Interference Cancelation for MISO Downlink Cellular Systems

Resource Allocation for Device-to-Device Communications based on Graph-Coloring

Interference Avoidance Resource Allocation for D2D Communication Based on Graph-Coloring

An Efficient Signaling Procedure for Stationary Machine Type Devices

Dynamic RACH Preamble Allocation Scheme

A Joint Adaptive Beamforming and User Scheduling Algorithm for Downlink Network MIMO Systems

Adaptive Beamforming Selection Methods for Inter-cell Interference Cancellation in Multicell Multiuser Systems

Antenna Placement Designs for Distributed Antenna Systems with Multiple-Antenna Ports

“Pseudo washout” sign in high-flow hepatic hemangioma on gadoxetic acid contrast-enhanced MRI mimicking hypervascular tumor

Prediction of liver cirrhosis, using diagnostic imaging tools

Comparison of the methods for tumor response assessment in patients with hepatocellular carcinoma undergoing transarterial chemoembolization

Increased selenoprotein p levels in subjects with visceral obesity and nonalcoholic Fatty liver disease

Doppler US and CT diagnosis of nutcracker syndrome

Using MRI to evaluate anatomic significance of aortic bifurcation, right renal artery, and conus medullaris when locating lumbar vertebral segments

Identification and prediction of transitional vertebrae on imaging studies: anatomical significance of paraspinal structures

Typical and atypical imaging findings of intrahepatic cholangiocarcinoma using gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid–enhanced magnetic resonance imaging

Radiologic-pathologic correlation of hepatobiliary phase hypointense nodules without arterial phase hyperenhancement at gadoxetic acid–enhanced MRI: a multicenter study

Chemotherapy-induced focal hepatopathy in patients with gastrointestinal malignancy: gadoxetic acid–enhanced and diffusion-weighted MR imaging with clinical-pathologic correlation

Using intravoxel incoherent motion (IVIM) MR imaging to predict lipiodol uptake in patients with hepatocellular carcinoma following transcatheter arterial chemoembolization: a …

Malignant glomus tumor of the stomach with multiorgan metastases: report of a case

Paradoxical high signal intensity of hepatocellular carcinoma in the hepatobiliary phase of Gd-EOB-DTPA enhanced MRI: initial experience

Usefulness of standard deviation on the histogram of ultrasound as a quantitative value for hepatic parenchymal echo texture; preliminary study

Usefulness of controlled aliasing in parallel imaging results in higher acceleration in gadoxetic acid–enhanced liver magnetic resonance imaging to clarify the hepatic arterial …

Cutoff points of abdominal obesity indices in screening for non‐alcoholic fatty liver disease in Asians

Transmural migration of surgical sponge evacuated by defecation: mimicking an intraperitoneal gossypiboma

Comparison of non-contrast abbreviated MRI and ultrasound as surveillance modalities for HCC

Multiparametric MR index for the diagnosis of non-alcoholic steatohepatitis in patients with non-alcoholic fatty liver disease

Paradoxical uptake of Gd-EOB-DTPA on the hepatobiliary phase in the evaluation of hepatic metastasis from breast cancer: is the “target sign” a common finding?

Changhee Lee

OBJECTIVE. The purpose of this article is to describe the “pseudo washout” sign of high-flow hepatic hemangioma that mimics hypervascular tumor on gadoxetic acid–enhanced MRI. CONCLUSION. High-flow hemangiomas might show relatively low signal intensity because of gadoxetic acid contrast uptake in the surrounding normal liver parenchyma during the equilibrium (3-minute delay) phase. Such findings are called pseudo washout and can mimic hypervascular hepatic tumors. However, high-flow hemangioma can be diagnosed by observing bright signal intensity on T2-weighted imaging, arterial phase–dominant enhancement, pseudo washout sign during the equilibrium phase, and isointense or slightly increased signal intensity on subtraction images.

Early diagnosis of liver cirrhosis is important. Ultrasound-guided liver biopsy is the gold standard for diagnosis of liver cirrhosis. However, its invasiveness and sampling bias limit the applicability of the method. Basic imaging for the diagnosis of liver cirrhosis has developed over the last few decades, enabling early detection of morphological changes of the liver by ultrasonography (US), computed tomography, and magnetic resonance imaging (MRI). They are also accurate diagnostic methods for advanced liver cirrhosis, for which early diagnosis is difficult. There are a number of ways to compensate for this difficulty, including texture analysis to more closely identify the homogeneity of hepatic parenchyma, elastography to measure the stiffness and elasticity of the liver, and perfusion studies to determine the blood flow volume, transit time, and velocity. Amongst these methods, elastography using US and MRI was …

BACKGROUND & AIMS Recently, new methods, including the concept of viable enhancing tumor such as EASL and mRECIST, have been proposed for substitution of the conventional WHO and RECIST criteria in hepatocellular carcinoma (HCC) undergoing transarterial chemoembolization (TACE). Herein, we evaluated the differences of four methods and compared the association of these methods with the prognosis of HCC patients undergoing TACE. METHODS We retrospectively reviewed 114 consecutive newly diagnosed HCC patients who underwent TACE as initial treatment. We evaluated the intermethod agreement (κ values) between the methods and compared their association with the prognosis of HCC patients. RESULTS The κ values for EASL vs. WHO, EASL vs. RECIST, mRECIST vs. WHO, and mRECIST vs. RECIST were low, of 0.102, 0.088, 0.112, and 0.122, respectively. However, good …

Background: Selenoprotein P (SeP) has recently been reported as a novel hepatokine that regulates insulin resistance and systemic energy metabolism in rodents and humans. We explored the associations among SeP, visceral obesity, and nonalcoholic fatty liver disease (NAFLD). Methods: We examined serum SeP concentrations in subjects with increased visceral fat area (VFA) or liver fat accumulation measured with computed tomography. Our study subjects included 120 nondiabetic individuals selected from participants of the Korean Sarcopenic Obesity Study. In addition, we evaluated the relationship between SeP and cardiometabolic risk factors, including homeostasis model of insulin resistance (HOMA-IR), high sensitivity C-reactive protein (hsCRP), adiponectin values, and brachial-ankle pulse wave velocity (baPWV). Results: Subjects with NAFLD showed increased levels of HOMA-IR, hsCRP, VFA, and several components of metabolic syndrome and decreased levels of adiponectin and high density lipoprotein cholesterol than those of controls. Serum SeP levels were positively correlated with VFA, hsCRP, and baPWV and negatively correlated with the liver attenuation index. Not only subjects with visceral obesity but also those with NAFLD exhibited significantly increased SeP levels (P< 0.001). In multiple logistic regression analysis, the subjects in the highest SeP tertile showed a higher risk for NAFLD than those in the lowest SeP tertile, even after adjusting for potential confounding factors (odds ratio, 7.48; 95% confidence interval, 1.72 to 32.60; P= 0.007). Conclusion: Circulating SeP levels were increased in subjects with NAFLD …

Objective To compare the diagnostic accuracy of dynamic computed tomography (CT) and gadoxetate-enhanced magnetic resonance imaging (MRI) for characterization of hepatic lesions by using the Liver Imaging Reporting and Data System (LI-RADS) in a multicenter, off-site evaluation. Materials and Methods In this retrospective multicenter study, we evaluated 231 hepatic lesions (114 hepatocellular carcinomas [HCCs], 58 non-HCC malignancies, and 59 benign lesions) confirmed histologically in 217 patients with chronic liver disease who underwent both gadoxetate-enhanced MRI and dynamic CT at one of five tertiary hospitals. Four radiologists at different institutes independently reviewed all MR images first and the CT images 4 weeks later. They evaluated the major and ancillary imaging features and categorized each hepatic lesion according to the LI-RADS v2014. Diagnostic performance was calculated and compared using generalized estimating equations. Results MRI showed higher sensitivity and accuracy than CT for diagnosing hepatic malignancies; the pooled sensitivities, specificities, and accuracies for categorizing LR-5/5V/M were 59.0% vs. 72.4% (CT vs. MRI; p < 0.001), 83.5% vs. 83.9% ( p = 0.906), and 65.3% vs. 75.3% ( p < 0.001), respectively. CT and MRI showed comparable capabilities for differentiating between HCC and other malignancies, with pooled accuracies of 79.9% and 82.4% for categorizing LR-M, respectively ( p = 0.139). Conclusion Gadoxetate-enhanced MRI showed superior accuracy for categorizing LR-5/5V/M in hepatic malignancies in comparison with dynamic CT. Both modalities had …

OBJECTIVE. This study evaluated the anatomic significance and sites of aortic bifurcation, right renal artery, and conus medullaris when locating lumbar vertebral segments on MRI. SUBJECTS AND METHODS. The study group was composed of 210 patients who underwent MRI of the lumbosacral spine. The position of aortic bifurcation was evaluated using MRI. Midline and sagittal MRI were also assessed to identify the proximal portion of the right renal artery and the tip of the conus medullaris. These locations were recorded in relation to the upper or lower half of the adjacent vertebral body or the adjacent intervertebral disk space. RESULTS. The most common site of aortic bifurcation was at the L4 vertebra (83%). In most patients two separated iliac arteries were apparent at the level of the L4–L5 disk space (93%), and the right renal artery was located between the lower half of the L1 vertebra and the upper half …

The aim of our study was to examine the locational distribution of paraspinal structures on MRI and to determine any predictable parameters that may be used for the identification of transitional vertebra (TV). We enrolled 534 patients who underwent MRI of their lumbosacral spine. The locations of the paraspinal structures, such as aortic bifurcation (AB), IVC confluence (IC), right renal artery (RRA), celiac trunk (CT), SMA root (SR), and iliolumbar ligament (ILL), were determined using “cross link” in PACS. We also assessed the morphology of the TV. The MRI showed that the most common site of the paraspinal structures in the normal group was AB at the lower L4, IC at the L4‐5 disc space, RRA at the L1‐2 disc space, CT at the T12‐L1 disc space, SR at the upper L1, and ILL at the L5. The frequency of TV was 23.8% (lumbarization, 9.9%; sacralization, 13.9%). The paraspinal structures of the S1 lumbarization were …

Purpose The objective of this study was to examine the imaging features of classic mass-forming intrahepatic cholangiocarcinoma (MICC) and nonclassic hypervascular MICC on gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid–enhanced magnetic resonance imaging. Methods Twenty pathologically confirmed MICCs were included. Two radiologists retrospectively reviewed the imaging characteristics on T2-weighted imaging, diffusion-weighted imaging, dynamic contrast-enhanced images, and hepatobiliary phase (HBP) of each MICC. For the morphologic feature of defect, HBP signal intensity (SI) ratio was calculated by dividing the SI of the MICC by nearby normal liver parenchyma SI. Results Classic MICCs (n= 14) showed classic rim or peripheral enhancement at arterial dominant phase with centripetal enhance in the delayed phases. Hypervascular MICCs (n= 6) showed complete (n= 4) or near …

Background Hepatobiliary phase (HBP) hypointense nodules without arterial phase hyperenhancement (APHE) at gadoxetic acid–enhanced MRI may indicate hepatocellular carcinoma (HCC) or nonmalignant cirrhosis-associated nodules. Purpose To assess the distribution of pathologic diagnoses of HBP hypointense nodules without APHE at gadoxetic acid–enhanced MRI and to evaluate clinical and imaging features in differentiating their histologic grades. Materials and Methods This retrospective multicenter study included pathologic analysis–confirmed HBP hypointense nodules without APHE (≤30 mm) in patients with chronic liver disease or cirrhosis …

Purpose To retrospectively evaluate findings of chemotherapy-induced focal hepatopathy (CIFH) on gadoxetic acid–enhanced magnetic resonance (MR) and diffusion-weighted (DW) images and to determine imaging features that are most helpful in differentiating CIFH from metastasis. Materials and Methods This retrospective study was approved by the institutional review board, and informed consent was waived. MR images, including DW images and gadoxetic acid–enhanced images, from 12 patients (four men, eight women; age range, 25–64 years) with 15 CIFHs were reviewed independently and in consensus by two radiologists and were compared with those obtained in 20 control patients (12 men, eight women; age range, 32–84 years) with 30 hepatic metastasis who were matched for tumor size, primary organ, and chemotherapy regimen. Interobserver agreement was assessed with κ statistics, and …

Purpose To assess the usefulness of intravoxel incoherent motion diffusion weighted imaging (IVIM-DWI) for predicting lipiodol uptake in patients with hepatocellular carcinoma (HCC) after transcatheter arterial chemoembolization (TACE). Materials and methods The institutional review board approved this study. 44 HCC patients underwent IVIM-DWI and Gd-EOB-DTPA-enhanced MRI prior to TACE. Using post-TACE CT as a reference standard, each HCC was classified into either lipiodol good uptake (LGU) or poor uptake (LPU) group. Apparent diffusion coefficient ( ADC ), true diffusion coefficient ( D ), perfusion coefficient ( D* ), and perfusion fraction ( f ) in HCC were calculated. Arterial enhancement ratio (AER) and IVIM parameters were compared between those two groups using the Mann-Whitney U test. Results Of the 51 HCCs, 37 (72.5%) were LGU group and 14 (27.5%) were LPU group. AER of HCC was …

This report presents the case of a 65-year-old female patient with a malignant glomus tumor of the stomach, with metastases to the kidney and brain. The clinical presentation and imaging results are discussed, and the literature on malignant glomus tumors is reviewed.

PURPOSE To describe the paradoxical high signal intensity of hepatocellular carcinoma (HCC) in the hepatobiliary phase on gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)-enhanced magnetic resonance imaging (MRI). MATERIALS AND METHODS A database search was performed to identify cases of HCC that showed unusual prolonged enhancement in the hepatobiliary phase of Gd-EOB-DTPA MRI. All patients received 3.0-T liver MRI including precontrast T1-weighted images, T2-weighted images and a post Gd-EOB-DTPA-enhanced dynamic study. The signal intensity of HCC was measured at pre-enhanced, arterial, portal, delayed and hepatobiliary phase using regions of interest. Radiologic and pathologic correlation was performed for the paradoxically prolonged enhancing portion of HCC in the hepatobiliary phase. RESULTS Four patients (all male, age range 44–70 …

This study examined whether the standard deviation (SD) values on the histogram measured in the picture archiving and communication system (PACS) views of an ultrasound (US) can be used as an index to determine the homogeneity or heterogeneity of a hepatic parenchymal echo. Patients (n = 202) underwent an abdominal US examination. One US unit was used with a convex (3 to 5 MHz) transducer. The echogenicity of the hepatic parenchyma was measured in four different regions-of-interest (ROIs) of the liver with all vessels, bile ducts and calcification being avoided. The SDs were calculated automatically on the histogram of the ROI in the PACS views. The echo patterns of the liver were classified into normal, fatty liver (FL) and chronic liver disease (CLD). The distribution of the SD calculated from the ROI histogram of these three groups was examined, and statistical analysis for multiple comparisons of …

Purpose We aimed to determine whether the controlled aliasing in parallel imaging results in higher acceleration (CAIPIRINHA) technique could improve the image quality of the hepatic arterial phase of gadoxetic acid–enhanced liver magnetic resonance (MR) imaging. Materials and Methods A total of 320 patients underwent gadoxetic acid–enhanced liver MR imaging: a conventional protocol (a fixed scan delay and 2-mL/s injection) using a standard 3-T MR system (Trio-Tim; Siemens, commercialized since 2005)(group A), an optimized protocol (bolus tracking and 1-mL/s injection) using a standard 3-T MR (group B), an optimized protocol using a new 3-T MR (Skyra; Siemens, commercialized since 2012)(group C), and an optimized protocol with CAIPIRINHA using a new 3-T MR (group D). The image quality of the hepatic arterial phase was graded using a 4-point rating scale from 1 (no artifacts) to 4 points (non …

Background/aims: Abdominal obesity is associated with metabolic syndrome and non‐alcoholic fatty liver disease (NAFLD). Although there have been many studies to determine the optimal cutoff points of waist circumference or visceral fat area in screening for metabolic syndrome, there have been no reports to establish adequate cutoff points of abdominal obesity indices in screening for NAFLD. Therefore, we examined the appropriate cutoff points of abdominal obesity indices associated with NAFLD in Korean men and women using receiver operating characteristic (ROC) curve analysis. Furthermore, we compared the usefulness of various abdominal obesity indices measured using computed tomography (CT), dual‐energy X‐ray absorptiometry (DXA) and anthropometric parameters for detecting NAFLD. Methods: We analysed the baseline data of an ongoing prospective, observational cohort study, including …

The spontaneous defecation of the surpical retained sponge is very rare. Here, we report a case of migrating surgical sponge that was retained in the colon and it was evacuated by defecation.

Background & Aims Ultrasound (US) is recommended for HCC surveillance in high-risk patients but has limited performance in detecting early-stage HCC. We aimed to compare the diagnostic performance of biannual US and annual non-contrast abbreviated magnetic resonance imaging (NC-AMRI) as HCC surveillance modalities in high-risk patients. Methods This prospective, multicenter cohort study enrolled participants with an estimated annual risk of HCC greater than 5% between October 2015 and April 2017. Participants underwent six rounds of HCC surveillance at 6-month intervals, with both US and NC-AMRI at rounds 1, 3, and 5, and only US at rounds 2, 4, and 6. The sensitivity, diagnostic yield (DY), and false referral rate (FRR) for HCC detection by US and NC-AMRI were compared. Results In total, 208 participants underwent 980 US and 516 NC-AMRI examinations during 30 months of follow-up …

Non-alcoholic steatohepatitis (NASH) is a complex disease consisting of various components including steatosis, lobular inflammation, and ballooning degeneration, with or without fibrosis. Therefore, it is difficult to diagnose NASH with only one imaging modality. This study was aimed to evaluate the feasibility of magnetic resonance imaging (MRI) for predicting NASH and to develop a non-invasive multiparametric MR index for the detection of NASH in non-alcoholic fatty liver disease (NAFLD) patients. This prospective study included 47 NAFLD patients who were scheduled to undergo or underwent ultrasound-guided liver biopsy within 2 months. Biopsy specimens were graded as NASH or non-NASH. All patients underwent non-enhanced MRI including MR spectroscopy (MRS), MR elastography (MRE), and T1 mapping. Diagnostic performances of MRS, MRE, and T1 mapping for grading steatosis, activity, and …

PURPOSE The purpose was to describe magnetic resonance imaging (MRI) findings of breast cancer liver metastasis using gadoxetic acid (Gd-EOB-DTPA) with an emphasis on the added value of the hepatobiliary phase (HBP). MATERIAL AND METHODS Nine patients with 13 liver metastases were included in the study after the medical records of 29 breast cancer patients who underwent Gd-EOB-DTPA-enhanced MRI between February 2008 and June 2010 were reviewed. The diagnoses of liver metastasis were established by percutaneous liver biopsy or surgery and on the basis of image findings. Two radiologists retrospectively evaluated signal intensity (SI) and sizes of metastases and patterns of enhancement in an HBP. The SI ratio was calculated as the SI of the central hyperintense portion in “target” lesions divided by the SI of nearby normal liver parenchyma on the HBP. We also measured apparent …

The AI-LAB

교수님 소개

이창희

연구실 소개

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

TimeSeg: An Information-Theoretic Segment-Wise Explainer for Time-Series Predictions

Deep Language Model-based Early Recognition of Out-of-Hospital Cardiac Arrest from Real-Time Emergency Calls

FA-DeepMSM: A Few-shot Adapted Interpretable Multimodal Survival Model for Improved Prognostic Prediction in Glioblastoma

연구실 보유 특허

대학별 입시끈

우리 연구실 모집

교수님 소개

이창희

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

TimeSeg: An Information-Theoretic Segment-Wise Explainer for Time-Series Predictions

Deep Language Model-based Early Recognition of Out-of-Hospital Cardiac Arrest from Real-Time Emergency Calls

FA-DeepMSM: A Few-shot Adapted Interpretable Multimodal Survival Model for Improved Prognostic Prediction in Glioblastoma

연구실 보유 특허

최근 3년간 연구실적 높은 연구실

대학별 입시끈

우리 연구실 모집