Minseok Kim

Minseok Kim

Applied Scientist II at Amazon AGI

Amazon AGI

Biography

I am an applied scientist II at Amazon AGI. I work on overcoming imbalanced and scarce real-world data for large language model based conversational AI. Throughout my research, I have studied various kinds of topics to help AI deal with real-world problems. Specifically, I am interested in advancing AI to better serve people, such as in recommender systems or AI assistants that became more than essential in everyday life.

Interests

  • Large Language Model based Conversational AI
  • Recommendation System
  • Information Retrieval
  • Trustworthy ML
  • Real-World ML/AI Challenges

Education

  • PhD in Graduate School of Data Science

    Korea Advanced Institute of Science and Technology (KAIST)

Recent Publications

Quickly discover relevant content by filtering publications.

Aligning Large Language Models via Fine-grained Supervision (ACL 2024)

Pre-trained large-scale language models (LLMs) excel at producing coherent articles, yet their outputs may be untruthful, toxic, or …
Aligning Large Language Models via Fine-grained Supervision (ACL 2024)

Toward Robustness in Multi-label Classification: A Data Augmentation Strategy against Imbalance and Noise (AAAI 2024)

Multi-label classification poses challenges due to imbalanced and noisy labels in training data. We propose a unified data augmentation …
PDF
Toward Robustness in Multi-label Classification: A Data Augmentation Strategy against Imbalance and Noise (AAAI 2024)

Debiasing Neighbor Aggregation for Graph Neural Network in Recommender Systems (CIKM 2022)

Graph neural networks (GNNs) have achieved remarkable success in recommender systems by representing users and items based on their …
PDF DOI
Debiasing Neighbor Aggregation for Graph Neural Network in Recommender Systems (CIKM 2022)

Learning from Noisy Labels with Deep Neural Networks: A Survey (TNNLS 2022)

Deep learning has achieved remarkable success in numerous domains with help from large amounts of big data. However, the quality of …
PDF
Learning from Noisy Labels with Deep Neural Networks: A Survey (TNNLS 2022)

Meta-Learning for Online Update of Recommender Systems (AAAI 2022)

Online recommender systems should be always aligned with users' current interest to accurately suggest items that each user would like. …
PDF
Meta-Learning for Online Update of Recommender Systems (AAAI 2022)

COVID-EENet: Predicting Fine-Grained Impact of COVID-19 on Local Economies (AAAI 2022)

Assessing the impact of the COVID-19 crisis on economies is fundamental to tailor the responses of the governments to recover from the …
Code
COVID-EENet: Predicting Fine-Grained Impact of COVID-19 on Local Economies (AAAI 2022)

Task-Agnostic Undesirable Feature Deactivation Using Out-of-Distribution Data (NeurIPS 2021)

A deep neural network (DNN) has achieved great success in many machine learning tasks by virtue of its high expressive power. However, …
PDF Code
Task-Agnostic Undesirable Feature Deactivation Using Out-of-Distribution Data (NeurIPS 2021)

Robust Learning by Self-Transition for Handling Noisy Labels (KDD 2021)

Real-world data inevitably contains noisy labels, which induce the poor generalization of deep neural networks. It is known that the …
PDF
Robust Learning by Self-Transition for Handling Noisy Labels (KDD 2021)

PREMERE: Meta-Reweighting via Self-Ensembling for Point-of-Interest Recommendation (AAAI 2021)

Point-of-interest (POI) recommendation has become an important research topic in these days. The user check-in history used as the …
PDF
PREMERE: Meta-Reweighting via Self-Ensembling for Point-of-Interest Recommendation (AAAI 2021)

Carpe Diem, Seize the Samples Uncertain "At the Moment" for Adaptive Batch Selection (CIKM 2020)

The performance of deep neural networks is significantly affected by how well mini-batches are constructed. In this paper, we propose a …
PDF
Carpe Diem, Seize the Samples Uncertain "At the Moment" for Adaptive Batch Selection (CIKM 2020)

Ada-Boundary: Accelerating DNN Training via Adaptive Boundary Batch Selection (Machine Learning 2020, (SCI Expanded, impact factor: 2.672)

Neural networks converge faster with help from a smart batch selection strategy. In this regard, we propose Ada-Boundary, a novel and …
Video Source Document DOI
Ada-Boundary: Accelerating DNN Training via Adaptive Boundary Batch Selection (Machine Learning 2020, (SCI Expanded, impact factor: 2.672)

Hi-COVIDNet: Deep Learning Approach to Predict Inbound COVID-19 Patients and Case Study in South Korea (KDD 2020)

The escalating crisis of COVID-19 has put people all over the world in danger. Owing to the high contagion rate of the virus, COVID-19 …
PDF Code DOI
Hi-COVIDNet: Deep Learning Approach to Predict Inbound COVID-19 Patients and Case Study in South Korea (KDD 2020)

How Does Early Stopping Help Generalization against Label Noise? (ICMLW 2020)

Noisy labels are very common in real-world training data, which lead to poor generalization on test data because of overfitting to the …
PDF Code
How Does Early Stopping Help Generalization against Label Noise? (ICMLW 2020)

TRAP: Two-level Regularized Autoencoder-based Embedding for Power-law Distributed Data (TheWebConf 2020)

Finding low-dimensional embeddings of sparse high-dimensional data objects is important in many applications such as recommendation, …
PDF
TRAP: Two-level Regularized Autoencoder-based Embedding for Power-law Distributed Data (TheWebConf 2020)

SELFIE: Refurbishing Unclean Samples for Robust Deep Learning (ICML 2019)

Owing to the extremely high expressive power of deep neural networks, their side effect is to totally memorize training data even when …
PDF Code Dataset
SELFIE: Refurbishing Unclean Samples for Robust Deep Learning (ICML 2019)

News

News for TAUFE (NeurIPS 2021)

Our 2021 NeurIPS paper was reported in a newspaper.
Dec 6, 2021

News for Recency Bias (CIKM 2020)

Our CIKM paper is reported in aitimes.
Oct 20, 2020

News for Hi-COVIDNet (KDD 2020)

Our KDD paper on the prediction of inbound COVID-19 confirmed cases received considerable attention from Government (Ministry of Science and ICT) and media coverage in KBS, YTN, MBC, and TJB as well as in over 40 daily newspapers .
Aug 20, 2020

Lectures & Mentoring

AI Computer Vision

KAIST Dept. of Mathematical Sciences Jul 2022 – Jul 2022
Lecturer

Special AI Academy Program AI-based Project Mentor

Seocho-gu office, Korea Jul 2022 – Sep 2022
Project Mentor

Deep Learning

KAIST IT Academy Jul 2022 – Jul 2022
Lecturer

Special AI Academy Program AI-based Project Mentor

Seocho-gu office, Korea Aug 2021 – Sep 2021
Project Mentor

Deep Learning

KAIST IT Academy Jul 2021 – Aug 2021
Lecturer

Special AI Academy Program Lecturer

Seocho-gu office, Korea Jun 2021 – Aug 2021
Lecturer

AI Lecture Materials Development

KAIST IT Academy Oct 2020 – Dec 2020
Team Leader

Deep Learning

DSME Sep 2020 – Dec 2020
Course Mentor

Machine Learning

DSME Sep 2020 – Dec 2020
Course Mentor

Analytical Methodologies for Big Data

KAIST Sep 2020 – Dec 2020
KSE526 2020 Fall lecture TA

Deep Learning

KAIST IT Academy Dec 2019 – Jan 2020
Lecturer

AI Program

Hankook Tire Oct 2019
Project Mentor

Special AI Academy Program (Stage 4: AI-based project)

Seocho-gu office, Korea Sep 2019 – Dec 2019
Project Mentor

Special AI Academy Program (Stage 3: Deep Learning)

Seocho-gu office, Korea Aug 2019
Lecturer

Special AI Academy Program (Stage 2: Data Structure & Algorithm)

Seocho-gu office, Korea Jul 2019
Lecturer

Data Preprocessing & Visualization

KAIST IT Academy Jul 2019 – Aug 2019
Lecturer

Special AI Academy Program (Stage 1: Python programming)

Seocho-gu office, Korea Jun 2019
Lecturer

Advanced AI Project

LG Group Mar 2019 – Jun 2019
AI project mentor

Data Mining and Knowledge Discovery

KAIST Mar 2019 – Jun 2019
KSE525 2019 Spring lecture TA

Analytical Methodologies for Big Data

KAIST Sep 2018 – Dec 2018
KSE526 2018 Fall lecture TA

Big Data for FinTech

Korea Bank Financial Group Jul 2017
Project mentor

Contact

  • minseokkim0630@gmail.com
  • 10450 NE 10th St, Bellevue, Washington 98004
  • (SEA107) Tahoe