Faculty

Jaesik Choi

Director of Explainable Artificial Intelligence Center

Professor of Graduate School of AI at KAIST
Graduate School of Artificial Intelligence, KAIST
PhD from University of Illinois at Urbana-Champaign, 2012
BS from Seoul National University, 2004
KAIST Endowed Chair Professor (KAIST 지정석좌교수)
Associate Member of the National Academy of Engineering of Korea (공학한림원 일반회원)
Director of Explainable Artificial Intelligence Center
CEO of INEEJI (The Government certified National Strategic Technology)
Seongnam: 8 Seongnam-daero 331, Bundang, Seongnam, Gyeonggi, 13558, Republic of Korea (main office)
Daejeon: 291 Daehak-ro, N24 , Yuseong-gu, Daejeon 34141, Republic of Korea
Secretary: (031) 8022-7530

Documents

Brief bio

Jaesik Choi is a KAIST Endowed Chair Professor from March 2025 and a full professor of Graduate School of Artificial Intelligence at KAIST since September 2024. He is also a director of Explainable Artificial Intelligent Center established by the Ministry of Science and ICT, Republic of Korea. He had been an associate professor of the Graduate School of Artificial Intelligence at KAIST since September 2019. He had been the Rising-Star Distinguished associate professor in the School of Electrical and Computer Engineering at UNIST (Ulsan, Korea) and an affiliate researcher of the Lawrence Berkeley National Laboratory (the Berkeley Lab) until August 2019. He had been an assistant professor at UNIST from July 2013 to August 2017. He was a Computer Scientist Postdoctoral Fellow of Computational Research Division at the Berkeley Lab. His research focuses on learning and inference with large scale, complex systems, explainable machine learning models and spatio-temporal data analysis. He received his Ph.D. in Computer Science from the University of Illinois at Urbana-Champaign (2012) and received B.S. degree in Computer Engineering from Seoul National University (2004).

Research interest

I am interested in Artificial Intelligence and Machine Learning. Specifically, I am interested in Efficient Inference with Large-Scale Probabilistic Graphical Models, Intelligent Hybrid (Cyber-Physical) Systems, Robot Task Planning and Multimedia Retrieval.

Thesis

Jaesik Choi, Lifted Inference for Relational Hybrid Models, Ph.D. Thesis, University of Illinois at Urbana-Champaign, Department of Computer Science, 2012 (pdf).

Talks

강연관련정보 (약력, 초록 및 발표자료 등)

News

I co-organized an ICML-2026 Workshop Philosophy Meets Machine Learning: What Counts As Trustworthy? (July 2026).
I was appointed as a Member of the National AI Strategy Committee of Korea and the leader of Manufacturing AI Task Force (November 2025).
I co-organized an ACM CIKM-2026 Workshop Human-Centric AI: From Explainability and Trustworthiness to Actionable Ethics (November 2025).
A paper, XAI—Explainable artificial intelligence, that I coauthored receives the 2025 Frontiers of Science Award by the International Congress of Basic Science (August 2025).The ministry of science and technology of Korea selected INEEJI as the first AI company to hold a national strategic technology, the predictive explainable artificial intelligence technology. (March 2025).
I am appointed as the KAIST Endowed Chair Professor (March 2025).
Dr. Haedong Jeong (my second PhD student) is appointed at Sogang University as an assistant professor (March 2025).
I become an associate member of Korean National Academy of Engineering (January 2025).
I am promoted to the full professor of Kim Jaechul Graduate School of Artificial Intelligence at KAIST (September 2024).
A paper, Explainable Artificial Intelligence (XAI) 2.0: A manifesto of open challenges and interdisciplinary research directions, co-authored with Luca Longo and many others is published at Information Fusion (June 2024) and received the best paper award (December 2025).
Dr. Anh Tong (my first PhD student) is appointed at Korea University as an assistant professor (March 2024).
INEEJI received a Series A funding, 11.1 Billion KRW (December 2023).
I co-organized and present at the IPAM Explainable AI for the Sciences: Towards Novel Insights Workshop (January 2023).
We received the appreciation plaque from Samsung Electronics with the explainable speech recognition explainable artificial intelligence (June 2022).
INEEJI received a pre-Series A funding, 5.5 Billion KRW (May 2022).
I will lead a New XAI project with a budget of 13 Billion KRW, Development of Plug and Play Explainable Artificial Intelligence Platform, which is granted by the ministry of science and ICT for the 57 months (April 2022).
My tenure application is approved by the KAIST Faculty Personnel Committee and the KAIST Board of Trustees (September 2021). Thankfully, so many people (graduate students, staff, department head and external recommendations) supported the procedure (September 2021).
A book, An Introduction to Lifted Probabilistic Inference, which I contributed two chapters are published by MIT press. (August 2021).
The Industrial Bank of Korea and KAIST XAI Research Center agrees a MOU for AI research collaboration (August 2021).
I received the POSCO O&C (Open and Collaboration) Award from POSCO (August 2021).
The first PhD student, Anh Tong, successfully defended his PhD thesis (December 2020).
Our team won (1st place) the AI Electricity Generation Competition of Korea East West Power (December 2020).
I presented a tutorial on Interpreting and Explaining Deep Neural Networks: A Perspective on Time Series Data at KDD-2020 (August 2020).
A paper, XAI - Explainable artificial intelligence, co-authored with David Gunning and other colleagues is published at Science Robotics (December 2019).
I am organizing/chairing 2019 ICCV Workshop on Interpreting and Explaining Visual Artificial Intelligence Models (November 2019).
I moved to KAIST and join Graduate School of Artificial Intelligence (September 2019).
The joint project with POSCO is selected as a 'Lighthouse Factory' by World Economic Forum (July 2019).
The smart blast furnace technology (deep learning based prediction and control of hot melting iro) is selected as a national core technologies of Korea (June 2019).
I have received a prime minister’s commendation (April 2019).
I found a company, INEEJI (January 2019).
I served a co-program chair for 2018 International Explainable Artificial Intelligence Symposium (November 2018).
I have been named as one of the Rising Star Distinguished professors of UNIST (September 2018).
I have received a prestigious award, POSCO Innovation Award 2018, from POSCO (August 2018).
Our SW (joint work with Kyowoon Lee and Sol-A Kim) won two international digital Curling competitions held at Japan (2017 November and 2018 March). The research paper, joint work with Prof. Seong-Whan Lee at Korea University, is presented at ICML 2018.
I have been named as the POSCO steel fellow professor (August, 2017).
I become a director of the explainable artificial intelligence (XAI) center to lead the 5 years of project that is supported by the ministry of science and ICT (July 2017).
A fast and effective way which could achieve up to 100X of compression ratio of many real-world big data. "Novel Data Reduction Based on Statistical Similarity" with Dongeun Lee, Alex Sim and John Wu is accepted at SSDBM-16 (May 2016).
A new perspective to analyze the success (failure) probability in dynamical compressive sensing framework. "Improving Imprecise Compressive Sensing" with Dongeun Lee and Rafael de Lima is accepted at UAI-16 (May 2016).
An improved way to generate human readable reports from time series data, "The Relational Automatic Statistician System" with Yunseong Hwang (now at NAVER Search solution) and Anh Tong is accepted at ICML-16 (April 2016).
I serve on Programming Committee of AAAI 2016, KR 2016, UAI 2016, ACML 2016 and StarAI 2016, and Reviewer for NIPS 2016.
Best Paper award at DataCom 2015. A predictive model for US household electricity data with Taehoon Kim, Dongeun Lee and colleagues in the Berkeley Lab receives the best paper award at DataCom 2015 (December 2015).
A new pragmatic archiving scheme for spatio-temporal sensor data! "A Scalable and Flexible Repository for Big Sensor Data" with Dongeun Lee and Prof. Heonshik Shin (Seoul National University) is accepted at IEEE Sensors Journal (August 2015).
A report on machine learning for Legislative Knowledge Service is published in the National Assembly Library of Korea.
A large scale data reduction framework, “Data Reduction Method with Locally Exchangeable Measures” with Alex Sim (Berkeley National Lab) is now patent pending (May 2015).
A new National IT Industry Promotion Agency (NIPA) grant is acknowledged (January 2015 ~ December 2016, Role: Co-PI, 200M KRW about 183K USD for my portion). This is a collaborative effort with the City of Ulsan (PI: Ulsan Economic Promotion Agency) and Daedo Techra, a local manufacturing company.
A New Efficient Method for Estimating the Partition Function of Exponential Random Graph Models! Paper, “A Deterministic Partition Function Approximation for Exponential Random Graph Models”, with Wen Pu (Linked In), Yunseong Hwang and Eyal Amir (University of Illinois) is accepted at the International Joint Conference on Artificial Intelligence 2015 (IJCAI-15) (April 2015).
New Machine Learning Algorithm that Monitors Memory Behaviors for Secure Embedded Systems! Paper, “Memory Heat Map: Anomaly Detection in Real-Time Embedded Systems Using Memory Behavior”, with Man-Ki Yoon, Sibin Mohan, and Lui Sha is accepted at the Design Automation Conference (DAC 2015). This research is a collaborative effort with Cyber Physical System Integration Lab at University of Illinois at Urbana-Champaign (February 2015).
I serve on Programming Committee of AAAI 2015, IJCAI 2015, UAI 2015, ACML 2015 and StarAI (Statistical Relational AI) 2015 and Reviewer for NIPS 2015.
New dynamic Compressive Sensing models! Paper, “Learning Compressive Sensing Models for Big Spatio-Temporal Data” with Dongeun Lee is accepted at the 2015 SIAM International Conference on Data Mining (SDM 2015) (December 2014).
A new Korean National Research Foundation (NRF) grant is acknowledged (PI: Jaesik Choi, December 2014 ~ November 2017, 600M KRW about 557K USD). This is a collaborative effort with Dr. R. Vilim and Dr. Y-S Park at the Nuclear Division of the Argonne National Laboratory. The new research project, “Development of Fault-Diagnosis Algorithms for Current and Next-Generation Nuclear Systems” is officially initiated. Many thanks, Korean NRF and US Department of Energy (DOE), and the selection committee!
The first paper (in the literature) on learning parameters in relational continuous probabilistic models! Paper, “Learning Relational Kalman Filtering” with Eyal Amir, Tianfan Xu and Albert Valocchi is accepted at 2015 AAAI Conference (November 2014).
A face detection algorithm robust on signal noises. Paper, “Semi-Local Structure Patterns for Robust Face Detection” with Kyungjoong Jeong and Gil-Jin Jag is accepted IEEE Signal Processing Letters (November 2014).
A new fast and efficient algorithm for big spatio-temporal sensor data archiving. Paper, “Low Complexity Sensing for Big Spatio-Temporal Data“ with Dongeun Lee is accepted at 2014 IEEE Big Data Conference (September 2014).
Slides for the Machine Learning Tutorial 2014, “Statistical Inference with Graphical Models“, at Yonsei University in Seoul are now available (January 2014) (Download slides).
I serve on Programming Committee of AAAI 2014 and ICML 2014.
A fast linear-time Gaussian Process for online monitoring of the electricity market. Paper, “Fast Change Point Detection for Electricity Market Analysis” with William Gu, Ming Gu, Horst Simon, and Kesheng Wu is accepted at IEEE Big Data Workshop: Scalable Machine Learning: Theory and Applications (September 2013).
A new lifted inference algorithm for transitive relations (e.g., p(Friends(A,B) and Friends(B,C) → Friends(A,C)) = 0.3). Paper, “Lifted Inference on Transitive Relations“ with Wen Pu and Eyal Amir is accepted at AAAI-13 Workshop: Star AI(Statistical Relational AI) (July 2013).
As of July 29th, I joined the Ulsan National Institute of Science and Technology (UNIST) as an assistant professor in the School of Electrical and Computer Engineering.
Machine Learning algorithms for accurate Groundwater Models! Paper, “Use of Machine Learning Methods to Reduce Predictive Error of Groundwater Models” with Tianfang Xu, Albert Valocchi and Eyal Amir is accepted at Groundwater (March 2013).
A new way of searching videos without tags! Paper, “A Spatio-Temporal Pyramid Matching for Video Retrieval” with Ziyu Wang, Sang-Chul Lee and Won J. Jeon is accepted at CVIU (February 2013).
Statistical Inference meets Real-Time Security! Paper, “SecureCore: A Multicore based Intrusion Detection Architecture for Real-time Embedded Systems” with Man-Ki Yoon, Sibin Mohan, Jung-Eun Kim and Lui Sha is accepted at RTAS 2013 (January 2013).
In January 2013, I joined the Lawrence Berkeley Lab as a postdoctoral fellow to solve machine learning problems with big data.
I serve on Programming Committee of IJCAI 2013, AAAI 2013, and ICML 2013.

32 notes

Research

All research

2025
Granular Concept Circuits
GCC discovers fine-grained circuits tied to specific visual concepts by assessing functional dependencies and semantic alignment between neurons.
2025
Granular Concept Circuits: Toward a Fine-Grained Circuit Discovery for Concept Representations
Deep vision models have achieved remarkable classification performance by leveraging a hierarchical architecture in which human-interpretable concepts emerge through the composition of individual neurons across layers. Given the distributed nature of representations, pinpointing where specific visual concepts are encoded within a model remains a crucial yet challenging task. In this paper, we introduce an effective circuit discovery method, called Granular Concept Circuit (GCC), in which each circuit represents a concept relevant to a given query. To construct each circuit, our method iteratively assesses inter-neuron connectivity, focusing on both functional dependencies and semantic alignment. By automatically discovering multiple circuits, each capturing specific concepts within that query, our approach offers a profound, concept-wise interpretation of models and is the first to identify circuits tied to specific visual concepts at a fine-grained level. We validate the versatility and effectiveness of GCCs across various deep image classification models.
2025
Counterfactual Activation Editing
A model-agnostic approach for inference-time control of prosody and pronunciation in pretrained text-to-speech systems.
2025
Counterfactual Activation Editing for Post-hoc Prosody and Mispronunciation Correction in TTS Models
Recent advances in Text-to-Speech (TTS) have significantly improved speech naturalness, increasing the demand for precise prosody control and mispronunciation correction. Existing approaches for prosody manipulation often depend on specialized modules or additional training, limiting their capacity for post-hoc adjustments. Similarly, traditional mispronunciation correction relies on grapheme-to-phoneme dictionaries, making it less practical in low-resource settings. We introduce Counterfactual Activation Editing, a model-agnostic method that manipulates internal representations in a pre-trained TTS model to achieve post-hoc control of prosody and pronunciation. Experimental results show that our method effectively adjusts prosodic features and corrects mispronunciations while preserving synthesis quality. This opens the door to inference-time refinement of TTS outputs without retraining, bridging the gap between pre-trained TTS models and editable speech synthesis.
2025
Local Manifold Approximation and Projection
LoMAP is a training-free method that keeps guided diffusion planning samples near the feasible data manifold in long-horizon tasks.
2025
Local Manifold Approximation and Projection for Manifold-Aware Diffusion Planning
Recent advances in diffusion-based generative modeling have demonstrated significant promise in tackling long-horizon, sparse-reward tasks by leveraging offline datasets. While these approaches have achieved promising results, their reliability remains inconsistent due to the inherent stochastic risk of producing infeasible trajectories, limiting their applicability in safety-critical applications. We identify that the primary cause of these failures is inaccurate guidance during the sampling procedure, and demonstrate the existence of manifold deviation by deriving a lower bound on the guidance gap. To address this challenge, we propose Local Manifold Approximation and Projection (LoMAP), a training-free method that projects the guided sample onto a low-rank subspace approximated from offline datasets, preventing infeasible trajectory generation. We validate our approach on standard offline reinforcement learning benchmarks that involve challenging long-horizon planning. Furthermore, we show that, as a standalone module, LoMAP can be incorporated into the hierarchical diffusion planner, providing further performance enhancements.
2025
Creative Concept Catalyst
C3 selectively amplifies features during denoising to improve creativity in Stable Diffusion-based models without additional training.
2025
Enhancing Creative Generation on Stable Diffusion-based Models
Recent text-to-image generative models, particularly Stable Diffusion and its distilled variants, have achieved impressive fidelity and strong text-image alignment. However, their creative capability remains constrained, as including `creative' in prompts seldom yields the desired results. In this paper, we introduce C3 (Creative Concept Catalyst), a training-free approach designed to enhance creativity in Stable Diffusion-based models. C3 selectively amplifies features during the denoising process to foster more creative outputs. We offer practical guidelines for choosing amplification factors based on two main aspects of creativity. C3 is the first study to enhance creativity in diffusion models without extensive computational costs. We demonstrate its effectiveness across various Stable Diffusion-based models.
2025
Context-aware Speech Correction
Automatic detection and correction of abnormal encoder context vectors improves synthesized speech without extra data or model retraining.
2025
An Automated Method to Correct Artifacts in Neural Text-to-Speech Models
Recent advances in deep learning technology have enabled high-quality speech synthesis, and text-to-speech models are widely used in a variety of applications. However, even stateof-the-art models still produce artificial speech, highlighting the need to correct errors in synthesized speech. Traditional speech correction methodologies face a number of challenges, such as the inefficiency of manually specifying errors, the need to retrain models, and the need for additional data to correct synthesized speech errors. In this paper, we present a novel approach that detects and corrects contextual errors within a model to improve synthetic speech without requiring additional resources or model retraining. Specifically, we analyze the inherent limitations of neural network encoders responsible for contextualizing input sentences and propose a method that automatically identifies abnormal encoder context vectors. We also introduce a correction algorithm that enhances the quality of speech by correcting the incorrect relationships between phonemes that cause abnormal encoder context vectors. Our algorithm demonstrated a 25.86% and 4.69% reduction in alignment errors and a 2.25% and 0.63% improvement in objective metrics such as Frechet Wav2Vec ´ distance for the Tacotron2 and VITS models. In addition, our algorithm improved the comparative mean opinion scores, a subjective evaluation, by 1.34 and 0.52. The results support the feasibility of a novel approach to identify, correct, and improve defects automatically in neural speech synthesis models. The results support the feasibility of a novel approach to identify, correct, and improve defects automatically in neural speech synthesis models.
2025
Causal Interactions for Payoff Allocation
A game-theoretic reformulation makes feature interactions explicit and improves attribution when negative interactions are present.
2025
Rethinking Shapley Value for Negative Interactions in Non-convex Games
We study causal interactions for payoff allocation in cooperative game theory, including quantifying feature attribution for deep learning models. Most feature attribution methods mainly stem from the criteria of the Shapley value, which assigns fair payoffs to players based on their expected contribution in a cooperative game. However, interactions between players in the game do not explicitly appear in the original formulation of the Shapley value. In this work, we reformulate the Shapley value to clarify the role of interactions and discuss implicit assumptions from a game-theoretical perspective. Our theoretical analysis demonstrates that when negative interactions exist—common in deep learning models—the efficiency axiom can lead to the undervaluation of attributions or payoffs. We suggest a new allocation rule that decomposes contributions into interactions and aggregates positive parts for non-convex games. Furthermore, we propose an approximation algorithm to reduce the cost of interaction computation which can be applied to differentiable functions such as deep learning models. Our approach mitigates counterintuitive attribution outcomes observed in existing methods, ensuring that features critical to a model’s decision receive appropriate attribution.
2025
Neural ODE Transformers: Analyzing Internal Dynamics and Adaptive Fine-tuning
Recent advancements in large language models (LLMs) based on transformer architectures have sparked significant interest in understanding their inner workings. In this paper, we introduce a novel approach to modeling transformer architectures using highly flexible non-autonomous neural ordinary differential equations (ODEs). Our proposed model parameterizes all weights of attention and feedforward blocks through neural networks, expressing these weights as functions of a continuous layer index. Through spectral analysis of the model’s dynamics, we uncover an increase in eigenvalue magnitude that challenges the weightsharing assumption prevalent in existing theoretical studies. We also leverage the Lyapunov exponent to examine token-level sensitivity, enhancing model interpretability. Our neural ODE transformer demonstrates performance comparable to or better than vanilla transformers across various configurations and datasets, while offering flexible fine-tuning capabilities that can adapt to different architectural constraints.
2025
Non-autonomous Neural ODE Transformers
Continuous-depth parameterization and spectral analysis reveal transformer dynamics while preserving flexible fine-tuning.
2025
xPatch
A dual-stream time-series forecasting architecture combines exponential decomposition, an MLP linear stream, and a CNN nonlinear stream.
2025
xPatch: Dual-Stream Time Series Forecasting with Exponential Seasonal-Trend Decomposition
In this work, we design eXponential Patch (xPatch for short), a novel dual-stream architecture that utilizes exponential decomposition. Inspired by the classical exponential smoothing approaches, xPatch introduces the innovative seasonal-trend exponential decomposition module. Additionally, we propose a dual-flow architecture that consists of an MLP-based linear stream and a CNN-based non-linear stream. This model investigates the benefits of employing patching and channel-independence techniques within a non-transformer model. Finally, we develop a robust arctangent loss function and a sigmoid learning rate adjustment scheme, which prevent overfitting and boost forecasting performance.
2025
Diverse Rare Image Generation
Latent optimization and feature-space density estimation generate rare images with controllable rarity, diversity, and reference similarity.
2025
Diverse Rare Sample Generation with Pretrained GANs
This study proposes a novel approach for generating diverse rare samples from high-resolution image datasets with pretrained GANs. Our method employs gradient-based optimization of latent vectors within a multi-objective framework and utilizes normalizing flows for density estimation on the feature space. This enables the generation of diverse rare images, with controllable parameters for rarity, diversity, and similarity to a reference image. We demonstrate the effectiveness of our approach both qualitatively and quantitatively across various datasets and GANs without retraining or fine-tuning the pretrained GANs.
2024
Understanding Distributed Representations of Concepts in Deep Neural Networks without Supervision
Understanding intermediate representations of the concepts learned by deep learning classifiers is indispensable for interpreting general model behaviors. In this paper, we propose a novel unsupervised method for discovering distributed representations of concepts by selecting a principal subset of neurons. Our empirical findings demonstrate that instances with similar neuron activation states tend to share coherent concepts. Based on the observations, the proposed method selects principal neurons that construct an interpretable region, namely a Relaxed Decision Region (RDR), encompassing instances with coherent concepts in the feature space.
2024
Relaxed Decision Regions
Principal neuron subsets reveal coherent distributed concepts in the intermediate representations of image classifiers.
2024
Towards Dynamic Trend Filtering through Trend Point Detection with Reinforcement Learning
We are finding a temporal pooling automatically reflecting the diverse characteristics of time series data in time series classification task. Previous temporal poolings exhibit data dependency based on segmentation type, and each excels with different time series data characteristics. Selection over Multiple Temporal Poolings (SoM-TP) employs selection ensemble learning to utilize the perspective of each temporal pooling in a data driven way. The diverse perspective learning of SoM-TP addresses the drawbacks of a single pooling perspective while also incorporating the benefits of the best-performing temporal pooling.
2024
Pathwise Explanation of ReLU Neural Networks
We decompose the ReLU neural network into components, called paths, based on their locally linear attributes. We identify a piecewise linear model corresponding to each path, and use it to generate input attributions that explain the model’s decisions. Configuration of the paths allow explanations to be generated by concentrating on the important parts, and decomposed across different paths to explain each object and feature.
2024
Towards Diverse Perspective Learning with Selection over Multiple Temporal Poolings
We are finding important points automatically reflecting extreme values of time series data to improve time series forecasting performance. Previous trend filtering methods ignore extreme values as outliers, which have notable information. Dynamic Trend Filtering network (DTF-net) utilizes RL to capture important points that should be reflected in the trend, including extreme values. With the trend from DTF-net, time series forecasting performance is improved.
2024
Memorizing Documents with Guidance In Large Language Models
In AI, data plays a pivotal role in training. Large language models (LLMs) are trained with massive amounts of documents, and their parameters hold document contents. Recently, several studies identified content-specific locations in LLMs by examining the parameters. Instead of the post hoc interpretation, we propose another approach, document-wise memories, which makes document-wise locations for neural memories in training. The proposed architecture maps document representation to memory entries and filters memory selections in the forward process of LLMs.
2023
Impact of Co-occurrence on Factual Knowledge of Large Language Models
We hypothesize that relying heavily on simple co-occurrence statistics of the pre-training corpora is one of the main factors that cause factual errors. Our results reveal that LLMs are vulnerable to the co-occurrence bias, defined as preferring frequently co-occurred words over the correct answer. Consequently, LLMs struggle to recall facts whose subject and object rarely co-occur in the pre-training dataset although they are seen during finetuning.
2023
Refining Diffusion Planner for Reliable Behavior Synthesis by Automatic Detection of Infeasible Plans
We propose a method to refine unreliable plans generated by diffusion-based planners for long-horizon tasks. We introduce a "restoration gap" metric to guide and improve these plans and uses an attribution map regularizer to prevent adversarial guidance. The approach is validated on benchmarks and enhances explainability by highlighting error-prone transitions in the generated plans.
2023
Explainable AI-Based Interface System for Weather Forecasting Model
As machine learning (ML) becomes more prevalent in meteorological decision-making, user-centered explainable artificial intelligence (XAI) studies have yet to extend into this domain. This study identifies three key requirements for explaining black-box models in meteorology: model bias in different rainfall scenarios, model reasoning, and output confidence. Appropriate XAI methods are mapped to these requirements and evaluated both quantitatively and qualitatively. The results show that intuitive explanations improve decision-making utility and user trust, highlighting the need for user-centered XAI to enhance the usability of AI systems in practice.
2023
Beyond Single Path Integrated Gradients for Reliable Input Attribution via Randomized Path Sampling
In this paper, we tackle this issue by estimating the distribution of the possible attributions according to the integrating path selection. We show that such noisy attribution can be reduced by aggregating attributions from the multiple paths instead of using a single path. Using multiple input attributions obtained from randomized path, we propose a novel attribution measure using the distribution of attributions at each input features.
2023
Variational Curriculum Reinforcement Learning for Unsupervised Discovery of Skills
We introduce Variational Curriculum RL (VCRL), a method for unsupervised skill discovery in reinforcement learning by using intrinsic rewards and curriculum learning. The proposed Value Uncertainty Variational Curriculum (VUVC) accelerates skill acquisition and improves state coverage compared to uniform approaches. The approach is validated on complex tasks, showing improved sample efficiency and successful real-world robotic navigation in a zero-shot setup.
2023
Rarity Score: A New Metric to Evaluate the Uncommonness of Synthesized Images
In this work, we propose a new evaluation metric, called ‘rarity score’, to measure the individual rarity of each image synthesized by generative models. We first show empirical observation that common samples are close to each other and rare samples are far from each other in nearest-neighbor distances of feature space. We then use our metric to demonstrate that the extent to which different generative models produce rare images can be effectively compared. We also propose a method to compare rarities between datasets that share the same concept such as CelebA-HQ and FFHQ.
2023
Adaptive and Explainable Deployment of Navigation Skills via Hierarchical Deep Reinforcement Learning
We propose a hierarchical framework for robotic navigation that learns a family of diverse low-level policies and a high-level policy to deploy them adaptively. Instead of relying on a fixed reward function, the method learns multiple navigation skills with varied reward functions and selects the most suitable one in real time. The approach is validated in simulation and real-world settings, demonstrating both adaptability and explainability of the agent's behavior.
2020
An Efficient Explorative Sampling Considering the Generative Boundaries of Deep Generative Neural Networks
In this paper, we present an explorative sampling algorithm to analyze generation mechanism of DGNNs. Our method efficiently obtains samples with identical attributes from a query image in a perspective of the trained model. We define generative boundaries which determine the activation of nodes in the internal layer and probe inside the model with this information.

127 papers

Publications

Full record

2026

[115]
Unified Backbone Refinement for Diffusion Models via Internal-Latent Analysis
Haksoo Lim, Myeongjin Lee, Wonjoon Chang, and Jaesik Choi
European Conference on Computer Vision (ECCV) 2026.Generative models
arXiv
[114]
Diffusion Integrated Gradients: Controllable Path Generation for Flexible Feature Attribution
Soyeon Kim, Kyowoon Lee, and Jaesik Choi
European Conference on Computer Vision (ECCV) 2026.Explainable AI Generative models
arXiv
[113]
SPICE: Simple Polysemantic Feature Interpretation via Clustering-based Explanations
Sehyun Lee, Dahee Kwon, Damin Lee, and Jaesik Choi
European Conference on Computer Vision (ECCV) 2026.Computer vision Explainable AI
[112]
Spectral Integrated Gradients for Coarse-to-Fine Feature Attribution
Soyeon Kim, Seongwoo Lim, Kyowoon Lee, and Jaesik Choi
Knowledge Discovery and Data Mining (KDD) 2026.Explainable AI Data mining
arXiv
[111]
DistMatch: Adaptive Binning via Distribution Matching for Robust Sequential Conformal Prediction
Enver Menadjiev, Jihyeon Seong, Jisu Yeo, and Jaesik Choi
International Conference on Machine Learning (ICML), 2026.Time series
Paper Code
[110]
Breaking the Lock-in: Diversifying Text-to-Image Generation via Representation Modulation
Dahee Kwon, Ha-Eun Lee, and Jaesik Choi
International Conference on Machine Learning (ICML), 2026.Generative models
Paper
[109]
Manifold-Aligned Guided Integrated Gradients for Reliable Feature Attribution
Soyeon Kim, Seongwoo Lim, Kyowoon Lee, and Jaesik Choi
International Conference on Machine Learning (ICML), 2026.Explainable AI
Paper
[107]
Incomplete Prompt Jailbreaks in Large Language Models
Yeonjea Kim, Bumjin Park, and Jaesik Choi
Findings of the Annual Meeting of the Association for Computational Linguistics (ACL), 2026.Language models
Paper
[106]
K-MetBench: A Multi-Dimensional Benchmark for Fine-Grained Evaluation of Expert Reasoning, Locality, and Multimodality in Meteorology
Soyeon Kim, Cheongwoong Kang, Myeongjin Lee, Eun-Chul Chang, Jaedeok Lee, and Jaesik Choi
Findings of the Annual Meeting of the Association for Computational Linguistics (ACL), 2026.Language models Weather and climate
arXiv
[105]
Towards Transparent Time Series Analysis: Exploring Methods and Enhancing Interpretability
Youngjin Park, Anh Tong, Sehyun Lee, Jihyeon Seong, Qin Xie and Jaesik Choi
ACM Computing Surveys, 2026.Explainable AI Time series
Paper

2025

[105]
State-Covering Trajectory Stitching for Diffusion Planners
Kyowoon Lee and Jaesik Choi
Conference on Neural Information Processing Systems (NeurIPS), 2025.Generative models Reinforcement learning
Paper
[104]
On the Relationship between Populated Regions and Adversarial Robustness in Deep Neural Networks
Seongjin Park, Haedong Jeong, Tair Djanibekov, Giyoung Jeon, Jinseok Seol and Jaesik Choi
IEEE International Conference on Data Mining (ICDM), 2025.Data mining
arXiv
[103]
When Format Changes Meaning: Investigating Semantic Inconsistency of Large Language Models
Cheongwoong Kang, Jongeun Baek, Yeonjea Kim, and Jaesik Choi
Findings of EMNLPLanguage models
Paper
[102]
Unsupervised Concept Discovery for Deep Weather Forecast Models with High-Resolution Radar Data
Soyeon Kim, Junho Choi, Subeen Lee and Jaesik Choi
PLOS Climate, 2025.Explainable AI Weather and climate
arXiv
[101]
Amortized Baseline Selection via Rank-Revealing QR for Efficient Model Explanation
Chanwoo Lee, Youngjin Park, Hyeongeun Lee, Yeeun Yoo, Daehee Han, Junho Choi, Geonhyeong Kim, Nari Kim and Jaesik Choi
Conference on Information and Knowledge Management (CIKM), 2025.Other
Paper
[100]
Deep-DFVAR: Dynamic Factor Vector Autoregression with Deep Learning for Regional House Price Index Forecasting
Jisu Yeo, Artyom Stitsyuk and Jaesik Choi
Neurocomputing, 2025.Weather and climate
Paper
[99]
Granular Concept Circuits: Toward a Fine-Grained Circuit Discovery for Concept Representations
Dahee Kwon*, Sehyun Lee* and Jaesik Choi
International Conference on Computer Vision (ICCV), 2025.Explainable AI
Paper
[98]
Context-aware Negative Sampling for Sequential Recommendation
Jinseok Seol and Jaesik Choi
IEEE Access, 2025.Other
Paper
[97]
Example-Based Concept Analysis Framework for Deep Weather Forecast Models
Soyeon Kim, Junho Choi, Subeen Lee and Jaesik Choi
Artificial Intelligence for the Earth Systems (AIES), 2025.Explainable AI Weather and climate
Paper
[96]
Counterfactual Activation Editing for Post-hoc Prosody and Mispronunciation Correction in TTS Models
Kyowoon Lee, Artyom Stitsyuk, Gunu Jho, Inchul Hwang and Jaesik Choi
Annual Conference of the International Speech Communication Association (Interspeech), 2025.Explainable AI
Paper
[95]
Deontological Keyword Bias: The Impact of Modal Verbs on Normative Judgments of Language Models
Bumjin Park, Jinsil Lee and Jaesik Choi
Annual Meeting of the Association for Computational Linguistics (ACL), 2025.Language models
Paper
[94]
Local Manifold Approximation and Projection for Manifold-Aware Diffusion Planning
Kyowoon Lee and Jaesik Choi
International Conference on Machine Learning (ICML), 2025.Generative models Reinforcement learning
Paper
[93]
Enhancing Creative Generation on Stable Diffusion-based Models
Jiyeon Han*, Dahee Kwon*, Gayoung Lee, Junho Kim and Jaesik Choi
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025 (* contributed equally).Generative models
Paper
[92]
An Automated Method to Correct Artifacts in Neural Text-to-Speech Models
Seongyeop Jeong, June Sig Sung, Inchul Hwang and Jaesik Choi
IEEE/ACM Transactions on Audio, Speech and Language Processing (T-ASLP), 2025.Other
Paper
[91]
A Perspective on Foundation Models in Chemistry
Junyoung Choi, Gunwook Nam, Jaesik Choi and Yousung Jung
Journal of the American Chemical Society (JACS AU), 2025.Other
Paper
[90]
Rethinking Shapley Value for Negative Interactions in Non-convex Games
Wonjoon Chang, Myeongjin Lee and Jaesik Choi
International Conference on Learning Representations (ICLR), 2025.Other
Paper
[89]
Neural ODE Transformers: Analyzing Internal Dynamics and Adaptive Fine-tuning
Anh Tong, Thanh Nguyen-Tang, Dongeun Lee, Duc Nguyen, Toan Tran, David Leo Wright Hall, Cheongwoong Kang and Jaesik Choi
International Conference on Learning Representations (ICLR), 2025.Other
Paper
[88]
xPatch: Dual-Stream Time Series Forecasting with Exponential Seasonal-Trend Decomposition
Artyom Stitsyuk and Jaesik Choi
AAAI Conference on Artificial Intelligence (AAAI), 2025.Time series Weather and climate
Paper Code
[87]
Diverse Rare Sample Generation with Pretrained GANs
Subeen Lee, Jiyeon Han, Soyeon Kim and Jaesik Choi
AAAI Conference on Artificial Intelligence (AAAI), 2025.Generative models
Paper

2024

[108]
LoCo: Low-rank Compositional Rotation Fine-tuning
An Nguyen, Jaesik Choi, Anh Tong
International Joint Conference on Artificial Intelligence (IJCAI), 2024.Other
arXiv
[86]
Towards Dynamic Trend Filtering through Trend Point Detection with Reinforcement Learning
Jihyeon Seong, Sekwang Oh and Jaesik Choi
International Joint Conference on Artificial Intelligence (IJCAI), 2024.Time series Reinforcement learning
Paper
[85]
Memorizing Documents with Guidance in Large Language Models
Bumjin Park and Jaesik Choi
International Joint Conference on Artificial Intelligence (IJCAI), 2024.Language models
Paper
[84]
Explainable Artificial Intelligence (XAI) 2.0: A manifesto of open challenges and interdisciplinary research directions
Luca Longo, Mario Brcic, Federico Cabitza, Jaesik Choi, Roberto Confalonieri, Javier Del Ser, Riccardo Guidotti, Yoichi Hayashi, Francisco Herrera, Andreas Holzinger, Richard Jiang, Hassan Khosravi, Freddy Lecue, Gianclaudio Malgieri, Andrés Páez, Wojciech Samek, Johannes Schneider, Timo Speith and Simone Stumpf
Information Fusion, 2024.Explainable AI
Paper
[83]
Pathwise Explanation of ReLU Neural Networks
Seongwoo Lim, Won Jo, Joohyung Lee and Jaesik Choi
International Conference on Artificial Intelligence and Statistics (AISTATS), 2024.Other
Paper
[82]
Towards Diverse Perspective Learning with Select over Multiple Temporal Poolings
Jihyeon Seong, Jungmin Kim and Jaesik Choi
AAAI Conference on Artificial Intelligence (AAAI), 2024.Time series
Paper
[81]
Understanding Distributed Representations of Concepts in Deep Neural Networks without Supervision
Wonjoon Chang, Dahee Kwon and Jaesik Choi
AAAI Conference on Artificial Intelligence (AAAI), 2024.Explainable AI
Paper
—
Capsule Neural Networks as Noise Stabilizer for Time Series Data
김소연, 성지현, 한현경, 최재식
정보과학회 논문지(Journal of KIISE), 2024.Time series

2023

—
멀티 에이전트 강화학습에서의 게이트 기반 메시지 교환
박범진, 강청웅, 최재식
제어·로봇·시스템학회 논문지
[80]
Impact of Co-occurrence on Factual Knowledge of Large Language Models
Cheongwoong Kang and Jaesik Choi
Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP), 2023.Language models
Paper
[79]
CR-COPEC: Causal Rationale of Corporate Performance Changes to learn from Financial Reports
Yeeun Chun, Soonjae Kwon, Kyunghwan Sohn, Nakwon Sung, Junyoup Lee, Byoung Ki Seo, Kevin Compher, Seung-won Hwang and Jaesik Choi
Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP), 2023.Language models
Paper
[78]
SigFormer: Signature Transformers for Deep Hedging
Anh Tong, Thanh Nguyen-Tang, Dongeun Lee, Toan Tran and Jaesik Choi
Association for Computing Machinery International Conference on Artificial Intelligence in Finance (ACM ICAIF), 2023.Other
Paper
[77]
Refining Diffusion Planner for Reliable Behavior Synthesis by Automatic Detection of Infeasible Plans
Kyowoon Lee*, Seongun Kim* and Jaesik Choi
Conference on Neural Information Processing Systems (NeurIPS), 2023 (* contributed equally).Generative models Reinforcement learning
Paper
[75]
Explainable AI-Based Interface System for Weather Forecasting Model
Soyeon Kim, Junho Choi, Yeji Choi, Subeen Lee, Artyom Stitsyuk, Minkyoung Park, Seongyeop Jeong, Youhyun Baek and Jaesik Choi
International Conference on Human-Computer Interaction International (HCII), 2023.Explainable AI Weather and climate
Paper
[74]
Beyond Single Path Integrated Gradients for Reliable Input Attribution via Randomized Path Sampling
Giyoung Jeon, Haedong Jeong and Jaesik Choi
International Conference on Computer Vision (ICCV), 2023.Explainable AI
Paper
[73]
Variational Curriculum Reinforcement Learning for Unsupervised Discovery of Skills
Seongun Kim*, Kyowoon Lee* and Jaesik Choi
International Conference on Machine Learning (ICML), 2023 (* contributed equally).Reinforcement learning
Paper
[72]
Rarity Score : A New Metric to Evaluate the Uncommonness of Synthesized Images
Jiyeon Han, Hwanil Choi, Yunjey Choi, Junho Kim, Jung-Woo Ha and Jaesik Choi
International Conference on Learning Representations (ICLR), 2023.Other
Paper
[71]
Adaptive and Explainable Deployment of Navigation Skills via Hierarchical Deep Reinforcement Learning
Kyowoon Lee*, Seongun Kim* and Jaesik Choi
International Conference on Robotics and Automation (ICRA), 2023 (* contributed equally) (pdf).Explainable AI Reinforcement learning
Paper
—
Capsule Neural Networks as Noise Stabilizer for Time Series Data
김소연, 성지현, 한현경, 최재식
한국컴퓨터종합학술대회(KCC), 2023. (우수논문상)Time series

2022

[70]
South Korea’s Response to Surging AI Use in the US and China
Jaesik Choi
Global Asia, 17 (4), 2022.Other
Paper
[69]
Learning Fractional White Noises in Neural Stochastic Differential Equations
Anh Tong, Thanh Nguyen-Tang, Toan Tran and Jaesik Choi
Conference on Neural Information Processing Systems (NeurIPS), 2022.Other
Paper
[68]
Distilled Gradient Aggregation: Purify Features for Input Attribution in the Deep Neural Network
Giyoung Jeon*, Haedong Jeong* and Jaesik Choi
Conference on Neural Information Processing Systems (NeurIPS), 2022 (* contributed equally).Explainable AI
Paper
[67]
Can We Find Neurons that Cause Unrealistic Images in Deep Generative Networks?
Hwanil Choi, Wonjoon Chang and Jaesik Choi
International Joint Conference on Artificial Intelligence (IJCAI), 2022.Generative models
Paper
[66]
An Unsupervised Way to Understand Artifact Generating Internal Units in Generative Neural Networks
Haedong Jeong, Jiyeon Han and Jaesik Choi
AAAI Conference on Artificial Intelligence (AAAI), 2022.Generative models
Paper
[65]
Cooperative Multi-Robot Task Allocation with Reinforcement Learning
Bumjin Park, Cheongwoong Kang, Jaesik Choi
Applied Sciences, 12 (1), 2022.Reinforcement learning
Paper
[64]
Scheduling PID Attitude and Position Control Frequencies for Time-Optimal Quadrotor Waypoint Tracking under Unknown External Disturbances
Cheongwoong Kang, Bumjin Park, Jaesik Choi
Sensors, 22 (1), 2022.Other
Paper

2021

[63]
Lifted Inference for Hybrid Relational Models
Jaesik Choi, Kristian Kersting and Yuqiao Chen
An Introduction to Lifted Probabilistic Inference, MIT Press, August 2021.Probabilistic inference
Paper
[62]
Lifted Aggregation and Skolemization for Directed Models
Wannes Meert, Jaesik Choi, Jacek Kisynski, Hung Bui, Guy Van den Broeck, Adnan Darwiche, Rodrigo de Salvo Braz and David Poole
An Introduction to Lifted Probabilistic Inference, MIT Press, August 2021.Probabilistic inference
Paper
[61]
Predicting Potentially Hazardous Chemical Reactions Using Explainable Neural Network
Juhwan Kim, Geun Ho Gu, Juhwan Noh, Seongun Kim, Suji Gim, Jaesik Choi* and Yousung Jung*
Chemical Science (Chem), 2021.Explainable AI
Paper
[60]
Explaining the Decisions of Deep Policy Networks for Robotic Manipulations
Seongun Kim and Jaesik Choi
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021.Explainable AI Reinforcement learning
Paper
[59]
Semi-Supervised Training of Deep Generative Models for High-Dimensional Anomaly Detection
Qin Xie, Peng Zhang, Boseon Yu* and Jaesik Choi*
IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2021 (* co-corresponding author).Generative models Data mining
Paper
[57]
Interpreting Internal Activation Patterns in Deep Temporal Neural Networks by Finding Prototypes
Sohee Cho*, Wonjoon Chang*, GinKyeng Lee and Jaesik Choi
ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2021 (* contributed equally).Explainable AI Time series Data mining
Paper
[56]
Conditional Temporal Neural Processes with Covariance Loss
Boseon Yu, Jiwoo Lee, Janghoon Ju, Seijun Chung, Soyeon Kim and Jaesik Choi
International Conference on Machine Learning (ICML), 2021.Time series
Paper
[55]
Automatic Correction of Internal Units in Generative Neural Networks
Ali Tousi*, Haedong Jeong*, Jiyeon Han, Hwanil Choi and Jaesik Choi
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021 (* contributed equally).Generative models
Paper
[54]
Learning Compositional Sparse Gaussian Processes with a Shrinkage Prior
Anh Tong, Toan Tran, Hung Bui and Jaesik Choi
AAAI Conference on Artificial Intelligence (AAAI), 2021.Probabilistic inference
Paper
[53]
Characterizing Deep Gaussian Processes via Nonlinear Recurrence Systems
Anh Tong and Jaesik Choi
AAAI Conference on Artificial Intelligence (AAAI), 2021.Probabilistic inference
Paper
[52]
Interpreting Deep Neural Networks with Relative Sectional Propagation by Analyzing Comparative Gradients and Hostile Activations
Woo Jeoung Nam, Jaesik Choi and Seong-Whan Lee
AAAI Conference on Artificial Intelligence (AAAI), 2021.Explainable AI
Paper
[51]
Spatiotemporal characterization of GPCR activity and function during endosomal trafficking pathway
Hyunbin Kim, Hae Nim Lee, Jaesik Choi*, Jihye Seong*
ACS Analytical Chemistry, 2021. (IF: 6.78)Time series
Paper

2020

[50]
Visualization of autophagy progression by Red-Green-Blue autophagy sensor
Heejung Kim, Hyunbin Kim, Jaesik Choi, Kyung-Soo Inn and Jihye Seong
ACS Sensors, 2020. (IF: 7.33)Other
Paper
[49]
Interpreting and Explaining Deep Neural Networks: A Perspective on Time Series Data
Jaesik Choi
ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD), 2020 [tutorial link, video]Explainable AI Time series Data mining
Paper
[48]
Improved Predictive Deep Temporal Neural Networks with Trend Filtering
Youngjin Park, Deokjun Eom and Jaesik Choi
ACM International Conference on AI in Finance (ICAIF-2020), 2020 [slides].Time series
Paper
[47]
HetPipe: Enabling Large DNN Training on (Whimpy) Heterogeneous GPU Clusters through Integration of Pipelined Model Parallelism and Data Parallelism
Jay H. Park, Gyeongchan Yun, Chang M. Yi, Nguyen T. Nguyen, Seungmin Lee, Jaesik Choi, Sam H. Noh and Young-ri Choi
USENIX Annual Technical Conference (USENIX ATC), 2020.Other
Paper
[46]
A Single Multi-Task Deep Neural Network with Post-Processing for Object Detection with Reasoning and Robotic Grasp Detection
YongHyeok Seo, Dongju Shin, Jaesik Choi and Se Young Chun
International Conference on Robotics and Automation (ICRA), 2020.Reinforcement learning
Paper
[45]
An Efficient Explorative Sampling Considering the Generative Boundaries of Deep Generative Neural Networks
Giyoung Jeon*, Haedong Jeong* and Jaesik Choi
AAAI Conference on Artificial Intelligence (AAAI), 2020 (* contributed equally) (Oral).Generative models
Paper
[44]
Relative Attributing Propagation: Interpreting the Comparative Contributions of Individual Units in Deep Neural Networks
Woo-Jeoung Nam, Shir Gur, Jaesik Choi, Lior Wolf and Seong-Whan Lee
AAAI Conference on Artificial Intelligence (AAAI), 2020.Explainable AI
Paper
—
시계열 심층학습 모델의 은닉 노드에 대한 시각화
조소희, 최재식
정보과학회논문지, 2020.5, 제 47권 제5호, 445-453.Other
Paper
—
섀플리 값 추정을 이용한 주가 변화 원인 설명 문장 분류 모델 분석
천예은, 박영진, 성낙원, 최재식
정보과학회 컴퓨팅의 실제 논문지, 2020.4, 제 26권 제4호, 195-201.Other
Paper

2019

[43]
XAI—Explainable artificial intelligence
David Gunning, Mark Stefik, Jaesik Choi, Timothy Miller, Simone Stumpf and Guang-Zhong Yang
Science Robotics, 4(37), 2019. (IF: 19.4)Explainable AI Reinforcement learning
Paper
[42]
Markov Information Bottleneck to Improve Information Flow in Stochastic Neural Networks
Thanh T. Nguyen and Jaesik Choi
Entropy, 21(10), 976, 2019.Reinforcement learning
Paper
[41]
Confirmatory Bayesian Online Change Point Detection in the Covariance Structure of Gaussian Processes
Jiyeon Han*, Kyowoon Lee*, Anh Tong and Jaesik Choi
International Joint Conference on Artificial Intelligence (IJCAI), 2019 (* contributed equally).Time series Probabilistic inference
Paper
[40]
Discovering Latent Covariance Structure for Multiple Time Series
Anh Tong and Jaesik Choi
International Conference on Machine Learning (ICML), 2019.Time series
Paper
[39]
Implementation of End-to-End Training of Deep Visuomotor Policies for Manipulation of a Robotic Arm of Baxter Research Robot
Seongun Kim, Sol-A Kim, Rafael Lima, Jaesik Choi
Journal of Korea Robotics Society, 2019.Reinforcement learning
Paper
—
백스터 로봇의 시각기반 로봇 팔 조작 딥러닝을 위한 강화학습 알고리즘 구현
김성운, 김솔아, 하파엘 리마, 최재식
로봇학회 논문지, 2019.3, 제51호, 40-49.Other
Paper

2018

[38]
Dynamic Online Performance Optimization in Streaming Data Compression
Kade Gibson, Dongeun Lee, Jaesik Choi and Alexander Sim
IEEE International Conference on Big Data (IEEE Big Data), 2018.Other
Paper
[36]
Deep Reinforcement Learning in Continuous Action Spaces: a Case Study in the Game of Simulated Curling
Kyowoon Lee*, Sol-A Kim*, Jaesik Choi and Seong-Whan Lee
International Conference on Machine Learning (ICML), 2018 (* contributed equally).Reinforcement learning
Paper

2017

[34]
Generative Design of Electromagnetic Structures Through Bayesian Learning
Ramesh Patel, Kallol Roy, Jaesik Choi and Ki Jin Han
IEEE Transactions on Magnetics, 2017 (pdf).Generative models Probabilistic inference
Paper
[33]
Expanding Statistical Similarity Based Data Reduction to Capture Diverse Patterns
Dongeun Lee, Alex Sim, Jaesik Choi and John Wu
Data Compression Conference (DCC), 2017 (pdf).Other
Paper
[32]
Learning Execution Contexts from System Call Distribution for Anomaly Detection in Smart Embedded System
Man-Ki Yoon, Sibin Mohan, Jaesik Choi, Mihai Christodorescu and Lui Sha
The ACM/IEEE International Conference on Internet-of-Things Design and Implementation (IoTDI), 2017.Data mining
Paper
[31]
Improving Statistical Similarity Based Data Reduction for Non-Stationary Data
Dongeun Lee, Alex Sim, Jaesik Choi and John Wu
International Conference on Scientific and Statistical Database Management (SSDBM), 2017 (pdf).Other
Paper

2016

[30]
Global Deconvolutional Networks for Semantic Segmentation
Vladimir Nekrasov, Janghoon Ju and Jaesik Choi
The British Machine Vision Conference (BMVC), 2016 (source codes).Language models
Paper
[29]
Novel Data Reduction Based on Statistical Similarity
Dongeun Lee, Alex Sim, Jaesik Choi and John Wu
International Conference on Scientific and Statistical Database Management (SSDBM), 2016 (pdf, source codes).Other
Paper
[28]
Improving Imprecise Compressive Sensing Models
Dongeun Lee, Rafael Lima and Jaesik Choi
Conference on Uncertainty in Artificial Intelligence (UAI), 2016 (pdf).Other
Paper
[27]
Automatic Construction of Nonparametric Relational Regression Models for Multiple Time Series
Yunseong Hwang, Anh Tong and Jaesik Choi
International Conference on Machine Learning (ICML), 2016 (source codes).Time series Probabilistic inference
Paper
—
지식베이스 확장을 위한 자동 관계 추출
임성우, 한지연, 이교운, 최재식
한국정보과학회, 정보과학회지, 제34권 제9호 2016.9, 39-46.Other
Paper
—
분산환경에서의 기계학습
최재식
한국정보과학회, 정보과학회지, 제34권 제9호, 2016.9Other
Paper
—
딥러닝 기반 실시간 안개제거 알고리즘
최재식, 강상훈
한국특허, 출원번호:10-2016-0023568, 2016년 2월 26일.Other
—
인공지능시대와 제4차 산업혁명
최재식
울산매일, 2016.7.19.Other

2015

[26]
Extracting Baseline Electricity Usage Using Gradient Tree Boosting
Taehoon Kim, Dongeun Lee, Jaesik Choi, Anna Spurlock, Alex Sim, Annika Todd and Kesheng Wu
International Conference on Big Data Intelligence and Computing (DataCom), 2015 (Best Paper Award).Other
Paper
[25]
Reading Documents for Bayesian Online Change Point Detection
Taehoon Kim and Jaesik Choi
Conference on Empirical Methods on Natural Language Processing (EMNLP), 2015. (pdf, source codes)Language models Time series Probabilistic inference
Paper
[24]
A Scalable and Flexible Repository for Big Sensor Data
Dongeun Lee, Jaesik Choi and Heonshik Shin
IEEE Sensors Journal, 15(12), 2015, 7284-7294.Other
Paper
[23]
A Deterministic Partition Function Approximation for Exponential Random Graph Models
Wen Pu, Jaesik Choi, Yunseong Hwang and Eyal Amir
International Joint Conference on Artificial Intelligence (IJCAI), 2015 (pdf, slides).Data mining
Paper
[22]
Memory Heat Map: Anomaly Detection in Real-Time Embedded Systems Using Memory Behavior
Man-Ki Yoon, Sibin Mohan, Jaesik Choi and Lui Sha
Design Automation Conference (DAC), 2015 (pdf, BibTeX).Data mining
Paper
[21]
Best Predictive Generalized Linear Mixed Model with Predictive Lasso for High-Speed Network Data Analysis
Kejia Hu, Jaesik Choi, Alex Sim and Jiming Jiang
International Journal of Statistics and Probability, 4(2), 2015, 132-148.Other
Paper
[20]
Learning Compressive Sensing Models for Big Spatio-Temporal Data
Dongeun Lee and Jaesik Choi
2015 SIAM International Conference on Data Mining (SDM), 2015 (pdf).Time series Data mining
Paper
[19]
Learning Relational Kalman Filtering
Jaesik Choi, Eyal Amir, Tianfang Xu and Albert Valocchi
AAAI Conference on Artificial Intelligence (AAAI), 2015 (pdf, slides).Probabilistic inference
Paper
[18]
Semi-Local Structure Patterns for Robust Face Detection
Kyoungjoong Jeong, Jaesik Choi and Giljin Jang
IEEE Signal Processing Letters, 22(9), 2015, 1400-1403.Other
Paper
—
딥러닝 이론과 응용
최재식
최원익, , 한국정보과학회, 정보과학회지, 제33권 제8호, 2015.8Other
Paper
—
머신러닝(Machine Learning)
최재식
국회도서관, 전문가보고서, 미래창조과학방송통신위원회, 2015.6.Other
—
다양한 머신러닝 알고리즘
최재식
TECH M(머니투데이). 23. 26-29. 2015.Other
Paper

2014

[17]
Low-complexity compressive sensing with downsampling
Dongeun Lee, Jaesik Choi and Heonshik Shin
IEICE Electronic Express, 11(3), 2014, 20130947.Other
Paper
[16]
Use of Machine Learning Methods to Reduce Predictive Error of Groundwater Models
Tianfang Xu, Albert J. Valocchi, Jaesik Choi and Eyal Amir
Groundwater, 52(3), 2014, 448-460.Other
Paper
[15]
Low Complexity Sensing for Big Spatio-Temporal Data
Dongeun Lee and Jaesik Choi
IEEE 2014 International Conference on Big Data (IEEE BigData), 2014 (pdf).Time series
Paper
—
얼굴정렬과 AdaBoost를 이용한 얼굴 표정 인식
정경중, 최재식
장길진, , 전자공회논문지, 제51권 11호, 2014.11, 193-201.Other
Paper
—
빅데이터를 위한 확률 관계형 학습 모델의 변분 추론
최재식
한국정보과학회, 정보과학회지 , 제32권 제7호 2014.7, 15-20.Other
Paper

2013

[14]
A Spatio-Temporal Pyramid Matching for Video Retrieval
Jaesik Choi, Ziyu Wang, Sang-Chul Lee and Won J. Jeon
Computer Vision and Image Understanding, 117(6), 2013, 660-669.Time series
Paper
[13]
SecureCore: A Multicore based Intrusion Detection Architecture for Real-time Embedded Systems
Man-Ki Yoon, Sibin Mohan, Jaesik Choi, Jung-Eun Kim and Lui Sha
IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS), 2013.Other
Paper

2012

[12]
Lifted Relational Variational Inference
Jaesik Choi and Eyal Amir
Conference on Uncertainty in Artificial Intelligence (UAI), 2012 (pdf).Probabilistic inference
Paper
[11]
Improving Groundwater Flow Model Prediction Using Complementary Data-Driven Models
Tianfang Xu, Albert J. Valocchi,Jaesik Choi and Eyal Amir
XIX International Conference on Computational Methods in Water Resources (CMWR), 2012.Data mining
Paper

2011

[10]
Efficient Methods for Lifted Inference with Aggregate Factors
Jaesik Choi, Rodrigo de Salvo Braz and Hung Bui
AAAI Conference on Artificial Intelligence (AAAI), 2011 (pdf,).Probabilistic inference
Paper
[9]
Lifted Relational Kalman Filtering
Jaesik Choi, Abner Guzman Rivera and Eyal Amir
International Joint Conference on Artificial Intelligence (IJCAI), 2011 (pdf).Probabilistic inference
Paper

2010

[8]
Lifted Inference for Relational Continuous Models
Jaesik Choi, Eyal Amir and David J. Hill
Conference on Uncertainty in Artificial Intelligence (UAI), 2010 (pdf).Probabilistic inference
Paper

2009

[7]
Greedy Algorithms for Sequential Sensing Decisions
Hannaneh Hajishirzi, Afsaneh Shirazi, Jaesik Choi and Eyal Amir
International Joint Conference on Artificial Intelligence (IJCAI), 2009 (pdf).Other
Paper
[6]
Safe navigation of a mobile robot considering visibility of environment
Woojin Chung, Seokgyu Kim, Minki Choi, Jaesik Choi, Hoyeon Kim, Changbae Moon and Jae-bok Song
IEEE Transactions on Industrial Electronics, 56 (10), 2009.Reinforcement learning
Paper
[5]
Combining Planning and Motion Planning
Jaesik Choi, Eyal Amir
IEEE International Conference on Robotics and Automation (ICRA), 2009 (pdf).Reinforcement learning
Paper

2008

[4]
Spatio-Temporal Pyramid Matching for Sports Videos
Jaesik Choi, Won J. Jeon, Sang-Chul Lee
ACM International Conference on Multimedia Information Retrieval (MIR), 2008 (pdf).Time series
Paper
[3]
Virtual Sensor-Powered Spatiotemporal Aggregation and Transformation: A Case Study Analyzing Near-Real-Time NEXRAD and Precipitation Gage Data in a Digital Watershed
Yong Liu, David J. Hill, Tarek Abdelzaher, Jin Heo,Jaesik Choi, Barbara Minsker, David Fazio
Environmental Information Management (EIM), 2008.Time series Weather and climate

2007

[2]
Factor-Guided Motion Planning for a Robot Arm
Jaesik Choi and Eyal Amir
IEEE International Conference on Intelligent Robots and Systems (IROS), 2007 (pdf).Reinforcement learning
Paper

2005

[1]
Efficient navigation of mobile robot based on the robot’s experience in human co-existing environment
Jaesik Choi, Woojin Chung and Jae-Bok Song
International Conference on Control, Automation and Systems (ICCAS), 2005.Reinforcement learning

Jaesik Choi

Brief bio

Research interest

Thesis

Talks

News

Research

Granular Concept Circuits

Granular Concept Circuits: Toward a Fine-Grained Circuit Discovery for Concept Representations

Counterfactual Activation Editing

Counterfactual Activation Editing for Post-hoc Prosody and Mispronunciation Correction in TTS Models

Local Manifold Approximation and Projection

Local Manifold Approximation and Projection for Manifold-Aware Diffusion Planning

Creative Concept Catalyst

Enhancing Creative Generation on Stable Diffusion-based Models

Context-aware Speech Correction

An Automated Method to Correct Artifacts in Neural Text-to-Speech Models

Causal Interactions for Payoff Allocation

Rethinking Shapley Value for Negative Interactions in Non-convex Games

Neural ODE Transformers: Analyzing Internal Dynamics and Adaptive Fine-tuning

Non-autonomous Neural ODE Transformers

xPatch

xPatch: Dual-Stream Time Series Forecasting with Exponential Seasonal-Trend Decomposition

Diverse Rare Image Generation

Diverse Rare Sample Generation with Pretrained GANs

Understanding Distributed Representations of Concepts in Deep Neural Networks without Supervision

Relaxed Decision Regions

Towards Dynamic Trend Filtering through Trend Point Detection with Reinforcement Learning

Pathwise Explanation of ReLU Neural Networks

Towards Diverse Perspective Learning with Selection over Multiple Temporal Poolings

Memorizing Documents with Guidance In Large Language Models

Impact of Co-occurrence on Factual Knowledge of Large Language Models

Refining Diffusion Planner for Reliable Behavior Synthesis by Automatic Detection of Infeasible Plans

Explainable AI-Based Interface System for Weather Forecasting Model

Beyond Single Path Integrated Gradients for Reliable Input Attribution via Randomized Path Sampling

Variational Curriculum Reinforcement Learning for Unsupervised Discovery of Skills

Rarity Score: A New Metric to Evaluate the Uncommonness of Synthesized Images

Adaptive and Explainable Deployment of Navigation Skills via Hierarchical Deep Reinforcement Learning

An Efficient Explorative Sampling Considering the Generative Boundaries of Deep Generative Neural Networks