Publications

Overview figure of Visual Compliance via Executable Safety Rule Entailment — **Figure 1**Overview, from the paper

ProblemEnd-to-end visual safeguards struggle to adapt to evolving, complex safety rules and cannot explain verdicts.
ApproachGuardEn compiles safety policies into atomic propositions as executable code, grounded on scene graphs at test time.
Result+9.8 F1 over the strongest baseline on SafetyVisionBench, with interpretable rule-level reasoning.

EMNLP · 2026

Esc close · ↑↓ move

EMNLP Conference

MA-Mesh: Resilient Multi-Agent Collaboration via Role-based Self-Healing

Junyong Lee, Woo Kyung Kim, Honguk Woo

Code Demo

Overview figure of MA-Mesh: Resilient Multi-Agent Collaboration via Role-based Self-Healing — **Figure 1**Overview, from the paper

ProblemLLM multi-agent systems break down under agent failures, prompt errors and corrupted messages.
ApproachSynthesize a self-healing layer: cross-agent reflection spots degraded peers and role-based failover replaces them.
ResultBetter task performance and robustness than existing multi-agent systems on software, tool-use and science tasks.

EMNLP · 2026

Esc close · ↑↓ move

EMNLP Conference

ParentingLLM: Pacing Reasoning Maturity for Bidirectional Test-Time Scaling under Token Budgets

Seunghwan Jeong, Sugi Lee, Wei-Jin Park, Ikjun Yeom, Honguk Woo

Code Demo

arXiv Paper

Overview figure of ParentingLLM: Pacing Reasoning Maturity for Bidirectional Test-Time Scaling under Token Budgets — **Figure 1**Overview, from the paper

ProblemUnder fixed token budgets, reasoning LLMs either commit too early or never reach an answer.
ApproachA forced-commit probe estimates reasoning maturity, and Wait/Hurry cues keep it on a target schedule without fine-tuning.
ResultOn AIME 24+25 at a 1,024-token budget: 22.6 points lower burst rate and +4.0 to +7.3 accuracy.

EMNLP · 2026

Esc close · ↑↓ move

AACL-IJCNLP Conference Oral

Amortizing Self-Correction for Continual Embodied Adaptation Oral

Woo Kyung Kim, Jooyoung Kim, Minjong Yoo, Honguk Woo

Code Demo

arXiv

OpenReview

Overview figure of Amortizing Self-Correction for Continual Embodied Adaptation — **Figure 1**Overview, from the paper

ProblemEmbodied agents must keep adapting to new tasks without large expert supervision or reliable feedback to self-correct.
ApproachASC trains CoT jointly with correcting a prior policy's rationales, and stops reasoning once it is confident in the action.
Result+19.11% over the strongest baseline on unseen VirtualHome and ALFWorld tasks (5-shot), with 46% fewer generated tokens.

AACL-IJCNLP · 2026

Esc close · ↑↓ move

CIKM Conference

Code-as-Auditor: Executable Compliance Reasoning via Regulation-to-Code

Jisoo Kim, TaeYoon Kwack, Jinwoo Jang, Woo Kyung Kim, Honguk Woo

Code Demo

OpenReview

Overview figure of Code-as-Auditor: Executable Compliance Reasoning via Regulation-to-Code — **Figure 1**Overview, from the paper

ProblemLLM compliance judgments rarely ground themselves in explicit legal logic and evidence.
ApproachRegulations become checklists and executable decision trees, expanded into factual and counterfactual questions.
ResultMore accurate, evidence-backed audits on privacy and data-protection scenarios.

CIKM · 2026

Esc close · ↑↓ move

ECCV Conference

Multi-scale Mixture of World Models for Embodied Agents in Evolving Environments

Jinwoo Jang, Daniel J. Rho, Sihyung Yoon, Hyunsuk Cho, Honguk Woo

Code Demo

Overview figure of Multi-scale Mixture of World Models for Embodied Agents in Evolving Environments — **Figure 1**Overview, from the paper

ProblemMixture-of-world-model agents have no notion of scale, so knowledge at every scale is updated the same way.
ApproachScale-aware routing via experiential distance, with per-scale forgetting rates and gated inter-scale transfer.
ResultBeats state-of-the-art baselines on EmbodiedBench and HAZARD for multi-scale reasoning and adaptation.

ECCV · 2026 DOI

Esc close · ↑↓ move

IROS Conference

RoboBRIDGE: A Modular Framework for Bridging Policies to Robust Real-World Robotic Agents

Sihyung Yoon, Minjong Yoo, Sanghyun Ahn, Seojeong Choi, Honguk Woo

Code Demo

OpenReview

Overview figure of RoboBRIDGE: A Modular Framework for Bridging Policies to Robust Real-World Robotic Agents — **Figure 1**Overview, from the paper

ProblemPretrained VLAs predict actions well but lack failure recovery and long-horizon robustness as agents.
ApproachA five-module orchestration layer (Monitor, Perceptor, Planner, Controller, Robot Interface) around any VLA.
ResultOutperforms standalone and augmented VLAs on LIBERO, RoboCasa and multiple real robots.

IROS · 2026

Esc close · ↑↓ move

ICML Conference

Knothe-Rosenblatt Quantile Regression for Risk-sensitive Multi-objective Reinforcement Learning

Gwangpyo Yoo, Woo Kyung Kim, Honguk Woo

Code Demo

arXiv

Overview figure of Knothe-Rosenblatt Quantile Regression for Risk-sensitive Multi-objective Reinforcement Learning — **Figure 1**Overview, from the paper

ProblemReal decisions trade off several objectives with asymmetric risks, but multi-objective RL has no tractable way to handle risk.
ApproachKnothe-Rosenblatt quantile regression extends IQN to vector-risk measures, with a transformer critic without positional encoding and MO-TQC.
ResultImproved performance on MO-Gymnasium benchmarks, with critic convergence guaranteed under the distributional Bellman operator.

ICML · 2026

Esc close · ↑↓ move

ICML Conference

Functional Cache Grafting: Robust and Rapid Code-Policy Synthesis for Embodied Agents

Saehun Chun, Wonje Choi, Sera Choi, Sanghyun Ahn, Honguk Woo

Code Demo

Overview figure of Functional Cache Grafting: Robust and Rapid Code-Policy Synthesis for Embodied Agents — **Figure 1**Overview, from the paper

ProblemCode-policy LLMs are slow (repeated prefill) and brittle (API mismatches, missing safety guards).
ApproachRetrieve validated function skeletons with their KV caches, then stitch and patch them into a new policy.
Result+18.31% task success and 2.3× faster policy synthesis than RAGCache.

ICML · 2026

Esc close · ↑↓ move

ICML Conference

Efficient Skill Grounding via Code Refactoring with Small Language Models

Sera Choi, Wonje Choi, Saehun Chun, Daehee Lee, Jooyoung Kim, Chaeun Lee, Honguk Woo

Code Demo

Overview figure of Efficient Skill Grounding via Code Refactoring with Small Language Models — **Figure 1**Overview, from the paper

ProblemSmall LMs cannot ground reusable skills to new embodiments, and large LMs are impractical on-robot.
ApproachSkills as code: keep the control structure, refactor only the execution bindings locally.
ResultBest sLM-based Code-as-Policies method, matching LLM-based CaP across robot embodiments.

ICML · 2026

Esc close · ↑↓ move

CVPR Conference

Cross-Domain Demo-to-Code via Neurosymbolic Counterfactual Reasoning

Jooyoung Kim, Wonje Choi, Younguk Song, Honguk Woo

Code Demo

Overview figure of Cross-Domain Demo-to-Code via Neurosymbolic Counterfactual Reasoning — **Figure 1**Overview, from the paper

ProblemVideo-to-code robot programming breaks when demo and deployment domains differ procedurally.
ApproachAbstract demos into symbolic trajectories and use counterfactual states to propose verified procedure revisions.
Result+31.14% task success over Statler in simulated and real-world manipulation.

CVPR · 2026

Esc close · ↑↓ move

ICLR Conference

Test-Time Mixture of World Models for Embodied Agents in Dynamic Environments

Jinwoo Jang, Minjong Yoo, Sihyung Yoon, Honguk Woo

Code Demo

Overview figure of Test-Time Mixture of World Models for Embodied Agents in Dynamic Environments — **Figure 1**Overview, from the paper

ProblemMixture-of-Experts routing is frozen after deployment, so embodied agents cannot adapt to unseen domains.
ApproachTest-time routing over world models: multi-granular prototypes, test-time refinement, distilled augmentation.
ResultStrong zero-shot adaptation and few-shot expansion on VirtualHome, ALFWorld and RLBench.

ICLR · 2026

Esc close · ↑↓ move

2025

NeurIPS Conference SpotlightTop 3.2%

Towards Reliable Code-as-Policies: A Neuro-Symbolic Framework for Embodied Task Planning SpotlightTop 3.2%

Sanghyun Ahn, Wonje Choi, Junyong Lee, Jinwoo Park, Honguk Woo

Code

ProblemLLM-written robot code fails in dynamic, partially observable scenes due to weak grounding.
ApproachNeuro-symbolic verification plus exploratory code that probes the environment for missing observations.
Result+46.2% success over Code-as-Policies and 86.8% executability on RLBench and real robots.

NeurIPS · 2025 DOI

Esc close · ↑↓ move

NeurIPS Conference SpotlightTop 3.2%

Policy Compatible Skill Incremental Learning via Lazy Learning Interface SpotlightTop 3.2%

Daehee Lee, Dongsu Lee, TaeYoon Kwack, Wonje Choi, Honguk Woo

Code Demo

Overview figure of Policy Compatible Skill Incremental Learning via Lazy Learning Interface — **Figure 1**Overview, from the paper

ProblemAs an agent keeps learning new skills, existing skill-based policies stop being compatible.
ApproachA bilateral lazy-learning interface maps policy subtasks to the evolving skill space by trajectory similarity.
ResultImproved skills lift downstream policies with no policy retraining.

NeurIPS · 2025 DOI

Esc close · ↑↓ move

NeurIPS Conference

NeSyPr: Neurosymbolic Proceduralization For Efficient Embodied Reasoning

Wonje Choi, Jooyoung Kim, Honguk Woo

Code Demo

Overview figure of NeSyPr: Neurosymbolic Proceduralization For Efficient Embodied Reasoning — **Figure 1**Overview, from the paper

ProblemOn-robot LMs cannot call large reasoners or symbolic planners under latency and connectivity limits.
ApproachCompile symbolic plans into composable procedures that fold multi-step planning into single-step LM inference.
ResultOut-reasons large reasoning models and a symbolic planner on PDDLGym, VirtualHome and ALFWorld with compact LMs.

NeurIPS · 2025 DOI

Esc close · ↑↓ move

ICML Conference

World Model Implanting for Test-time Adaptation of Embodied Agents

Minjong Yoo, Jinwoo Jang, Sihyung Yoon, Honguk Woo

Code

ProblemEmbodied agents need to adapt to new domains without heavy data collection or retraining.
ApproachImplant domain-specific world models into an LLM policy at test time via retrieval and compound attention.
ResultSuperior zero- and few-shot performance on VirtualHome and ALFWorld unseen domains.

ICML · 2025

Esc close · ↑↓ move

ICLR Conference

NeSyC: A Neuro-symbolic Continual Learner For Complex Embodied Tasks in Open Domains

Wonje Choi, Jinwoo Park, Sanghyun Ahn, Daehee Lee, Honguk Woo

Code Demo

Overview figure of NeSyC: A Neuro-symbolic Continual Learner For Complex Embodied Tasks in Open Domains — **Figure 1**Overview, from the paper

ProblemEmbodied agents fail when a domain shift breaks the actionable knowledge they learned from limited experience.
ApproachLLM-generated hypotheses are validated contrastively with symbolic tools, and memory-based monitoring triggers knowledge refinement.
Result+33.6–53.7% task success over AutoGen on ALFWorld, VirtualHome, Minecraft and RLBench, plus a real robot.

ICLR · 2025

Esc close · ↑↓ move

ICLR Conference

Model Risk-sensitive Offline Reinforcement Learning

Gwangpyo Yoo, Honguk Woo

Code Demo

arXiv

Overview figure of Model Risk-sensitive Offline Reinforcement Learning — **Figure 1**Overview, from the paper

ProblemRisk-sensitive offline RL misjudges risk when small return-estimation errors meet distribution shift.
ApproachMinimize the worst-case risk over plausible scenarios found by a critic ensemble, with Fourier-feature quantile regression.
Result11.2–18.5% lower risk than the best risk-sensitive offline RL baseline in finance and self-driving.

ICLR · 2025

Esc close · ↑↓ move

AAAI Conference

In-Context Policy Adaptation via Cross-Domain Skill Diffusion

Minjong Yoo, Woo Kyung Kim, Honguk Woo

Code Demo

Overview figure of In-Context Policy Adaptation via Cross-Domain Skill Diffusion — **Figure 1**Overview, from the paper

ProblemSkill-based RL policies must adapt across domains with no model updates and little target data.
ApproachCross-domain skill diffusion: shared prototype skills plus a domain-grounded adapter, steered by domain prompts.
ResultSuperior in-context adaptation on Metaworld and CARLA across dynamics, embodiment and horizon shifts.

AAAI · 2025 DOI

Esc close · ↑↓ move

2024

NeurIPS Conference

Incremental Learning of Retrievable Skills For Efficient Continual Task Adaptation

Daehee Lee, Minjong Yoo, Woo Kyung Kim, Wonje Choi, Honguk Woo

Code Demo

Overview figure of Incremental Learning of Retrievable Skills For Efficient Continual Task Adaptation — **Figure 1**Overview, from the paper

ProblemAdapter-based continual imitation isolates tasks, so demonstrations cannot share knowledge.
ApproachIncrementally learn retrievable skills through prototype-based memory, each on its own adapter.
ResultRobust, sample-efficient adaptation on Franka-Kitchen and Meta-World, and it supports task unlearning.

NeurIPS · 2024 DOI

Esc close · ↑↓ move

NeurIPS Conference

LLM-based Skill Diffusion for Zero-shot Policy Adaptation

Woo Kyung Kim, Youngseok Lee, Jooyoung Kim, Honguk Woo

Code Demo

arXiv

Overview figure of LLM-based Skill Diffusion for Zero-shot Policy Adaptation — **Figure 1**Overview, from the paper

ProblemSkill-based policies cannot adapt to unseen contexts such as a target speed or low power use.
ApproachAn LLM writes loss functions that guide a skill diffusion planner with sequential in-painting, refined in a closed loop.
Result18.2% higher success than Diffuser+Guidance on MetaWorld, adapting zero-shot to new contexts.

NeurIPS · 2024 DOI

Esc close · ↑↓ move

NeurIPS Conference

Exploratory Retrieval-Augmented Planning For Continual Embodied Instruction Following

Minjong Yoo, Jinwoo Jang, Wei-Jin Park, Honguk Woo

Code

ProblemLLM planners lose track of a changing world when following many continual instructions at once.
ApproachInformation-based exploration builds an environment memory that is queried, with temporal-consistency refinement.
ResultHigher goal success and efficiency than LLM planners on VirtualHome, ALFRED and CARLA.

NeurIPS · 2024 DOI

Esc close · ↑↓ move

EMNLP Findings Conference

LLM-Based Offline Learning for Embodied Agents via Consistency-Guided Reward Ensemble

Yujeong Lee, Sangwoo Shin, Wei-Jin Park, Honguk Woo

Code Demo

Overview figure of LLM-Based Offline Learning for Embodied Agents via Consistency-Guided Reward Ensemble — **Figure 1**Overview, from the paper

ProblemUsing LLMs directly as embodied agents is costly and poorly grounded.
ApproachUse the LLM only at training time: an ensemble of spatio-temporally consistent rewards for offline RL.
ResultA 117M-param policy rivals 8B LLM agents on VirtualHome.

EMNLP Findings · 2024 DOI

Esc close · ↑↓ move

ACL Findings Conference

Semantic Skill Grounding for Embodied Instruction-Following in Cross-Domain Environments

Sangwoo Shin, SeungHyun Kim, Youngsoo Jang, Moontae Lee, Honguk Woo

Code Demo

Overview figure of Semantic Skill Grounding for Embodied Instruction-Following in Cross-Domain Environments — **Figure 1**Overview, from the paper

ProblemPretrained skills are entangled with domain knowledge, so LM planners fail to ground them in new domains.
ApproachIteratively decompose skills top-down through a semantic hierarchy until each is executable in the target domain.
ResultEffective across 300 cross-domain instruction-following scenarios in VirtualHome.

ACL Findings · 2024 DOI

Esc close · ↑↓ move

ICML Conference

Embodied CoT Distillation From LLM To Off-the-shelf Agents

Wonje Choi, Woo Kyung Kim, Minjong Yoo, Honguk Woo

Code Demo

Overview figure of Embodied CoT Distillation From LLM To Off-the-shelf Agents — **Figure 1**Overview, from the paper

ProblemLLM reasoning is too heavy for timely decisions on off-the-shelf devices.
ApproachDistill LLM embodied reasoning into small-LM reasoning and planning policies, aided by an embodied knowledge graph.
ResultSurpasses leading language-planning and distillation methods on ALFRED.

ICML · 2024

Esc close · ↑↓ move

IJCAI Conference

Offline Policy Learning via Skill-step Abstraction for Long-horizon Goal-Conditioned Tasks

Donghoon Kim, Minjong Yoo, Honguk Woo

Code Demo

Overview figure of Offline Policy Learning via Skill-step Abstraction for Long-horizon Goal-Conditioned Tasks — **Figure 1**Overview, from the paper

ProblemLong-horizon goal-conditioned RL suffers from sparse rewards and goal distribution shift.
ApproachLearn skill-step abstractions offline and a policy hierarchy that allows parameter-efficient fine-tuning.
ResultCompetitive zero- and few-shot adaptation on maze and Franka Kitchen.

IJCAI · 2024 DOI

Esc close · ↑↓ move

IJCAI Conference

Pareto Inverse Reinforcement Learning for Diverse Expert Policy Generation

Woo Kyung Kim, Minjong Yoo, Honguk Woo

Code Demo

Overview figure of Pareto Inverse Reinforcement Learning for Diverse Expert Policy Generation — **Figure 1**Overview, from the paper

ProblemFull Pareto-optimal policy sets rarely come with datasets covering every preference.
ApproachIRL regularised by reward distance builds a Pareto set from two datasets, distilled into one diffusion model.
ResultDense Pareto-frontier approximation, demonstrated on autonomous driving in CARLA.

IJCAI · 2024 DOI

Esc close · ↑↓ move

CVPR Conference

Model Adaptation for Time Constrained Embodied Control

Jaehyun Song, Minjong Yoo, Honguk Woo

Code Demo

Overview figure of Model Adaptation for Time Constrained Embodied Control — **Figure 1**Overview, from the paper

ProblemEmbodied control must meet task-specific inference deadlines on limited resources.
ApproachTreat model adaptation as dynamic routing over a modular network, with time limits as objectives.
ResultBetter performance and deadline adherence in manipulation and driving.

CVPR · 2024 DOI

Esc close · ↑↓ move

AAAI Conference

Robust Policy Learning via Offline Skill Diffusion

Woo Kyung Kim, Minjong Yoo, Honguk Woo

Code Demo

Overview figure of Robust Policy Learning via Offline Skill Diffusion — **Figure 1**Overview, from the paper

ProblemOffline-learned skills are tied to their dataset domain and transfer poorly.
ApproachA guided diffusion skill decoder disentangles domain-invariant behavior from domain variations.
ResultOutperforms skill-based imitation and RL in few-shot imitation and online RL.

AAAI · 2024 DOI

Esc close · ↑↓ move

AAAI Conference

Risk-Conditioned Reinforcement Learning: A Generalized Approach for Adapting to Varying Risk Measures

Gwangpyo Yoo, Jinwoo Park, Honguk Woo

Code Demo

arXiv Paper

Overview figure of Risk-Conditioned Reinforcement Learning: A Generalized Approach for Adapting to Varying Risk Measures — **Figure 1**Overview, from the paper

ProblemRisk-sensitive RL policies must be retrained whenever the preferred risk measure changes.
ApproachA unified risk representation (WV@R) and a risk proposal network let one policy condition on any risk measure.
ResultMatches per-measure baselines on locomotion, finance and self-driving, and often beats them, without retraining.

AAAI · 2024 DOI

Esc close · ↑↓ move

AAAI Conference

SemTra: A Semantic Skill Translator for Cross-Domain Zero-Shot Policy Adaptation

Sangwoo Shin, Minjong Yoo, Jeongwoo Lee, Honguk Woo

Code Demo

Overview figure of SemTra: A Semantic Skill Translator for Cross-Domain Zero-Shot Policy Adaptation — **Figure 1**Overview, from the paper

ProblemMulti-modal prompts should specify long-horizon tasks that run in a different domain.
ApproachExtract semantic skills, translate them with an LM, then adapt each skill to the target context.
ResultOne-shot task inference with zero-shot domain transfer on Meta-World, Franka Kitchen, RLBench and CARLA.

AAAI · 2024 DOI

Esc close · ↑↓ move

2023

NeurIPS Conference

Efficient Policy Adaptation with Contrastive Prompt Ensemble for Embodied Agents

Wonje Choi, Woo Kyung Kim, SeungHyun Kim, Honguk Woo

Code Demo

Overview figure of Efficient Policy Adaptation with Contrastive Prompt Ensemble for Embodied Agents — **Figure 1**Overview, from the paper

ProblemRL agents fail zero-shot under unseen visual and physical changes.
ApproachEnsemble contrastively learned visual prompts on a VLM via guided attention for robust state representations.
ResultBeats state-of-the-art methods in AI2THOR, egocentric Meta-World and CARLA, with better sample efficiency.

NeurIPS · 2023

Esc close · ↑↓ move

ICML Conference

One-shot Imitation in a Non-Stationary Environment via Multi-Modal Skill

Sangwoo Shin, Daehee Lee, Minjong Yoo, Woo Kyung Kim, Honguk Woo

Code Demo

Seunghwan Jeong, Honguk Woo

Overview figure of One-shot Imitation in a Non-Stationary Environment via Multi-Modal Skill — **Figure 1**Overview, from the paper

ProblemOne-shot imitation breaks down for complex tasks in non-stationary environments.
ApproachInfer a semantic skill sequence with a VLM and adapt each skill to hidden dynamics by meta-learning.
ResultSuperior one-shot imitation and dynamics generalization on multi-stage Meta-World.

ICML · 2023

Esc close · ↑↓ move

IEEE Access Journal

A Configurable Model-Based Reinforcement Learning Framework for Disaggregated Storage Systems

Code Demo

arXiv Paper

3-line summary coming soon.

IEEE Access · 2023 DOI

Esc close · ↑↓ move

Applied Sciences Journal

A Maturity Model for Trustworthy AI Software Development

Seunghwan Cho, Ingyu Kim, Jinhan Kim, Honguk Woo, Wanseon Shin

Code Demo

arXiv Paper

3-line summary coming soon.

Applied Sciences · 2023 DOI

Esc close · ↑↓ move

2022

NeurIPS Conference Oral

Skills Regularized Task Decomposition for Multi-task Offline Reinforcement Learning Oral

Minjong Yoo, SangWoo Cho, Honguk Woo

Code Demo

Woo Kyung Kim, Youngseok Lee, Honguk Woo

OpenReview

Overview figure of Skills Regularized Task Decomposition for Multi-task Offline Reinforcement Learning — **Figure 1**Overview, from the paper

ProblemMulti-task offline RL degrades when datasets come from behavior policies of mixed quality.
ApproachJointly embed skills and tasks with a WAE, using quality-weighted regularization and imaginary trajectories.
ResultRobust to mixed-quality data in robotic manipulation and drone navigation.

NeurIPS · 2022

Esc close · ↑↓ move

CCCE Conference

Federated Offline Reinforcement Learning for Autonomous Systems

Ju-eun Park, Honguk Woo

Code Demo

arXiv Paper

3-line summary coming soon.

CCCE · 2022 DOI

Esc close · ↑↓ move

Applied Sciences Journal

Sample-Efficient Deep Learning Techniques for Burn Severity Assessment with Limited Data Conditions

Hyunkyung Shin, Hyeonung Shin, Wonje Choi, Jaesung Park, Minjae Park, Euiyul Koh, Honguk Woo

Code Demo

arXiv Paper

3-line summary coming soon.

Applied Sciences · 2022 DOI

Esc close · ↑↓ move

Journal of Sensors Journal

Toward an Effective Bug Triage System Using Transformers to Add New Developers

Syed Farhan Alam Zaidi, Honguk Woo, Chan-Gun Lee

Code Demo

arXiv Paper

3-line summary coming soon.

Journal of Sensors · 2022 DOI

Esc close · ↑↓ move

ICCAI Conference

Integrating A Deep Learning-based Plane Detector in Mobile AR Systems for Improvement of Plane Detection

Yunho Kim, Honguk Woo

Code Demo

arXiv Paper 2022 the 8th International Conference on Computing and Artificial Intelligence (ICCAI)

3-line summary coming soon.

2022 the 8th International Conference on Computing and Artificial Intelligence (ICCAI) · 2022 DOI

Esc close · ↑↓ move

IEEE Access Journal

A Graph Convolution Network-Based Bug Triage System to Learn Heterogeneous Graph Representation of Bug Reports

Syed Farhan Alam Zaidi, Honguk Woo, Chan-Gun Lee

Code Demo

arXiv Paper

3-line summary coming soon.

IEEE Access · 2022 DOI

Esc close · ↑↓ move

ICMVA Conference

Transfer Learning based Precise Pose Estimation with Insufficient Data

Wonje Choi, Honguk Woo

Code Demo

arXiv Paper 2022 the 5th International Conference on Machine Vision and Applications (ICMVA)

3-line summary coming soon.

2022 the 5th International Conference on Machine Vision and Applications (ICMVA) · 2022 DOI

Esc close · ↑↓ move

ICMVA Conference

Iterative Pruning-based Model Compression for Pose Estimation on Resource-constrained Devices

SungHyun Choi, Wonje Choi, Youngseok Lee, Honguk Woo

Code Demo

arXiv Paper 2022 the 5th International Conference on Machine Vision and Applications (ICMVA)

3-line summary coming soon.

2022 the 5th International Conference on Machine Vision and Applications (ICMVA) · 2022 DOI

Esc close · ↑↓ move

ICMVA Conference

Mean-variance Based Risk-sensitive Reinforcement Learning with Interpretable Attention

Code Demo

arXiv Paper 2022 the 5th International Conference on Machine Vision and Applications(ICMVA)

3-line summary coming soon.

2022 the 5th International Conference on Machine Vision and Applications(ICMVA) · 2022 DOI

Esc close · ↑↓ move

ICOIN Conference

Reinforcement Learning based Load Balancing in a Distributed Heterogeneous Storage System

Jooyoung Park, Seunghwan Jeong, Honguk Woo

Code Demo

arXiv Paper International Conference on Information Networking (ICOIN)

3-line summary coming soon.

International Conference on Information Networking (ICOIN) · 2022 DOI

Esc close · ↑↓ move

AAAI Conference

Structure Learning-Based Task Decomposition for Reinforcement Learning in Non-stationary Environments

Honguk Woo, Gwangpyo Yoo, Minjong Yoo

Code Demo

arXiv Paper

3-line summary coming soon.

AAAI · 2022 DOI

Esc close · ↑↓ move

AAAI Conference

An Efficient Combinatorial Optimization Model Using Learning-to-Rank Distillation

Honguk Woo, Hyunsung Lee, Sangwoo Cho

Suin Lee, Youngseok Lee, Chan-Gun Lee, Honguk Woo

Code Demo

arXiv Paper IT Convergence and Security: ICITCS 2020

3-line summary coming soon.

IT Convergence and Security: ICITCS 2020 · 2021 DOI

Esc close · ↑↓ move

2020

IEEE Access Journal

rocorl: Transferable Reinforcement Learning-Based Robust Control for Cyber-Physical Systems With Limited Data Updates

Gwangpyo Yoo, Minjong Yoo, Ikjun Yeom, Honguk Woo

Code Demo

arXiv Paper

3-line summary coming soon.

IEEE Access · 2020 DOI

Esc close · ↑↓ move

arXiv Preprint

Fixed Priority Global Scheduling from a Deep Learning Perspective

Hyunsung Lee, Michael Wang, Honguk Woo

Code Demo