Our paper, From Token Correction to Claim Admission: Training-Free Hallucination Mitigation for Large Vision-Language Models, is now available on arXiv.
News
Lab Updates
Research, publication, project, and community updates from TML.
Our paper, Rethinking Memorization Mitigation in Diffusion Models: Reinforcing Text Conditioning, is now available on arXiv.
TML joined Jeju National University for the AI Star Fellowship project workshop at Camphortree Hotel, Jeju, from October 1–3, 2026.
Our paper, Continual Data Unlearning in Diffusion Models via Transition-based Regularization, is now available on arXiv.
Our paper, Mechanistic Circuit Identification for Controllable Data Generation, has been accepted to NeurIPS 2026.
Our paper, Mitigating Sequential Reappearance in Diffusion Data-Point Unlearning, coauthored with TML students at Konkuk University, is now available on arXiv.
Our lab has added two additional NVIDIA RTX PRO 6000 GPUs to its research computing infrastructure.
Our paper, Mechanistic Circuit Identification for Controllable Data Generation, is now available on arXiv.
Our paper, Function-Level Execution Feedback for Code Preference Optimization, has been accepted to EMNLP 2026 Findings.
Our paper, Geometry-Preserving Robust Neural Reconstruction via Statistical Reweighting, has been accepted to BMVC 2026.
Our team has been selected for the AI Star Fellowship 2026 in collaboration with Jeju National University, NC AI, AIVIS, and Metsakuur. Prof. Hong will serve as Project 2 Leader, focusing on proactive security against multi-unit deepfakes and secure authentication of AI agents. The six-year project has total funding of KRW 11.0 billion.
Our team has been selected to lead a KETI-funded project with access to a cluster of eight NVIDIA B200 GPUs with 1,536 GB total VRAM.
Our team at Konkuk University has been selected for the Group Research Workforce Development Program.
Our paper on Efficient Large Language Models has been accepted to IEEE Transactions on Circuits and Systems II: Express Briefs.
Our paper on Efficient VLA Pretraining has been accepted to ICML 2026.
Our team has been selected to lead the Global Talent Development in Copyright Technology project, supported by MCST and KOCCA.
Our team has been selected to lead the Advanced Industry Talent Training Bootcamp project, supported by KIAT.
Prof. Hong will serve as an Area Chair for NeurIPS 2026.
Our paper about Efficient Large Language Models has been accepted to ACL 2026.
Our paper about Fair and Efficient Models has been accepted to IEEE Transactions on Image Processing.
Our team has been selected to lead the Verifiable Privacy-Preserving Machine Unlearning project.
Our paper about Efficient Fidelity Estimation for Quantum State has been accepted to Physical Review A.
Our paper about communication using AI has been accepted to IEEE Wireless Communications Letters.
Our paper about Efficient VLA Pretraining has been released on arXiv.
Prof. Hong delivered invited talks on XAI and on generative AI and privacy.
Our lab introduced a GPU server equipped with two NVIDIA RTX Pro 6000 GPUs.
Our lab established a partnership with Theta Labs by signing an MOU.
Our paper about reliable data generation in generative AI has been accepted to AAAI 2026 as an oral presentation.
Our paper about enhancing data privacy in generative AI has been accepted to NeurIPS 2025.
Our team has been selected to lead a project on coded parallel tensor network contraction for robust and efficient quantum circuit simulation.
Our team at Konkuk University has been selected for the Research Group Fostering Program.
Our paper about Explainable AI has been accepted to AAAI 2025.
Our paper about Fair AI has been accepted to NeurIPS 2024, and our paper about 3D Surface Reconstruction has been accepted to ACCV 2024.
Our paper about defending adversarial attacks has been accepted to TIFS 2024.