google scholar

Preparation

• Sinkhorn distance for encrypted data
• Hosik Choi, Minje Park, Jaeseon Kim, Sungchul Shin, Cheolwoo Park, Changdae Oh, Kyungwoo Song, Jong-jun Jeon, Cheolwoo Park*(2025+) Group rank for encrypted data
• Heejin Kim et. al. (2026+). Sparse association rule analysis using combinatorial optimization.
• Dayoung Lee, Cholwoo Park, and Hosik Choi* (2026+). Learning rank axis in embedding space
• Jihyeon Hwang, Hyeseong Lee and Hosik Choi* (2026+) XGBoosting with LLM–Tabular Foundation Model
• Dawun Lee and Hosik Choi* (2026+) A Competence-Aware Cross-Model Verification for Medical Multiple-Choice Question Answering
• Junsik Kim and Hosik Choi* (2026+) Graph-CDM: Heterogeneous Graph Representation with Edge Imputation for ICU Sepsis Risk Stratification in OMOP-CDM Data
• Soo-Heang Abel Eo, Garam Lee, Hosik Choi(2026+) BoundaryVault: Confidential Policy-Boundary Precedent Retrieval for Multimodal Guardrails
• Jaekyeong Jung, Chanhyeok Yoon, Jihyeon Hwang, Shinjeong Hwang and Hosik Choi* (2026+). Auditing Agent Benchmarks with Cognitive Diagnostic Evidence

2026

• Dakyeong Gwak, Heejin Kim, Jihyeon Hwang, Juyoung Jeon, Minhoi Park, Kyungwoo Song and Hosik Choi* (2026+). CYREN: Temporal Dual-Prompt Optimization with Selective Tool Augmentation for Cyber-Scam Video Detection
• Jihyeon Hwang, Jaekyeong Jung, Shinjeong Hwang, BeomJin Park, Hosik Choi* (2026+). RAFT: Graph-Routed Adaptive Feature Transformation for Heterogeneous Tabular Learning, submitted.
• Shinjeong Hwang, Jihyeon Hwang, Sungman Lee, Gaeul Yang, Jinhee Choi and Hosik Choi* (2026+) From Answer Retrieval to Evidence-Guided KGQA: An SOP-Guided Framework for Toxicological Evidence Retrieval, submitted.
• Jiin Han, Hyeseong Lee, Hyeongboo Baek, and Hosik Choi* (2026+) Madness Is Already in the Room: A Black-Box Analysis of Answer Transitions, submitted.
• Jung et al.(2025+) FunctionalBERT: A language-model framework for tokenizing functional data in GMFCS level classification, submitted
• Hyungwoo Kim, Hosik, Choi, Woojoo Lee(2025+). Partially privacy-protected logistic regression with L1 penalty and sign constraint, submitted

published papers(conference)