---

ACL2026読み会@名大

---

概要

• 会場: 名古屋大学IB中棟011室
• 日時: 2026年9月25日(金) 10:00-18:00
• 形式: 1人15分（発表12分、質疑3分）

---

プログラム

• 10:00-10:05: Opening
• 10:05-10:35: Session1: モデル訓練・更新
  • Find Your Optimal Teacher: Personalized Data Synthesis via Router-Guided Multi-Teacher Distillation [木村]
  • Fisher-Driven Adaptive Locating for Knowledge Editing in Large Language Models [黄]
• 10:35-11:05: Session2: 言語・推論メカニズム
  • Memory efficiency and resource-rational encoding in sentence processing [木迫]
  • Immediate Inference: The Missing Foundation in Large Language Model Logical Reasoning [加藤]
• 11:05-11:25: Short Break
• 11:25-12:25: Session3: 拡散型言語モデル
  • Unlocking the Potential of Diffusion Language Models through Template Infilling [大畑]
  • DecoCal: Decoding with Calibration in Diffusion Large Language Models [樫森]
  • Efficient Self-Evaluation for Diffusion Language Models via Sequence Regeneration [山内]
  • d-TreeRPO: Towards More Reliable Policy Optimization for Diffusion Language Models [木下]
• 12:25-13:25: Lunch Break
• 13:25-13:55: Session4: 研究・学術応用
  • SciCoQA: Quality Assurance for Scientific Paper–Code Alignment [伊藤]
  • CoCoNUTS: Concentrating on Content while Neglecting Uninformative Textual Styles for AI-Generated Peer Review Detection [大鹿, slide]
• 13:55-14:25: Session5: 教育応用
  • EduBench: A Comprehensive Benchmarking Dataset for Evaluating Large Language Models in Diverse Educational Scenarios [清水]
  • Your Students Don’t Use LLMs Like You Wish They Did [笹野, slide]
• 14:25-14:45: Short Break
• 14:45-15:30: Session6: マルチモーダル
  • MMAC: A Multilingual, Multimodal Alignment Framework for Cultural Grounding Evaluation [徳久]
  • HowToNarrate: A General-Domain Benchmark for Synchronized Video Narration with External Knowledge [郭]
  • Semantic Hardness Is Not Visual Hardness: Sign-Aware Hard Negative Mining for Sign Language Retrieval [宇野]
• 15:30-15:50: Short Break
• 15:50-16:35: Session7: エージェント・記憶／知識
  • APEX-MEM: Agentic Semi-Structured Memory with Temporal Reasoning for Long-Term Conversational AI [西出]
  • AutoGraph-R1: End-to-End Reinforcement Learning for Knowledge Graph Construction [井田, slide]
  • Whose Facts Win? LLM Source Preferences under Knowledge Conflicts [辻本]
• 16:35-16:55: Short Break
• 16:55-17:25: Session8: エージェント・ツール利用
  • Lost in Execution: On the Multilingual Robustness of Tool Calling in Large Language Models [矢野]
  • SynthAgent: Adapting Web Agents with Synthetic Supervision [平松]
• 17:25-17:55: Session9: LLM評価・信頼性
  • ReTraceQA: Evaluating Reasoning Traces of Small Language Models in Commonsense Question Answering [大崎]
  • LLM-Powered Benchmark Factory: Reliable, Generic, and Efficient [小島]
• 17:55-18:00: Closing

---

発表者

• 木村優介（名大石川研）、加藤芳秀、伊藤滉一朗、黄晟昊（名大松原研）、笹野遼平、大鹿雅史、矢野千紘、木迫璃玖、郭慶安、辻本祥吾（名大笹野研）、徳久良子、清水綾太（愛工大徳久研）、井田龍希、西出隆盛、樫森信平、木下貴翔、大崎稜司、小島亮、大畑綾音、山内大暉、宇野駿希、平松優弥（豊田工大三輪研）

---

過去の読み会

• ACL2025読み会@名大
• ACL2024読み会@名大
• ACL2023読み会@名大
• ACL2019読み会@名大
• ACL2018読み会@名大

---