Running Composer's AI Cafe

Publications

Peer-Reviewed Journal Articles

  1. Taisei Hirayama, Kohei Yoshida, Hiroki Sakaji, Itsuki Noda, "Deliberate Assignment Deferral for Multi-Agent Pickup and Delivery with Deadlines," Systems, 14(5), 494, 2026. https://doi.org/10.3390/systems14050494
    Impact Factor 2024: 3.1; CiteScore 2024: 4.1; Q1, top 10% in "Social Sciences, Interdisciplinary".
  2. Tatsuto Ito, Taisei Hirayama, Naoki Hattori, Hiroki Sakaji, Itsuki Noda, "SL-LDA: LDA-Based Storage Location Assignment for Automated Warehouses Under MAPD Constraints," Systems, 14(5), 581, 2026. https://doi.org/10.3390/systems14050581
    Impact Factor 2024: 3.1; CiteScore 2024: 4.1; Q1, top 10% in "Social Sciences, Interdisciplinary".

Peer-Reviewed Conference Papers

  1. Taisei Hirayama, Kohei Yoshida, Hiroki Sakaji, Itsuki Noda, "Dynamic Haven Selection for Multi-Agent Pickup and Delivery in Constrained Warehouses," Joint Workshop on Planning for Complex Real-World Applications (CAIPI) and Bridging the Gap Between AI Planning and (Reinforcement) Learning (PRL), IJCAI-ECAI 2026, accepted, 2026.
  2. Taisei Hirayama, Kohei Yoshida, Hiroki Sakaji, Itsuki Noda, "Fixed-Haven Reservation for Online Multi-Agent Pickup and Delivery in Dense Warehouses," 14th Workshop on Planning and Robotics (PlanRob 2026), co-located with ICAPS 2026, accepted, 2026. Workshop page
  3. Taisei Hirayama, Kohei Yoshida, Hiroki Sakaji, Itsuki Noda, "SHARP: A Complete Multi-Task MAPD Planner for Realistic Warehouse Scenarios," Proceedings of the 18th International Conference on Agents and Artificial Intelligence (ICAART 2026), 2026. https://api.semanticscholar.org/CorpusID:286509981
  4. Taisei Hirayama, Kohei Yoshida, Hiroki Sakaji, Itsuki Noda, "Efficient Assignment of Immediate Tasks Using Deep Reinforcement Learning in Multi-Agent Pickup and Delivery," Advances in Practical Applications of Agents, Multi-Agent Systems, and Computational Social Science (PAAMS 2025), pp. 166-178, Springer, 2025. https://doi.org/10.1007/978-3-032-07638-0_14
  5. Taisei Hirayama, Itsuki Noda, Hiroki Sakaji, Norihiko Kato, "Optimal Assignment of Immediate Tasks in Multi-agent Pickup and Delivery," Technologies and Applications of Artificial Intelligence (TAAI 2024), Communications in Computer and Information Science, vol. 2415, Springer, 2025. https://doi.org/10.1007/978-981-96-4596-1_6

Domestic Conferences

  1. Taisei Hirayama, Akihiro Matsufuji, Yuki Ogawa, Hiroki Sakaji, Itsuki Noda, "Evidence-Complemented RAG for Regulatory Document QA Using Document Structure Graphs," JSAI2026, 2026.
  2. Taisei Hirayama, Akihiro Matsufuji, Yuki Ogawa, Hiroki Sakaji, Itsuki Noda, "Construction of Hierarchical Tree-Structured Datasets from Legal and Standards PDFs," NLP2026, 2026.
    Paper Poster
  3. Taisei Hirayama, Kohei Yoshida, Hiroki Sakaji, Itsuki Noda, "Improving Interrupt Handling Efficiency in Multi-Agent Pickup and Delivery Using Deep Reinforcement Learning," JSAI2025, 2025.

Awards & Certifications

Hackathon

Kaggle

Active as a Competitions Expert (profile)

Bronze medals

Signate

Active as an Advanced member

Bronze medal

Others

Prize winner

Certifications

Experience

Internships

September-October 2025 (5 weeks)

Recruit Co., Ltd.

RECRUIT INTERNSHIP for Data Specialists 2025

August 2025 (2 weeks)

Toyota Central R&D Labs., Inc.

R&D Positions SUMMER INTERNSHIP

Part-Time Work

September 2024-September 2025

Kurusugawa Computer Inc. (research part-time)

  • Surveyed methods for building LLM datasets with LangSmith
  • Evaluated AutoML OSS using Kaggle and SIGNATE
  • Participated in an image anomaly detection project for about 4 months
    • Handled paper surveys, implementation, experiments, tuning, and customer-facing materials
    • Worked on development while considering evaluation metric design and practical constraints such as compute resources and runtime

May 2021-August 2024

Sapporo Shingaku Plaza Group (tutor)

  • Taught group classes across multiple courses, instructing more than 200 students
  • Created curricula, selected teaching materials, and prepared instructor guides
  • Delivered multiple talks for new customer acquisition
  • Developed the ability to explain complex topics clearly

Projects

🎵 Music Visualizer

Visualize your music. Create a stylish music video just by uploading a track.

Open app

📑 Paper Survey

Related MAPF/MAPD papers are automatically summarized every day in Japanese using the Ochiai format.

Open app

Study Notes

Technical notes I compiled for learning. These notes are currently available in Japanese.

🔥 Deep Learning

🎮 Reinforcement Learning

📖 Other

Graduate Entrance Exam Prep

Notes for the entrance examination for the master's program at Hokkaido University Graduate School of Information Science and Technology. These notes are currently available in Japanese.