Please note that as I will be in Cambridge, UK until the end of April 2025, the supervision will be remote. However, full support will be provided.

I am recruiting graduate students as research assistants (RA) at the Artificial Intelligence Research Center (AIRC), National Institute of Advanced Industrial Science and Technology (AIST) in Tokyo, to study the following topics throughout the year. Research outcomes can be available as part of an academic degree.

Research Themes

A. Data-Driven Multi-Agent Path Planning

This is collaborative research with an industrial company.

Multi-robot automation has attracted particular attention in logistics systems. To build highly efficient systems, it is necessary to solve multi-agent path finding, which assigns collision-free paths to multiple agents. In recent years, dramatic progress has been made in this area, but its application is still limited to specific scenarios. We aim to establish versatile path planning technologies by incorporating machine learning into existing algorithms.

Required Skills

  • Experience with Python / C++
  • Knowledge of deep learning

Preferred Skills and Experience

  • Deep understanding of search algorithms.
  • Experience of using basic software to collaboration with other people, such as Git/GitHub, Docker, etc.
  • Publications at conferences or journals. Specific examples include:
    • AI related: AAAI, IJCAI, AAMAS, ICAPS, SoCS, AIJ, JAIR, JAAMAS
    • Robotics related: ICRA, IROS, RSS, CoRL, L4DC, DARS, MRS, T-RO, IJRR, RA-L
    • Workshops related to the above conferences

B. Applications of Multi-Agent Path Finding

Not limited to data-driven approaches, we are willing to explore scenarios where multi-agent path finding (MAPF) is useful. In particular, we aim to develop large-scale task assignment and lifelong operations in combination with MAPF.

Required / Preferred Skills

Similar to topic A, but the knowledge of machine learning can be minimal.

C. Development of Multi-Robot Hardware Platforms

You will be developing a research platform. The aim is to establish a stable platform for groups of small ground robots and small drones, and to conduct experiments to verify algorithms.

Required Skills

  • Hardware development (NB. KO does not have strong experience here.)

D. Collective Behavior Design for Swarm Robots

This is not a formal RA recruitment, but if you are interested please contact me. We will see what we can do. Research topics range from classical planning and control to automatic design of cooperative behaviour using machine learning.

Conditions

  • Target: Gradual students. Please consult with your academic advisor in advance. Bachelor’s students can also be accepted as technical staff.
  • Location: Tokyo Waterfront Center (or Tsukuba Center)
  • Salary:
    • Master student: 1,700 JPY per hour (standard is 7 working days, 4-14 days per month)
    • PhD student: 2,100 JPY per hour (standard is 14 working days, 10-14 days per month)
    • Bachelor student (technical stuff): 1,500 JPY per hour (standard is 7 working days, 4-14 days per month)
  • Working Hours: 9:30 AM - 6:15 PM (or 9:00 AM - 5:45 PM for Tsukuba)
  • Term: Assumed to be more than 3 months. Start and end dates can be adjusted.
  • Facilities / Supports: Laptop provided. For computing resources, GPU-equipped desktops or ABCI (cloud infrastructure) are available. Necessary expenses for research activities such as attendance at academic conferences will be covered.
  • Reference: Please also check [AIST: Research Assistant].

Others

Application

Please email your CV to okumura.k [at] aist.go.jp.

(I usually use Cambridge email address. Probably faster.)

Q&A

  • Can I use this as part of my master’s or doctoral thesis? — No problem. However, please consult with your academic advisor in advance.
  • What are the specific goals? — The goal is to submit papers to good venues. Regardless of the outcome, this will help your career.

References