Towards Achieving AI for Scientific Discovery: A Grand Challenge to Promote Strategic International Research in AI
Abstract
Scientific discovery is one of the most essential activities pushing our civilization. The goal of Nobel Turing Challenge is to deve lop AI scientist that can make major scientific discoveries autonomously some of which may worth Nobel Prize or beyond. Building AI scientist means that we will have an engine of scientific discovery, just like an invention of steam engine that triggered the indus trial revolution. It will reshape the way we do science and give rise to the alternative form of science. However, the Challenge req uires coordinated effort of international science community particularly artificial intelligence, robotics, computer science, biolog y, chemistry, and other relevant domains. The proposed project kick starts the effort by creating the minimum viable prototype to sh are the concept and forming the international network of scientists and technologists passionate in achieving the goal through a ser ies of workshop, mostly on-li ed by the relevant project by the Alan Turing Institute (supported by ONR Global), is expected to evolve to be the main thrust of th e challenge and evolve into a virtual big science project. The project also proposes to develop minimal viable prototype that can be the basis of AI scientist as well as providing concrete picture of AI scientist at the early stage. By developing a prototype based on open-API platform, ranges of development efforts can be integrated rather than fragmented to create an integrated AI scientist. Over next 3 years, the project will lay foundation of Nobel Truing Challenge and promote science and technology development with a s eries of spin-outs and transformative technological outcomes.The ultimate objective of the Nobel Turing Challenge is to develop art ificial intelligence system capable of making major scientific discoveries some of which may worthy of Nobel Prize and may win it. The research program to achieve this challenge (Nobel Turing Challenge Research Program; NTCRP) requires broad multi-institutional c ollaborations with flexible yet highly strategic planning and execution. It shall be the Apollo Project equivalent in AI. This propo sal aims to promote international collaboration toward NTCRP to accelerate research in AI for scientific discovery with emphasis on Indo-pacific region as well as robotics and systems development. A successful AI Scientist achieving the goal shall be a highly au tonomous system that closed-the-loop of hypothesis-generation, experimental design, and execution of experiments. In addition, it sh all be capable to making strategic choices of research topics and problems to be solved. AI, robotics, and wide range of devices sha ll be tightly integrated with consistent data accessible over the platform. Due to the complexity of the challenge, it is essential to form broad based international network of collaboration with shared vision. In addition, it is critical to develop a series of pr ototype demonstrating feasibility of the challenge that can be the basis of technology platform for NTCRP. A series of comprehensive documents describing directions, opportunities, outlining key milestones, scientific and technical issues shall be created so that opportunities and challenges will be widely shared in the science and technology community. With the success of an initial workshop hosted by the Alan Turing Institute, the proposed project is a logical expansion of the effort in to Japan and Asia region as well a s developing minimal viable prototype to shape concrete technological base. Interaction with ATI project will be maintained througho ut the period and will assume a series of joint events to be organized.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 20, 2021
- Source ID
- N629092112032
Entities
People
- Hiroaki Kitano
Organizations
- Office of Naval Research
- Systems Biology Institute
- United States Navy