The Nobel Turing Challenge is a grand challenge aiming at developing a highly autonomous AI and robotics system that can make major scientific discoveries, some which may be worthy of the Nobel Prize and even beyond. Accomplishing this challenge requires a development of a series of technologies and in-depth understanding on the process of scientific discoveries. From the system development perspective the challenge is to make a closed-loop system from knowledge acquisition and hypothesis generation and verification to full automation of experiments and data analytics.
This series of workshops aims to discuss a process of scientific discovery and how AI can make it happen highly autonomously to yield high impact discoveries. The first workshop will focus on sharing the current status of the field, and discuss what are possible high impact targets.
Vast scientific knowledge is recorded in publications described using natural language that is at best informal and less precise form of communication. The process of how knowledge can be acquired from publications, databases, and other sources and how hypotheses can be generated and tested. Not only does it lack precision due to the inherent nature of human language behaviors, it may contain errors, fabrications, and other issues that undermine the integrity of the body of knowledge. A series of challenges is envisioned that are:
- Extraction of knowledge from a body of publications
- Accuracy evaluation and maintenance of integrity while allowing options for revision
- Generation of hypotheses from the body of knowledge
- Generation of reasoning to support and refute generated hypothesis
- Formation of potentially multiple self-consistent system of knowledge that can be resolved by specific series of tests
On the automation front, development of fully connected laboratory and data analysis pipelines are essential. How can this be linked with hypothesis generation? How automation of experiments improve precision of the experiments, as well as accelerate the speed of discoveries? How can we combine a variety of experimental devices and enable flexible protocol generation and execution?
This is not just automation of laboratories. It is the automation of science.
This workshop is designed to be a small and intense forum of those who have a vision and experiences on the related subjects. We are expecting that this workshop will be the stepping stone for creation of the larger and more comprehensive forum of the field.
Date: 21st and 22nd, April 2022
Registration: Registration is required to attend this event. Registration for participation deadline is 20th April 2022.
After you register, the zoom link to join this workshop will be informed by email. If you need technical assistance for Zoom operation, the operation committee will assist you individually. Please inform us in advance at iosu-workshop@oist.jp.
See the workshop program and a list of the speakers here.