Astera Institute opens applications for projects with budgets up to $2 million, including AI evaluation for biology and new ways to observe brain activity
Astera Institute opens applications for projects with budgets up to $2 million, including AI evaluation for biology and new ways to observe brain activity
The spring 2027 call was announced on November 13; applications close on January 7, 2027. The program funds 12–18 months of work with an annual salary of $125–250 thousand. Funding for laboratory work, staff, and other project expenses is agreed separately.
Astera funds researchers while they are still working out what to build and who would benefit. The work could develop into a company, a nonprofit organization, or a publicly available research tool. From the program's original conception, the institute sought projects that could accelerate other research groups' work. Astera provides between zero and $2 million for project expenses, as agreed for each project. Terms for projects with commercial interests are discussed during selection.
In the current call, Astera lists specific challenges to attract applicants beyond its usual circles; other ideas are also welcome. One challenge is to develop open systems for evaluating AI for biology. Comparing model predictions with reliable data would help researchers choose which model to rely on in their work. The institute seeks systems that provide this dependable basis for comparison and keep pace with new models.
In neuroscience, Astera seeks methods that use light to record rapid changes in electrical voltage in neurons. Projects could improve sensors, optics, or signal processing to observe more cells, deeper in tissue, or for longer periods. The technology will need to be tested experimentally and made usable by other laboratories. In AI, the institute also proposes modeling the activity of interconnected neurons to study dementia and age-related memory impairment.
Another challenge is to represent scientific findings in a form that computer programs can use. This format should help researchers trace where a finding came from and how others have used it. Developers are invited to create these machine-readable findings, test them in an actual study, and assess their usefulness. The project also includes developing tools that other groups can use to present their work in the same format.
The projects admitted to the program in September already include Discourse Graphs, a system for sharing scientific questions, claims, and evidence. It allows researchers to link a claim to a finding that supports or contradicts it, then share that connection with another group while retaining information about the authors and the original work. Findings from subsequent experiments can be added to the same claim.
The basic rule is to share work as it progresses: make data, code, experimental methods, and unsuccessful attempts openly available. Other groups can then repeat an experiment, check a conclusion, or account for a previous unsuccessful attempt when planning their own research. The Astera open science page expresses the principle this way:
Once a result can be useful, it is ready to be shared.