Home
About WBAI
Annual Policies

WBAI Policy, FY2023

Introduction

Recent advances in artificial intelligence are particularly evident in large-scale generative models. The progress has increased the likelihood of early realization of artificial general intelligence (AGI) via ways that do not necessarily refer to the brain. Meanwhile, it is also facilitating our brain reference architecture (BRA)-driven development to build brain-inspired AGI. For example, it enables automated construction of brain information flows (BIFs) that reflect anatomical structures found in neuroscience papers and other sources.

The rapid technological progress in AI is stimulating the research field of AI alignment, which aims to align AI with humanity. As we have been pointing out, one of the advantages of brain-inspired AGI constructed with the WBA approach is “high affinity with humans,” and we believe that the significance of AGI referring to the brain, constructed to pursue survival as life, will increase.

In this light, we plan to continue our educational and R&D businesses in conjunction with BRA-driven development. As 2023 is the midpoint of the project, the technological roadmap will be reconstructed.

The following is the policy for our education and R&D businesses for the fiscal year 2023.

Education Business

The objective of the education business is to develop and increase the number of interdisciplinary personnel with different expertise in AI/ML, neuroscience, and cognitive science necessary for R&D in the WBA approach over the long-term. We continue our regular activities of holding seminars and an annual symposium to raise public interest.

We shall promote the progress of our organization by participating in and cooperating with external academic events and by cooperating and exchanging information with external academic organizations. In addition, we will also foster and involve human resources through cooperation with external researchers, engineers, and groups. For example, we shall continue our HCD data submission project with the Neurocomputing Group of the Institute of Electronics, Information and Communication Engineers (IEICE).

We shall invite applications to the WBAI Incentive Award to recognize those who have made significant contributions to the promotion of brain-inspired AGI technology, thereby invigorating the community. We also continue our efforts to make the brain-inspired AGI beneficial.

Research and Development Business

The R&D business aims to accelerate democratic AGI research via the WBA approach on open platforms by taking advantage of our being a non-profit organization to avoid competition with other research organizations in AGI research, and by promoting BRA-driven development.  The following activities will be carried out for FY2023.

  • Design: 
    • BIF creation: Automation is planned as BIF credibility evaluation will be on a large scale for the whole brain connectomes.
    • HCD design
  • Implementation: 
    • Architecture implementation: to implement architecture referring to HCDs on a software framework with structural fidelity evaluation
    • Component implementation: to implement software components referring to the process description in HCDs with activity reproducibility evaluation
    • Test environment development: to build test environments (with Gym, etc.) to evaluate the implemented architecture
  • Evaluation: 
    • HCD review flow: a platform and automated review
    • Structural fidelity evaluation: to evaluate the similarity of the implemented code structure with anatomical structure
    • Execution environment development: to develop environments in which brain-inspired software carries out tasks and is tested
    • Fidelity evaluation: activity reproducibility evaluation of functions/performance (and anomaly)
  • Operation: 
    • BRA data management system: construction, operation, and maintenance of the BRA database, submission review flow platform, information publication portal, automated review and data version management tools, etc.
    • BRA visualization
    • WholeBRA development: to review and register items in the WholeBIF, and to create the integrated HCD
  • Methodology: 
    • Research on methodologies for BRA-driven development
  • Other: 
    • Design of new functions: to design unknown computational mechanisms that can effectively solve specific tasks that can be solved by humans (and other animals)
    • Verification of new functions: to tentatively implement unknown computational mechanisms that can effectively solve tasks that can be solved by humans (and other animals)

Building Open R&D Communities

We continue to develop software and other research infrastructure to support research through the WBA approach, and to promote research using the infrastructure in collaboration with other organizations. In FY2023, we plan to make presentations at domestic conferences on the Whole Brain Probabilistic Generative Model (WB-PGM), which was proposed in collaboration with research organizations including Ritsumeikan University, OIST, the University of Electro-Communications, Tokyo Women’s Medical University, Tamagawa University, and the University of Tokyo.

In order to construct BRA useful for the development of brain-inspired software, we promote the construction of brain information flows (BIFs) that represent anatomical knowledge, and the assignment of functional hypotheses to BIFs for tasks and functions. In order to evaluate the implementation, we also develop techniques to evaluate the neuroscientific validity of brain-inspired AI with BRA.

Activities to align brain-inspired AGI with Humanity

A practical advantage of brain-inspired AGI is its ability to take advantage of its affinity with humans. This will be especially true in areas where interaction with humans is necessary. It will be useful when humans interact with AI that has different thought patterns and behaviors than humans. For example, it could serve as a simulator to deepen AI’s understanding of humans, to promote the significance of the human species, and to preserve human values and culture. Furthermore, it is expected to be valuable for creating AI, which is not only a tool, but also an autonomous entity or an ethical subject.

The 8th WBA Symposium (to be held in Japanese) will feature a panel discussion on the theme “the development on the whole brain architecture approach in the age of generative AI and its significance.” We also plan to utilize large-scale language models and the HCD of the amygdala-based fear and motivation mechanisms to achieve emotionally rich interactions. We continue to discuss the alignment issues in our roadmap as well.

Budget for the fiscal year

The planned income is approximately 0.89 million yen, which includes 0.69 million yen in membership fee income as well as fees from seminars (the membership fees were 1.67 million yen in FY 2022). The total of the planned income for the current fiscal year and the 8.31 million yen carried over from the previous fiscal year is approximately 9.2 million yen.

Expenditures will be approximately 1.89 million yen in total, including approximately 0.81 million yen for administrative expenses and approximately 1.08 million yen for operating expenses, including event expenses and R&D expenses (in relation to the planned income for the current fiscal year, there will be a deficit of 1.0 million yen). In the FY2022 budget, we planned to spend about 2.78 million yen in total, including about 0.83 million yen for administrative expenses and about 1.97 million yen for operating expenses.

Conclusion

From its inception in 2015, WBAI has been aiming to develop AGI by 2030, based on the Whole Brain Architecture approach.  Though, at the time of the founding, this seemed an ambitious goal, AI technology has advanced rapidly in the years since. Now, we can no longer ignore the possibility that AGI will be realized before 2030 through ways that do not refer to the brain.

On one hand, we have proposed a methodology with brain reference architecture, namely BRA-driven development. After overcoming the initial period of confusion, we have made steady progress in realizing brain-inspired AI, and our research collaboration with the fields of neuroscience and AI is expanding. The completion of the Whole Brain Reference Architecture (WBRA), the BRA of the entire brain, is an important milestone in this development process, and its construction is being accelerated by the use of large-scale generative models. Since other approaches to AGI construction may not necessarily be successful, there is still a possibility that the completion of AGI construction from our BRA-driven development approach will take precedence.

On the other hand, research on AI alignment to reconcile AI and human values is rapidly gaining momentum in order to avoid the great risks posed to humanity by AGI and the superintelligence that will emerge as a result of its development. Brain-inspired AGI, built through BRA-driven development, produces superintelligence with a high affinity for humans. It will be useful in a variety of ways when humans confront AI that thinks and behaves differently from humans. For example, it could serve as a simulator to help AI understand humans, to emphasize the significance of the human species, and to preserve human values and culture. More broadly, it may serve as a fundamental technology to produce AI as an autonomous, rather than instrumental, life form or as an ethical entity. Therefore, even if AGI and superintelligence are realized first by means other than the WBA approach, these roles played by brain-inspired AGI could continue to be of great value to humanity.

Considering the accelerating progress of AI technology, it becomes even more difficult to predict the future. However, as brain-inspired AGI can play a valuable role in the future of humanity, we work on a technology roadmap, in which 2023 is marked as the midpoint milestone of the WBA project, to effectively utilize the BRA-driven development methodology and WBRA as brain-inspired software specification.

WBAI continues to promote educational activities, strengthen R&D activities, and establish a new technology roadmap, aiming to link our technologies and achievements to activities that will have a beneficial impact on the future of humanity. Your understanding and support are greatly appreciated.