Carboncopies Research¶
At Carboncopies, we're conducting various types of research. Please see the list below and feel free to get involved!
Brain Emulation Challenge¶
About¶
The Brain Emulation Challenge will rapidly accelerate progress towards WBE through a series of challenges and prizes designed to focus, quantify, and grow the structure-to-function field. Through rigorous validation metrics and performance scores, we can determine what methods work best and provide debugging feedback where they go wrong.
Impact¶
Conversion of brain data to working emulations (operational models that satisfy WBE criteria) is almost non-existent. This ‘Translation’ from data to model parameters is the final step to end-to-end examples of brain emulation. End-to-end examples will cause explosive growth in the field, as more people see the real and immense potential of the technology. Unfortunately, Translation is very hard, because evaluating how well a Translation method works is very difficult without a known correct result, i.e. without fully known ground-truth.
Therefore, we propose the WBE challenge - directly addressing these issues. Challenge data sets with fully understood images (e.g. ImageNet) were critical to the rapid development of computer vision. Our generated virtual brain samples containing a fully understood functional neuronal network allow us to apply validation metrics based on rigorous success criteria. With that, we can score the performance of an attempted translation from virtual brain data to a running emulation, and provide precise feedback about errors made.
Our challenge will encourage friendly standardized academic competition within the field, driving innovation by setting clearly defined goals and rewards. With each iteration of the challenge, we create more-realistic and complex datasets that build participants up to the real-deal. This step-wise approach would allow participants to systematically debug and improve their methods, further accelerating their development.
BrainGenix Platform¶
About¶
The BrainGenix initiative exists to create and maintain a comprehensive software platform specifically designed to enable and accelerate WBE research.
Impact¶
Currently there are no existing platforms designed to address the interdisciplinary challenges presented by Whole Brain Emulation research. The BrainGenix platform seeks to change that by providing extensive libraries, tools, and other software needed to rapidly test and develop different aspects of WBE research.
In our BrainGenix group, software of the Generative Meta-Analysis platform is being developed to support the WBE Challenge, which requires both high-performance simulation of large-scale compartmental neuronal networks and a highly-realistic virtual brain data acquisition system. Many of these tasks are computationally non-trivial, which only adds to the difficulty in conducting WBE research.
We seek to change that by taking the computational complexity out of the equation - our platform focuses on being fast and reliable, removing the need for researchers to worry about making code scale.
Ethics Framework¶
Whole Brain Emulation (WBE) is the theoretical process of scanning, modeling, and reproducing the entire functional organization of a human brain in a computational substrate, allowing for the recreation of cognition, memory, and potentially consciousness. If realized, WBE would represent a historic leap in neuroscience, artificial intelligence, and human self-understanding. However, such potential comes with profound ethical questions. The development and application of WBE involves intimate engagement with the human mind — raising urgent concerns about identity, consent, consciousness, suffering, personhood, and social justice.
The Ethics Framework project exists to ensure that the advancement of WBE technology is aligned with human dignity, scientific integrity, and global justice. It provides a structured set of principles and guidelines for ethical research, responsible deployment, and inclusive governance of WBE systems.
Early insights from work on an Ethics Framework have shown that the presuppositions upon which such a framework rests are based in fundamental core questions about whole brain emulation of a metaphysical nature. Questions such as:
- What is consciousness, and what sorts of intelligent structures can manifest consciousness?
- How is it that subjective experience comes about, that there can be an inner world of awareness?
- What is the nature of personal identity? How is that affected by strange new possibilities that whole brain emulation can bring about?
These questions are not at all trivial, especially, from the point of view of the ethical consequences and impacts of specific choices in the development of whole brain emulation. Of course these questions are relevant in many ways beyond WBE, for example, in artificial intelligence, where - as in the case of non-human animals - we must understand when one should presume that conscious experience, and an accompanying sense of well being vs suffering can and may exist. Furthermore, when whole brain emulatin technologies are applied medically, offering future patients a procedure that results in living with non-biological (possibly digital) brain, one seeks high confidence that a neuroprosthetic brain provides not only consciousness, but also the inner experience of subjective awareness that defines what it is to be human.
For this reason, the Ethics Framework project group is partnering with our Consciousness & Subjective Experience journal club collaboration to carefully explore leading and most parsimonious theoretical explorations in these fields.
Whole Brain Emulation Roadmap¶
Our Whole Brain Emulation (WBE) Roadmap update group is a collaborative partnership between the Carboncopies Foundation and experts at universities in the US and Europe, and is working on an update of a forward-looking roadmap towards whole brain emulation. The first WBE roadmap manuscript was published in 2008 by the Future of Humanity Institute, Oxford University, edited by Dr. Anders Sandberg and Prof. Nick Bostrom. It was a compilation of articles contributed in a follow-up report to the First Whole Brain Emulation Workshop at Oxford University, by workshop participants John Fiala, Robin Hanson, Kenneth Jeffrey Hayworth, Todd Huffman, Eugene Leitl, Bruce McCormick, Ralph Merkle, Toby Ord, Peter Passaro, Nick Shackel, Randal A. Koene, Robert A. Freitas Jr and Rebecca Roache. You can download the 2008 manuscript here: https://www.fhi.ox.ac.uk/brain-emulation-roadmap-report.pdf
Many technical developments and advances in understanding have come along in the years since the 2008 publication. The WBE Roadmap update group aims to publish a significantly revised, updated, and for some topics, extended roadmap. This update requires input from leading experts in numerous fields, insights gathered through a series of workshops, and the selective publication of individual sections in peer reviewed academic journals. During this process, the WBE Roadmap update group will make available on our affiliate site wholebrainemulation.org the pre-print manuscripts of roadmap sections seeking public comment, as well as finalized and/or published sections.