1st Online Workshop on ASIM

ASIM June 5, 2010 Workshop

First Online Workshop on Advancing Substrate Independent Minds

June 5, 2010, Teleplace TeleXLR8, 16:00-20:00CET, 10:00-14:00EST, 07:00-11:00PST

To be held in Teleplace (http://www.teleplace.com/), contact giulio (AT) gmail.com

Advancing Substrate Independent Minds (ASIM) 2010-1 is the first in a series of workshops organized by the ASIM community. Some workshops, including this first one, will be held online through the Teleplace virtual meeting software (http://www.teleplace.com/). Others will be held in physical venues, often as satellite meetings at other events catering to compatible interest groups. As the title suggests, both the community and the associated workshop are strongly objective-oriented and organization-oriented rather than technology specific. This choice is a reflection of the interests of many in the community, who prioritize achieving and advancing the possibilities for substrate independent intelligent minds over the development of a particular technology track for its own sake. Candidate technologies, in this case, must best serve the objective. Candidates include technologies that are relevant to whole brain emulation, neuroprostheses, brain augmentation, cognitive simulation and artificial general intelligence.

This inaugural workshop (in some ways the continuation of previous annual whole brain emulation satellite meetings) necessarily has a foundational character. Results are intended to crystalize the nature of our community and its objectives, and to prepare future efforts.

The following topics will be addressed by speakers in the workshop:

• 1. Fundamentals of ASIM, such as: What are the minimum requirements for the ASIM objective that we can agree upon? What are proper definitions to work with? What are the main groups in terms of technological opinions / approaches? What are the main concrete problems that we face? What is some of the useful knowledge and technology that currently exists?

  2. Introductions to several of the possible technology directions. This includes structural preservation and data acquisition, functional data acquisition, neuroprosthetic devices, brain augmentation and gradual procedures, the usefulness and design of cognitive architectures, the interaction of ASIM topics with artificial general intelligence, and the interaction of ASIM topics with life extension and synthetic biology.

  3. An examination of needs: The networking and public relations needs of ASIM organizations. Needs in terms of basic research. The needs of specific projects. This includes analysis and discussion of relevant cultural and social boundary conditions and possible initiatives which may boost, or on the contrary be detrimental to acceptance of and interest in the ASIM programs.

  4. Discussion of funding and fund raising possibilities. What is it needed for? Which sort of funding is best applied to which need? What are the steps involved in accomplishing funding goals?

  5. Summary of results from the workshop and plans for the next meeting. (Depending on progress we can of course introduce a continuation online workshop ASIM 2010-1.5 on short notice, or a set of email discussion threads.)

PROGRAM

Introduction and Goals

Basic Objectives and Fundamentals of Advancing Substrate Independent Minds

Randal A. Koene

I would like to start off discussion at our first ASIM workshop with a much needed bit of attention to the big picture objectives that form the umbrella for our group discussions and efforts. Essentially, we need to determine some minima and fundamental agreements that we can adhere to and stand for as a group. Beyond this, there are many individual paths.

ASIM in Science, Enterprise and Society

Memetically Engineering Support for Uploading Technology

Giulio Prisco

In order to receive funding for uploading projects, it is important to create popular support for uploading, or at least a widespread interest and a perception of feasibility. Uploading should be seen as a cool science fiction technology that is slowly but steadily becoming reality, with positive images of a sexy and solar future. I will present some (rather unconventional and unPC) ideas on how to engineer such support.

Brain emulation ethics

Anders Sandberg

Are there particular ethical problems or constraints whole brain emulation needs to take into account? In this talk I will outline how brain emulation research intersects with current ethics in animal and human subjects research. One of the key issues is the fundamental philosophy of mind uncertainty: even if the eventual goal is the creation of software minds (or a disproof of their feasibility) early steps may create systems that may have a profoundly unclear mental, and hence moral, status. Given that uncertainty is often more fearful than actual risk, these uncertainties may pose a unique stumbling block for ethical approval. I will discuss various strategies for responsible brain emulation getting around the ethical quagmires.

ASIM: Can we recast the PR nightmare?

Suzanne Gildert

We all know from personal experience how difficult it is to get the message across when our carefully crafted explanation of ASIM is interrupted by the dreaded phrase 'You mean like in The Matrix, right?' No matter how interested we are in bringing our dream to fruition, we ultimately need to convince others that it is a worthwhile and noble venture. In this talk I will play devil's advocate by discussing some of the problems we face when presenting ASIM as both a serious and lucrative avenue of research and business. I will also discuss ways in which we can present ourselves and our organisation to give us the best chance at making the memes and products of ASIM not just accepted, but positively desired by society within the next couple of decades.

Structured and Scientific Exploration of ASIM

Are conceptual as well as technological/computational breakthroughs necessary for progress in brain emulation?

Yoonsuck Choe

In our lab, we are working on high-throughput microscopy technology and computational algorithms for automated reconstruction of the brain microcircuit at the whole brain scale. Eventually, we are aiming to understand the brain function and be able to emulate it in silico. There are three key components in this endeavour: (1) data acquisition technology (microscopy), (2) automated morphological tracing, and (3) theoretical framework for understanding/modeling. Although at first component 1 seemed the hardest, it immediately turned out that component 2 is a real challenge. Since we have not been able to carry out component 2 sufficiently, component 3 is yet out of our sight, but I strongly believe that once we conquer component 2, we will realize that component 3 is the real challenge. I will talk about this, and how thinking about conceptual/theoretical framework at an early stage can help avoid pursuing unproductive directions in stage 1 and 2.

A science of mind uploading

Ken Hayworth

Mind uploading must be put on a sturdy conceptual foundation if it is to be taken seriously by the scientific community; unfortunately its current foundation is as intellectually fragile as can be. Briefly, the current case for mind uploading is 1.) Our conscious self is ‘somehow’ generated by physical processes in the brain, 2.) Any physical process embodied in one substrate can be “simulated” to an arbitrary level of precision in a different substrate, and 3.) Given a sufficiently high-fidelity simulation of a person’s brain one can say that the original conscious self has been “transferred” to the new substrate.

While this argument is logically sound it leaves unanswered extremely basic questions: “What process in the physical brain gives rise to our consciousness?”, “What precision of simulation is necessary to duplicate this process in another substrate?”, “Will such a duplicate process be the ‘original’ person or will it be merely a ‘copy’?”, “If such a duplicate is only a ‘copy’ then will a gradual transfer from one substrate to another ensure that the ‘original’ is now transferred?”  If we cannot provide at least provisional answers to these questions (backed up by current scientific research) then how can we design a coherent strategy of research toward the goal of mind uploading?

In my talk I will attempt to lay out a “science of mind uploading” by starting with the best scientific theory of what the conscious self (that we are trying to transfer) actually is. I will discuss how this computational model of the self is supported within the best cognitive science model of the mind (ACT-R). I will then discuss how ACT-R is probably implemented in the neuronal circuits of the brain. This will allow the question of “fidelity of simulation” to be addressed head on. Of course there may be flaws in this story, but science has always progressed by putting forward hypotheses which are clear enough to be directly tested. A science of mind uploading will be based on the experiments necessary to test these hypotheses.

Projects and Practical Approaches to ASIM

Basic workflow of a typical WBE/ASIM procedure

Eugen Leitl

I'll be quite brief, will cover basic workflow of a typical WBE/ASIM procedure, mention the good things that are available already, and the list of known unknowns (hard stuff; to be done), as well as consider what targets need to be defined so that we can proceed to make ASIM happen.

Strategies for automated mapping and reverse engineering of neural tissue

Peter Passaro

I'll provide a general perspective of the software and recording systems that currently exist for large-scale extraction and analysis of neuronal spike data and other signals (at the single neuron or above level) that are used for information transmission in neural systems. I'll then give an overview of what technologies are required for automatic network (structural and dynamic) map construction, and information flow network reverse engineering using this data, and my perspective on how these could be scaled up for larger networks leading to whole CNS emulation.

To what extent could neural prosthetics allow staged transfer of functions?

(name withheld)

I'll outline one possible approach to achieving substrate independent minds. The approach is progressive replacement of or creation of parallel systems for parts of the nervous system, starting periferally and working inwards. I'll discuss some of the difficulties and some of the implications.

Neuronanorobotics:  A proposed technology for Advancing Substrate Independent Minds

Nuno Martins

Diamondoid medical neuronanorobotics offers the possibility of real-time structural and functional whole-brain monitoring of a living person, thus has the potential to represent the "gold standard" scanning technology for whole brain emulation.  In a preview of thesis work still in progress, this talk intends to: (1) compare scanning requirements with neuronanorobotics theoretical scanning capabilities, (2) estimate and justify the time-frame predicted for having neuronanorobots in clinical trials, and (3) argue why diamond neuronanorobotics are preferable to synthetic biology or other alternatives such as programmable microbes.

Post-workshop write-up: ASIM2010-1 First Online Workshop on Advancing Substrate Independent Minds

Brief video summary: on blip.tv, on youtube, on vimeo