Re: [agi] Self-maintaining Architecture first for AI
2008/5/11 Russell Wallace [EMAIL PROTECTED]: On Sat, May 10, 2008 at 10:10 PM, William Pearson [EMAIL PROTECTED] wrote: It depends on the system you are designing on. I think you can easily create as many types of sand box as you want in programming language E (1) for example. If the principle of least authority (2) is embedded in the system, then you shouldn't have any problems. Sure, I'm talking about much lower-level concepts though. For example, on a system with 8 gigabytes of memory, a candidate program has computed a 5 gigabyte string. For its next operation, it appends that string to itself, thereby crashing the VM due to running out of memory. How _exactly_ do you prevent this from happening (while meeting all the other requirements for an AI platform)? It's a trickier problem than it sounds like it ought to be. I'm starting to mod qemu (it is not a straightforward process) to add capabilities. The VM will have a set amount of memory and if a location outside this memory is referenced, it will throw a page fault inside the VM, not crash it directly. The system will be able to deal with it how it wants to, something smarter than, Oh no I have done a bad memory reference, I must stop all my work and lose everything!!! Hopefully. In the greater scheme of things the model that a computer has unlimited virtual memory has to go as well. Else you might get important things on the hard disk and have much thrashing and ephemera in main memory. You could still make high level abstractions but the virtual memory one is not the one to display to the low level programs. Will Pearson --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: Newcomb's Paradox (was Re: [agi] Goal Driven Systems and AI Dangers)
On Sun, May 11, 2008 at 4:06 AM, Matt Mahoney [EMAIL PROTECTED] wrote: Yes, but in this case the input to P is not (P,y), it is a self reference to whatever program P is running plus y. It's irrelevant, because description of P (or Q) could've been contained in the prefix that said simulate this on yourself: , and it could've been handled by the same machinery that in my example was printing the output is . The only problem is brackets, so if the description is always [finite prefix with machine specification]+[data parameters], it will work. I think you would agree that a virtual machine with n bits of memory can only be implemented on a machine with more than n bits of memory. If it needs to, it can reserve a finite number of additional bits just for this purpose. -- Vladimir Nesov [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] Self-maintaining Architecture first for AI
On Sun, May 11, 2008 at 7:45 AM, William Pearson [EMAIL PROTECTED] wrote: I'm starting to mod qemu (it is not a straightforward process) to add capabilities. So if I understand correctly, you're proposing to sandbox candidate programs by running them in their own virtual PC, with their own operating system instance? I assume this works recursively, so a qemu-sandboxed program can itself run qemu? --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: Newcomb's Paradox (was Re: [agi] Goal Driven Systems and AI Dangers)
--- Vladimir Nesov [EMAIL PROTECTED] wrote: On Sun, May 11, 2008 at 4:06 AM, Matt Mahoney [EMAIL PROTECTED] wrote: Yes, but in this case the input to P is not (P,y), it is a self reference to whatever program P is running plus y. It's irrelevant, because description of P (or Q) could've been contained in the prefix that said simulate this on yourself: , and it could've been handled by the same machinery that in my example was printing the output is . The only problem is brackets, so if the description is always [finite prefix with machine specification]+[data parameters], it will work. If a machine P could simulate two other machines Q and R (each with n bits of memory), then P needs n+1 bits, n to reproduce all the states of Q or R, and 1 to remember which machine it is simulating. You described a machine P that can simulate only P. -- Matt Mahoney, [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: Newcomb's Paradox (was Re: [agi] Goal Driven Systems and AI Dangers)
On Sun, May 11, 2008 at 8:57 PM, Matt Mahoney [EMAIL PROTECTED] wrote: If a machine P could simulate two other machines Q and R (each with n bits of memory), then P needs n+1 bits, n to reproduce all the states of Q or R, and 1 to remember which machine it is simulating. You described a machine P that can simulate only P. You are being obscure again. A machine that I described can simulate any machine of limited size and number of states. Basically, it is a universal machine that can run any other machine, one of which is P, the machine itself, but other machines are allowed too. Each machine can initiate loading of another machine on the underlying universal machine. -- Vladimir Nesov [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: Newcomb's Paradox (was Re: [agi] Goal Driven Systems and AI Dangers)
--- Vladimir Nesov [EMAIL PROTECTED] wrote: On Sun, May 11, 2008 at 8:57 PM, Matt Mahoney [EMAIL PROTECTED] wrote: If a machine P could simulate two other machines Q and R (each with n bits of memory), then P needs n+1 bits, n to reproduce all the states of Q or R, and 1 to remember which machine it is simulating. You described a machine P that can simulate only P. You are being obscure again. A machine that I described can simulate any machine of limited size and number of states. Basically, it is a universal machine that can run any other machine, one of which is P, the machine itself, but other machines are allowed too. Each machine can initiate loading of another machine on the underlying universal machine. You have only shown that P can simulate itself with no additional memory. If it simulates any other machine, then you need to show that no additional memory is needed for those machines too. Also, it doesn't make sense to talk about universal finite state machines, because there are only a finite number of unique programs it can simulate. Instead we can define a machine as general purpose if it can accept 2 or more program specifications (which can be as small as 1 bit). So let me restate my claim. We say that P simulates Q if for all x, P((Q,x)) = Q(x). We say P is general purpose if there exists Q and R such that P simulates both, and there is an x such that Q(x) != R(x) (i.e. Q and R are different programs). Then I claim there is no general purpose finite state machine P that can simulate itself. -- Matt Mahoney, [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com