Edit: never mind, I thought she was talking about a real machine until I read the paper. Even the Turing machine model assumes infinite storage space. So there must be more to her "super-turing" machine than just infinite storage.
I wonder if you could just build a ARNN and "fudge" the infinite precision with a reasonably large precision? Or with a big disk, you could compute and unreasonable amount of precision :) Or, store the infinite precision in a lazy way that only calculates as much as you need for a particular answer.
The thing is that you could do all that on a Turing machine, so any model just "fudging" infinite precision would be equivalent in power to a Turing machine.
Comments
Edit: never mind, I thought she was talking about a real machine until I read the paper. Even the Turing machine model assumes infinite storage space. So there must be more to her "super-turing" machine than just infinite storage.
I wonder if you could just build a ARNN and "fudge" the infinite precision with a reasonably large precision? Or with a big disk, you could compute and unreasonable amount of precision :) Or, store the infinite precision in a lazy way that only calculates as much as you need for a particular answer.
The thing is that you could do all that on a Turing machine, so any model just "fudging" infinite precision would be equivalent in power to a Turing machine.