Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail From: Bill Sloman Newsgroups: sci.electronics.design Subject: Re: yes! Date: Sun, 18 Aug 2024 01:56:32 +1000 Organization: A noiseless patient Spider Lines: 66 Message-ID: References: <2ugqbjhvlh9vrlmqhciaubcf64dbooph0o@4ax.com> <06jvbjp36khao0m5ot65a1o1krricoasre@4ax.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Injection-Date: Sat, 17 Aug 2024 17:56:34 +0200 (CEST) Injection-Info: dont-email.me; posting-host="77b32acd5a2fc5f6060a7eb8d34ddda6"; logging-data="2065549"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX18p0Rnwj6Uw7dAWknvBNOHzZOR2xkZ+1EQ=" User-Agent: Mozilla Thunderbird Cancel-Lock: sha1:hW3/zul8s9GpoRu0ofvyGsSeJsc= X-Antivirus-Status: Clean Content-Language: en-US In-Reply-To: X-Antivirus: Norton (VPS 240817-10, 17/8/2024), Outbound message Bytes: 3723 On 18/08/2024 12:00 am, john larkin wrote: > On Sat, 17 Aug 2024 06:26:27 GMT, Jan Panteltje > wrote: > >> On a sunny day (Fri, 16 Aug 2024 15:07:52 -0700) it happened john larkin >> wrote in <06jvbjp36khao0m5ot65a1o1krricoasre@4ax.com>: >> >>>>> It's just code. >>>> >>>> Not any more it isn't. >>> >>> Those giant computer networks don't run code? >>> >>>> Your lack of understanding is a handicap. >>> >>> Your lack of imagination ditto. >> >> Well, there is a bit of your lack of understanding > > So you understand how brains work? Nobody does, but it is being worked on. > Where are images stored, and how can one recognize and name one of > maybe a million storted images in a fraction of a second? Functional magnetic resonance imaging does give some insight into where the processing happens as various - short - times after the human brain has been exposed to a stimulus. Not a hell of lot so far, but it is being worked on. >> You know about analog computing >> So big neural networks are basically imitations of the analog brain. >> But you can do a lot in hardware such as storing the 'weights' and vector multiplication. communication. >> My suggestion is for you, just as a free time project, code some neural net. >> Or at least look up how it works: >> https://en.wikipedia.org/wiki/Neural_network >> >> No need for digital at all. >> https://research.ibm.com/projects/analog-ai >> >> OTOH my opinion is that our brain stores memory in RNA and DNA, strong hint is that >> newborn species of many types know how the move, find food, interpret what they see and feel, etc. >> Recent research found that in those neurons some data is stored in such a basic form as RNA, >> >> Nature .. we still invent thing nature already had millions of years ago. > > I've been in several situations where people wanted to use NN's. It > never actually worked. It doesn't make sense. They seem to work in speech recognition https://theaisummer.com/speech-recognition/ I knew a couple of people that worked on that, and lots of neural nets didn't work, but there seems to have been progress on finding variations that can be made to work progressively better. -- Bill Sloman, Sydney -- This email has been checked for viruses by Norton antivirus software. www.norton.com