Talk:Kak neural network

Merge with Subhash Kak

At a glance, from the references, it looks as though Kak Neworks are here to stay. Whether or not they become very useful, they are likely to be mentioned for some time.

Kak's latest suggestion about the Twins Paradox suggests that he may have temporarily gone slightly astray. This is something that happens all too easily when people are investigating relatively uncharted territory.

Given that Kak Networks are mentioned in their own right by other authors, and that this may be a developing area, there is some reason to keep the Kak Networks page and the Subhash Kak page separate.Davy p 18:52, 20 February 2007 (UTC)[reply]

Merge with Instantaneously trained neural networks

Merge with this page would be appropriate. As the Kak networks page seems to have been written mostly anonymously, it's a bit difficult to know how reliable the suggestion is that Kak networks have been used, for example, in finance; or how relevant they are/have been in modellng neurons. Davy p 02:50, 21 February 2007 (UTC)[reply]

Put the best of both into [Neural Network]

It's hard to see why for the majority of non-specialist readers there is any need for separate pages. Davy p 02:51, 22 February 2007 (UTC)[reply]

The "Kak Neural Network" is one of the tens of thousands of meaningless papers that are published in computer science each year. They are unrefereed. A simple google search revealed to me that the handful of authors who publish (in mostly unrefereed places) are Mr. Kak's own students. Try the keyword "Pritam Rajagopal" yourself.

Instantaneous Neural Networks was a stub until the IP address 68.159.27.135 (which traces back to Kak's university of Louisiana and also contributed to wikipedia entry on Kashmir, kak's home state) rather unceremoniously added an unpublished "research" paper again written by Kak's student "Abhilash Ponnath".

http://www.lsu.com/paws005.nsf/$unique/aponna1?OpenDocument

If you merge it with Neural Networks it will get deleted within days if not weeks. Just try it out yourself.

This puerile wikipedia entry should be merged with Subhash Kak.

Gobalan Achayan 18:47, 22 February 2007 (UTC)[reply]

Yes, in the main I agree. Perhaps naively I'd assumed that a request for comment meant just that; then realised I was having a 'conversation' with myself. After reading (thinly) two of the papers mentioned on the Kak networks page and a couple on the Instantaneous neural nets page I reckoned that there is scope for maybe a couple of lines or so on the neural nets page; though I couldn't see quite where to put them. It seems like a neat offshoot, even if is is of limited application.

Shouldn't the two pages on instantaneous nets perhaps now be listed for deletion? Isn't the description of Kak/instantaneous networks rather too detailed for the biog page? It's possibly interesting for anyone learning about neural nets, but only very brief mention is needed in the biog. Davy p 00:34, 23 February 2007 (UTC)[reply]

On second thoughts, isn't it appropriate to revert Instantaneous Nerual Neworks back to the original stub, which had been meaningful, delete Kak neworks and put a mention of them onto Kak's page with the reference to Instantaneous Neural Nets. (Or was this how things were before the self publicity started?) Davy p 00:39, 23 February 2007 (UTC)[reply]