Hi Uiux,
Was it only yesterday that I posted this drawing from WO2020092691 which shows Akida's STDP circuit?
As I discussed, the weights in the weight register 106 can be positive or negative. As the USPTO examiner's report points out STDP depends on the timing of the spikes relative to the output spike - the prior bits being positive, and the posterior bits being negative.
Ths drawing should be burned into the neurons of all AKIDA true believers.
Allowed: Neural processor based accelerator system and method, page-15
-
- There are more pages in this discussion • 62 more messages in this thread...
You’re viewing a single post only. To view the entire thread just sign in or Join Now (FREE)
Add BRN (ASX) to my watchlist
 (20min delay)
|
|||||
|
Last
16.3¢ |
Change
-0.003(1.52%) |
Mkt cap ! $313.9M | |||
| Open | High | Low | Value | Volume |
| 16.5¢ | 16.5¢ | 16.0¢ | $410.6K | 2.498M |
Buyers (Bids)
| No. | Vol. | Price($) |
|---|---|---|
| 76 | 2884003 | 16.0¢ |
Sellers (Offers)
| Price($) | Vol. | No. |
|---|---|---|
| 16.5¢ | 112387 | 8 |
View Market Depth
| Last trade - 12.30pm 18/09/2024 (20 minute delay) ? |
| BRN (ASX) Chart |