Hi RG,
The data Anil is talking about is produced from cameras, microphones etc.
In CNN, the data is, for example, standard 8 bit bytes.
Akida converts this into SNN format.
It is not stored in conventional memory in Akida.
Memory in Akida is distribute adjacent to each neurone to eliminate the "memory bottleneck" of von Neumann computers.
It is used for comparison with threshold values in the neurone memories to recognize a large number of spikes from associated neurones as, eg, a tiger.
BRN Podcast, page-406
-
- There are more pages in this discussion • 525 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
23.5¢ |
Change
0.010(4.44%) |
Mkt cap ! $463.5M | |||
| Open | High | Low | Value | Volume |
| 22.0¢ | 24.5¢ | 22.0¢ | $2.565M | 10.96M |
Buyers (Bids)
| No. | Vol. | Price($) |
|---|---|---|
| 6 | 227996 | 23.5¢ |
Sellers (Offers)
| Price($) | Vol. | No. |
|---|---|---|
| 24.0¢ | 343440 | 7 |
View Market Depth
| Last trade - 16.10pm 31/10/2024 (20 minute delay) ? |
| BRN (ASX) Chart |