Artificial Intelligence. Reimagined.
A new era in the field of Artificial Intelligence with KE's SNN(Sieve Neural Network), a non-recursive neural network algorithm.
All the advantages of a Deep Neural Network with none of its disadvantaes. No need for complicated tools like TensorFlow or expensive GPUs.
Addressing an AI problem, talk to us. Sieve Neural Network helps solve it in Days, what otherwise took endless months. Contact us below:
Back Propagation is slow! The human mind does not do back propagation. The current AI algorithms take too long to train and take too much of compute power.
Why KE's SNN ?
Closer approximation of how the mind works
We believe, the SNN is the correct representation of the how the mind works. In fact, we have mathematical proof that our representation is the most efficient representation.
One algortihm to rule them all
Image Recognition, speech recognition, NLP or any other domain, SNN adapts itself and works without any change. No need to do trial and error with weights or hidden layers. SNN always finds the best architecture itself.
You can add data to the SNN and it learns incrementally. There is no need to run the training process on the whole data again.
Fast. With no special GPUs
SNN is the fastest learning algorithm(O(NlogN) time complexity). There are cases where we have been able to get the training time down from days to seconds.
As simple as 1,2,3...
Single Ke's SNN Algorithm applied across domains
with no change to the base algorithm.
Extended Yale DB
Time to train: 40 secs
Time to train: 90 secs
Time to train: 2 secs
Human Activity Detection Dataset
Time to train: 3 secs
Google Speech Dataset
Time to train: 60 secs
Google Landmark Dataset
Time to train: 120 secs
CIFAR-10 - Object Recognition in Images
Time to train: 80 secs