http://www.asimovinstitute.org/neural-network-zoo/
- differentiation equations
- calculus
- python
removes the necessary for experts in the field for medical imaging like ultrasound, mamography, x-ray, MRI, PET scan
removes tumor, cancer cells, disease diagnosis
ucla staff machine learning facial recognition
overfitting - when your train model describe just your data set and it needs more data to des to describe the edge cases
CNNs convolution neural networks
universal approximator - 2 layer neural network with a finite number of hidden units can approximate any continuos function
deep architecture
for medical imaging you need to be able to work from small data sets
affine transformation Computer vision
- Greedy layerwise learning
- train the first stack
- leave the rest of the layer out,
- let it learn
- output its learning
- then add the next layer
Zoom - record talks
- Deep belief Network
- Deep Boltzmann Machine
- feature map
- convolution layer
- pooling layer
better accuracy with less parameters
- supervised learning unsupervised - let it create patterns reward - make it learn like humans do
GANS
- https://skymind.ai/wiki/generative-adversarial-network-gan
- https://www.google.com/search?q=gans&rlz=1C5CHFA_enUS818US818&source=lnms&tbm=vid&sa=X&ved=0ahUKEwif25bkt_rfAhWHHXwKHexmBHAQ_AUIECgD&biw=1022&bih=768
- art
- fashion mnist + POET Machine learning
https://jetware.io/versions/stanford-cs20-course:2018
weights and biases
- https://datascience.stackexchange.com/questions/19099/what-is-weight-and-bias-in-deep-learning
- https://www.wandb.com/
- https://www.youtube.com/results?search_query=Weights+and+Biases
For basics of neural network check this: http://www.asimovinstitute.org/neural-network-zoo/
- Human few-shot learning of compositinal instructions
- Attentive Neural Processes
- OPtimization Models for Machine Learning: A Survey
ELIR
ICLR
Artifial intelligence and Industry Open Gym AGI - Ben Gordsel
Android - podcast addict
Lex Freedman