관심있는것들

before this lecture we assumed prior distribution to be simple gaussian , but now we want something more complicate so we can map more complex manifold.it is hard to evaluate all the possible z, so in VAE we used encoder to find distribution z using q.instead of passing the z through these 2 NN and then sampling from a guassian, we directly transform the latent variable by applying a single dete..

1st piece is average log probability when both x part ,z part are observed.when z is infered from q model distribution.2nd piece is function of q and doesn't depend on our generative modelif q is chosen as posterior distribution z given x under our generative model, now there are no gap between inequalityposterior is hard to computeinvert neural network in probabilistic way which is very hard be..

latent variable = doesn't have notation , what we observe is just a pixel but it might be eye color, hair color something like thatsince it is hard to notate condition for latent variable, we use NNin this case z is not infitie, it is discrete and limitedin this case Z is categorical distirbution , in this case K was 2 and model was able to cluster datasz is now continous , having guassian distr..

using rnn with maksing to act like autoregressive model (looking back only , ordering conditional probabiltily)cnn with masked to be autoregressive, this masked convolution causes blind spot and need to be taken care of.one way is to use multiple convolution with different maskingPixelDefend can detect adversarial examplesBPD(X) ~= − log pCNN(X)/(I × J × K × log 2) for an image of resolution I ×..

위 3가지 조건을 만족하는 model distribution을 찾고 싶다.in order to use chain rule you need to pick order, and there is no ground rule for this.seperate probability distribution using chain rule.first one is conditional probability table(CPT) because it is the very first pixel and doesn't rely on anythingfrom second pixel we use logisitic regressionautoregressive = trying to predict parts of each data point, g..

represent 하는 state는 2^n 개이지만 joint pdf 를 나타내기위한 필요한 parameter 개수는 2^n-1 이다.마지막 1개는 1- (the other parameter's probabilty)를 해서 계산할 수 있으니까but when assuming RV are indepdent variable. we can calculate joint distribution by mulitpling each pdand then we only need N parameters , since we know how to calculate joint distribution.before example we didn't know how to calculate joint distrubiton so we lit..