qq_41103204的博客

NLP，Transformer，Bert

Deep Dream里面能够改变的参数不是filter，而是input image，通过调整input image，来使得output image as high as possible (excites the layer)大部分的参数来自于最后一层dense layer，通过global average pooling可以极大减少参数数量。

ContentsLecture 12: Lexical Semantics (part I) - Word Representations and Word Embeddings.Lecture 13: Machine Learning: Linear and Log-Linear ModelsLecture 14: Machine Learning: Feed-forward Neural Networks, Autoencoders/embeddings, Dense networksAutoencod

ContentsComputer systems and performance rules of thumbData centersDatabasesStorage and distributed file systemsDistributed systemsMapreduceDistributed analytics and streamingCachingAdditional informationComputer systems and performance rules of thumbLa

ContentsMultivariate GaussiansReview: Variance and covarianceMultivariate GaussiansPCA/ SVDMultivariate GaussiansReview: Variance and covarianceCovariance matrix: X⃗=(X1,X2,…,Xd)\vec{X} = (X_1, X_2, \dots, X_d)X=(X1​,X2​,…,Xd​), cov(X⃗)cov(\vec{X})cov

ContentsTranspose and Inverse(AB)−1=B−1A−1(AB)^{-1} = B^{-1} A^{-1} (AB)−1=B−1A−1(AT)−1=(A−1)T(A^T)^{-1} = (A^{-1})^{T}(AT)−1=(A−1)TAA−1=I, (A−1)T(AT)=IAA^{-1} = I,\ (A^{-1})^T(A^T) = I AA−1=I, (A−1)T(AT)=IPermutation matrixP−1=PT P^{-1} =

ContentLecture 1 IntroLecture 2 EstimationEstimation:Two estimation methodsOptimality in estimationLecture 3 Confidence intervals and hypothesis testingLecture 1 IntroLLN (Law of large numbers)CLT (Central Limit Theorem)CMT (Continuous Mapping Theorem

ContentsImportant algorithmsNoteLecture1: Insertion sort, efficient algorithmLecture2: Merge sortLecture3: Binary search, quicksortLecture5: Graphs, Breadth-First Search (BFS)Lecture6: Depth-first search, topological sortingLecture7&8: Strongly connect