Deep Learning(CS7015): Lec 13.5 Some Gory Details

preview_player
Показать описание
lec13mod05
  1. NPTEL-NOC IITM
  2. deep learning
  3. prof.mitesh khapra
  4. iitmadras
  5. nptel deep learning
Рекомендации по теме
Deep Learning(CS7015): Lec 0:08:44

Deep Learning(CS7015): Lec 13.1 Sequence Learning Problems

Deep Learning(CS7015): Lec 0:03:32

Deep Learning(CS7015): Lec 1.4 From Cats to Convolutional Neural Networks

Deep Learning(CS7015): Lec 0:13:21

Deep Learning(CS7015): Lec 5.7 Tips for Adjusting Learning Rate and Momentum

Deep Learning(CS7015): Lec 0:09:48

Deep Learning(CS7015): Lec 13.2 Recurrent Neural Networks

Deep Learning(CS7015): Lec 0:15:44

Deep Learning(CS7015): Lec 9.5 Batch Normalization

Deep Learning(CS7015): Lec 0:09:18

Deep Learning(CS7015): Lec 14.1 Selective Read, Selective Write, Selective Forget

Deep Learning(CS7015): Lec 0:36:36

Deep Learning(CS7015): Lec 10.4 Continuous bag of words model

Deep Learning(CS7015): Lec 0:14:24

Deep Learning(CS7015): Lec 13.3 Backpropagation through time

Deep Learning(CS7015): Lec 0:31:21

Deep Learning(CS7015): Lec 3.4 Learning Parameters: Gradient Descent

Deep Learning(CS7015): Lec 0:21:55

Deep Learning(CS7015): Lec 15.1 Introduction to Encoder Decoder Models

Early Stopping & 0:04:52

Early Stopping & Dropout: Ways to overcome Overfitting

Deep Learning(CS7015): Lec 0:18:45

Deep Learning(CS7015): Lec 5.4 Momentum based Gradient Descent

Deep Learning Part 0:02:55

Deep Learning Part - II (CS7015): Lec 16.9 Bayesian Networks : Formal Semantics

Deep Learning(CS7015): Lec 0:11:31

Deep Learning(CS7015): Lec 15.4 Attention over images

Deep Learning(CS7015): Lec 0:13:44

Deep Learning(CS7015): Lec 1.2 From Spring to Winter of AI

Deep Learning(CS7015): Lec 0:12:45

Deep Learning(CS7015): Lec 4.7 Backpropagation: Computing Gradients w.r.t. Parameters

Deep Learning(CS7015): Lec 0:13:39

Deep Learning(CS7015): Lec 2.2 McCulloch Pitts Neuron, Thresholding Logic

Deep Learning Part 0:14:01

Deep Learning Part - II (CS7015): Lec 18.7 Motivation for Sampling

Deep Learning Part 0:15:40

Deep Learning Part - II (CS7015): Lec 16.3 Can we represent the joint distribution more compactly

Deep Learning(CS7015): Lec 0:15:38

Deep Learning(CS7015): Lec 2.6 Proof of Convergence of Perceptron Learning Algorithm

Deep Learning(CS7015): Lec 0:07:54

Deep Learning(CS7015): Lec 8.10 Ensemble Methods

Deep Learning(CS7015): Lec 0:12:06

Deep Learning(CS7015): Lec 10.2 Distributed Representations of words

Deep Learning(CS7015): Lec 0:13:30

Deep Learning(CS7015): Lec 10.7 Hierarchical softmax

Deep Learning(CS7015): Lec 0:01:31

Deep Learning(CS7015): Lec 8.6 Parameter sharing and tying

Комментарии
Автор

Can someone summarize in simpler words what this lecture clip was about ?
I felt a bit lost, this was tough to follow :(

newbie
Автор

It is funny how he easily says derivative of a vector is tensor....actually, unlike scalars, derivative of a vector is almost certainly never a tensor! That is why we have notion of "Covariant derivatives" for vectors. This is why defining "Tensors" using Matrices is a very bad idea, no one can really appreciate what tensors actually are! An actual mathematician studying differential geom or physicist studying GR etc will jump off from roof after seeing this abuse of "Tensors"
Again, this is not a limitation of the teacher, this is how "Data science and its paraphernalia is" ...mathematically naive at this stage at best!
Another thing which is very peculiar in all ANN lectures i see is the way they convey the simple concept of back propagation in a horrendous way (computation networks etc.), why not teach it with theory of composite functions and total derivatives, chain rule etc.? Dimensions are self correcting, no need to even memorize them! I will even go on to say that teach students with the analogy of continuous functions first! Everything else will be so simple!!!!

paulhowrang