Главная
Study mode:
on
1
Statistical Physics Methods in Machine Learning
2
Deep Learning Applications
3
Tutorial: Deep Learning
4
Outline
5
Learning an unknown function
6
Learning an unknown function: like curve fitting
7
Learning a function: why?
8
Learning a function: How
9
Linear Regression: Line fitting
10
Minimize errorloss in prediction
11
Loss measures error in prediction
12
Gradient descent
13
Learning a function: Linear Regression x
14
Gradient update: BackPropagation.
15
Stochastic Gradient Descent: gradients over a few examples at a time.
16
Learning a function: Sigmoid, sign
17
Sigmoid, RELU
18
Logistic regression uses logloss
19
Deep Network. Allows rich representation Can express any function/circuit
20
Neurons
21
Network of Neurons
22
Hierarchical representation of Objects
23
Training w: SGD to Minimize loss
24
Backpropagation: Gradient Descent for one example
25
Softmax for multiclass output: just like max
26
Convergence of Gradient Descent for Model training
27
Applications
28
MNIST
29
Convolution and Pooling
30
Gradient-Based Learning Applied to Document Recognition
31
Goal
32
ImageNet
33
ILSVRC
34
Architecture
35
RELU Nonlinearity
36
96 Convolutional Kernels
37
Phone recognition on the TIMIT benchmark Mohamed, Dahl, & Hinton,
38
Word error rates from MSR, IBM, & Google Hinton et. al. IEEE signal Processing Magazine, Nov 2012
39
Speech recognition
40
RNN
41
Videos/tutorials on Deep learning applications
42
Theoretical Understanding? - Deep Learning
43
Nonconvex Optimization
44
Low rank Approximation
45
No local minima in linear networks [Kawaguchi, NIPS 16, Ge et al, ICML 17]
46
Deep Learning
47
Does well experimentally
48
With simplifications, our target functions f are...
49
Overview of Results
50
Q&A
Description:
Explore deep learning applications in this comprehensive lecture from the International Centre for Theoretical Sciences. Delve into the fundamentals of machine learning, including linear regression, gradient descent, and neural networks. Examine advanced topics such as convolutional neural networks, recurrent neural networks, and their applications in image and speech recognition. Gain insights into the theoretical aspects of deep learning, including non-convex optimization and low-rank approximation. Learn about cutting-edge research in the field, including recent findings on local minima in linear networks. Conclude with an overview of experimental results and a Q&A session to deepen your understanding of this rapidly evolving field.

Deep Learning Applications by Rina Panigrahy

International Centre for Theoretical Sciences
Add to list
#Computer Science #Deep Learning #Machine Learning #Artificial Intelligence #Neural Networks #Mathematics #Statistics & Probability #Linear Regression #Gradient Descent #Theoretical Computer Science #Distributed Computing #Distributed Algorithms #Science #Physics #Statistical Physics #Speech Recognition
0:00 / 0:00
1 of 50

Statistical Physics Methods in Machine Learning