Skip to content

Commit 0c806b2

Browse files
committed
update week 1
1 parent 25ebe5d commit 0c806b2

File tree

8 files changed

+551
-223
lines changed

8 files changed

+551
-223
lines changed

doc/pub/week1/html/week1-bs.html

Lines changed: 58 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -82,6 +82,22 @@
8282
2,
8383
None,
8484
'good-books-with-hands-on-material-and-codes'),
85+
('Twst yourself: Deep learning 1',
86+
2,
87+
None,
88+
'twst-yourself-deep-learning-1'),
89+
('Test yourself: Deep learning 2',
90+
2,
91+
None,
92+
'test-yourself-deep-learning-2'),
93+
('Test yourself: Optimization part',
94+
2,
95+
None,
96+
'test-yourself-optimization-part'),
97+
('Test yourself: Analysis of results',
98+
2,
99+
None,
100+
'test-yourself-analysis-of-results'),
85101
('Types of machine learning',
86102
2,
87103
None,
@@ -358,6 +374,10 @@
358374
<!-- navigation toc: --> <li><a href="#gaussian-processes-and-bayesian-analysis" style="font-size: 80%;">Gaussian processes and Bayesian analysis</a></li>
359375
<!-- navigation toc: --> <li><a href="#hpc-path" style="font-size: 80%;">HPC path</a></li>
360376
<!-- navigation toc: --> <li><a href="#good-books-with-hands-on-material-and-codes" style="font-size: 80%;">Good books with hands-on material and codes</a></li>
377+
<!-- navigation toc: --> <li><a href="#twst-yourself-deep-learning-1" style="font-size: 80%;">Twst yourself: Deep learning 1</a></li>
378+
<!-- navigation toc: --> <li><a href="#test-yourself-deep-learning-2" style="font-size: 80%;">Test yourself: Deep learning 2</a></li>
379+
<!-- navigation toc: --> <li><a href="#test-yourself-optimization-part" style="font-size: 80%;">Test yourself: Optimization part</a></li>
380+
<!-- navigation toc: --> <li><a href="#test-yourself-analysis-of-results" style="font-size: 80%;">Test yourself: Analysis of results</a></li>
361381
<!-- navigation toc: --> <li><a href="#types-of-machine-learning" style="font-size: 80%;">Types of machine learning</a></li>
362382
<!-- navigation toc: --> <li><a href="#main-categories" style="font-size: 80%;">Main categories</a></li>
363383
<!-- navigation toc: --> <li><a href="#the-plethora-of-machine-learning-algorithms-methods" style="font-size: 80%;">The plethora of machine learning algorithms/methods</a></li>
@@ -643,6 +663,44 @@ <h2 id="good-books-with-hands-on-material-and-codes" class="anchor">Good books w
643663
from Goodfellow, Bengio and Courville's text <a href="https://www.deeplearningbook.org/" target="_self">Deep Learning</a>
644664
</p>
645665

666+
<!-- !split -->
667+
<h2 id="twst-yourself-deep-learning-1" class="anchor">Twst yourself: Deep learning 1 </h2>
668+
669+
<ol>
670+
<li> Describe the architecture of a typical feed forward Neural Network (NN).</li>
671+
<li> What is an activation function and discuss the use of an activation function.</li>
672+
<li> Can you name and explain three different types of activation functions?</li>
673+
<li> You are using a deep neural network for a prediction task. After training your model, you notice that it is strongly overfitting the training set and that the performance on the test isn&#8217;t good. What can you do to reduce overfitting?</li>
674+
<li> How would you know if your model is suffering from the problem of exploding Gradients?</li>
675+
<li> Can you name and explain a few hyperparameters used for training a neural network?</li>
676+
</ol>
677+
<!-- !split -->
678+
<h2 id="test-yourself-deep-learning-2" class="anchor">Test yourself: Deep learning 2 </h2>
679+
680+
<ol>
681+
<li> Describe the architecture of a typical Convolutional Neural Network (CNN)</li>
682+
<li> What is the vanishing gradient problem in Neural Networks and how to fix it?</li>
683+
<li> When it comes to training an artificial neural network, what could the reason be for why the cost/loss doesn't decrease in a few epochs?</li>
684+
<li> How does L1/L2 regularization affect a neural network?</li>
685+
<li> What is(are) the advantage(s) of deep learning over traditional methods like linear regression or logistic regression?</li>
686+
</ol>
687+
<!-- !split -->
688+
<h2 id="test-yourself-optimization-part" class="anchor">Test yourself: Optimization part </h2>
689+
690+
<ol>
691+
<li> Which is the basic mathematical root-finding method behind essentially all gradient descent approaches(stochastic and non-stochastic)?</li>
692+
<li> And why don't we use it? Or stated differently, why do we introduce the learning rate as a parameter?</li>
693+
<li> What might happen if you set the momentum hyperparameter too close to 1 (e.g., 0.9999) when using an optimizer for the learning rate?</li>
694+
<li> Why should we use stochastic gradient descent instead of plain gradient descent?</li>
695+
<li> Which parameters would you need to tune when use a stochastic gradient descent approach?</li>
696+
</ol>
697+
<!-- !split -->
698+
<h2 id="test-yourself-analysis-of-results" class="anchor">Test yourself: Analysis of results </h2>
699+
<ol>
700+
<li> How do you assess overfitting and underfitting?</li>
701+
<li> Why do we divide the data in test and train and/or eventually validation sets?</li>
702+
<li> Why would you use resampling methods in the data analysis? Mention some widely popular resampling methods.</li>
703+
</ol>
646704
<!-- !split -->
647705
<h2 id="types-of-machine-learning" class="anchor">Types of machine learning </h2>
648706

doc/pub/week1/html/week1-reveal.html

Lines changed: 46 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -390,6 +390,52 @@ <h2 id="good-books-with-hands-on-material-and-codes">Good books with hands-on ma
390390
</p>
391391
</section>
392392

393+
<section>
394+
<h2 id="twst-yourself-deep-learning-1">Twst yourself: Deep learning 1 </h2>
395+
396+
<ol>
397+
<p><li> Describe the architecture of a typical feed forward Neural Network (NN).</li>
398+
<p><li> What is an activation function and discuss the use of an activation function.</li>
399+
<p><li> Can you name and explain three different types of activation functions?</li>
400+
<p><li> You are using a deep neural network for a prediction task. After training your model, you notice that it is strongly overfitting the training set and that the performance on the test isn&#8217;t good. What can you do to reduce overfitting?</li>
401+
<p><li> How would you know if your model is suffering from the problem of exploding Gradients?</li>
402+
<p><li> Can you name and explain a few hyperparameters used for training a neural network?</li>
403+
</ol>
404+
</section>
405+
406+
<section>
407+
<h2 id="test-yourself-deep-learning-2">Test yourself: Deep learning 2 </h2>
408+
409+
<ol>
410+
<p><li> Describe the architecture of a typical Convolutional Neural Network (CNN)</li>
411+
<p><li> What is the vanishing gradient problem in Neural Networks and how to fix it?</li>
412+
<p><li> When it comes to training an artificial neural network, what could the reason be for why the cost/loss doesn't decrease in a few epochs?</li>
413+
<p><li> How does L1/L2 regularization affect a neural network?</li>
414+
<p><li> What is(are) the advantage(s) of deep learning over traditional methods like linear regression or logistic regression?</li>
415+
</ol>
416+
</section>
417+
418+
<section>
419+
<h2 id="test-yourself-optimization-part">Test yourself: Optimization part </h2>
420+
421+
<ol>
422+
<p><li> Which is the basic mathematical root-finding method behind essentially all gradient descent approaches(stochastic and non-stochastic)?</li>
423+
<p><li> And why don't we use it? Or stated differently, why do we introduce the learning rate as a parameter?</li>
424+
<p><li> What might happen if you set the momentum hyperparameter too close to 1 (e.g., 0.9999) when using an optimizer for the learning rate?</li>
425+
<p><li> Why should we use stochastic gradient descent instead of plain gradient descent?</li>
426+
<p><li> Which parameters would you need to tune when use a stochastic gradient descent approach?</li>
427+
</ol>
428+
</section>
429+
430+
<section>
431+
<h2 id="test-yourself-analysis-of-results">Test yourself: Analysis of results </h2>
432+
<ol>
433+
<p><li> How do you assess overfitting and underfitting?</li>
434+
<p><li> Why do we divide the data in test and train and/or eventually validation sets?</li>
435+
<p><li> Why would you use resampling methods in the data analysis? Mention some widely popular resampling methods.</li>
436+
</ol>
437+
</section>
438+
393439
<section>
394440
<h2 id="types-of-machine-learning">Types of machine learning </h2>
395441

doc/pub/week1/html/week1-solarized.html

Lines changed: 54 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -109,6 +109,22 @@
109109
2,
110110
None,
111111
'good-books-with-hands-on-material-and-codes'),
112+
('Twst yourself: Deep learning 1',
113+
2,
114+
None,
115+
'twst-yourself-deep-learning-1'),
116+
('Test yourself: Deep learning 2',
117+
2,
118+
None,
119+
'test-yourself-deep-learning-2'),
120+
('Test yourself: Optimization part',
121+
2,
122+
None,
123+
'test-yourself-optimization-part'),
124+
('Test yourself: Analysis of results',
125+
2,
126+
None,
127+
'test-yourself-analysis-of-results'),
112128
('Types of machine learning',
113129
2,
114130
None,
@@ -554,6 +570,44 @@ <h2 id="good-books-with-hands-on-material-and-codes">Good books with hands-on ma
554570
from Goodfellow, Bengio and Courville's text <a href="https://www.deeplearningbook.org/" target="_blank">Deep Learning</a>
555571
</p>
556572

573+
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
574+
<h2 id="twst-yourself-deep-learning-1">Twst yourself: Deep learning 1 </h2>
575+
576+
<ol>
577+
<li> Describe the architecture of a typical feed forward Neural Network (NN).</li>
578+
<li> What is an activation function and discuss the use of an activation function.</li>
579+
<li> Can you name and explain three different types of activation functions?</li>
580+
<li> You are using a deep neural network for a prediction task. After training your model, you notice that it is strongly overfitting the training set and that the performance on the test isn&#8217;t good. What can you do to reduce overfitting?</li>
581+
<li> How would you know if your model is suffering from the problem of exploding Gradients?</li>
582+
<li> Can you name and explain a few hyperparameters used for training a neural network?</li>
583+
</ol>
584+
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
585+
<h2 id="test-yourself-deep-learning-2">Test yourself: Deep learning 2 </h2>
586+
587+
<ol>
588+
<li> Describe the architecture of a typical Convolutional Neural Network (CNN)</li>
589+
<li> What is the vanishing gradient problem in Neural Networks and how to fix it?</li>
590+
<li> When it comes to training an artificial neural network, what could the reason be for why the cost/loss doesn't decrease in a few epochs?</li>
591+
<li> How does L1/L2 regularization affect a neural network?</li>
592+
<li> What is(are) the advantage(s) of deep learning over traditional methods like linear regression or logistic regression?</li>
593+
</ol>
594+
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
595+
<h2 id="test-yourself-optimization-part">Test yourself: Optimization part </h2>
596+
597+
<ol>
598+
<li> Which is the basic mathematical root-finding method behind essentially all gradient descent approaches(stochastic and non-stochastic)?</li>
599+
<li> And why don't we use it? Or stated differently, why do we introduce the learning rate as a parameter?</li>
600+
<li> What might happen if you set the momentum hyperparameter too close to 1 (e.g., 0.9999) when using an optimizer for the learning rate?</li>
601+
<li> Why should we use stochastic gradient descent instead of plain gradient descent?</li>
602+
<li> Which parameters would you need to tune when use a stochastic gradient descent approach?</li>
603+
</ol>
604+
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
605+
<h2 id="test-yourself-analysis-of-results">Test yourself: Analysis of results </h2>
606+
<ol>
607+
<li> How do you assess overfitting and underfitting?</li>
608+
<li> Why do we divide the data in test and train and/or eventually validation sets?</li>
609+
<li> Why would you use resampling methods in the data analysis? Mention some widely popular resampling methods.</li>
610+
</ol>
557611
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
558612
<h2 id="types-of-machine-learning">Types of machine learning </h2>
559613

doc/pub/week1/html/week1.html

Lines changed: 54 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -186,6 +186,22 @@
186186
2,
187187
None,
188188
'good-books-with-hands-on-material-and-codes'),
189+
('Twst yourself: Deep learning 1',
190+
2,
191+
None,
192+
'twst-yourself-deep-learning-1'),
193+
('Test yourself: Deep learning 2',
194+
2,
195+
None,
196+
'test-yourself-deep-learning-2'),
197+
('Test yourself: Optimization part',
198+
2,
199+
None,
200+
'test-yourself-optimization-part'),
201+
('Test yourself: Analysis of results',
202+
2,
203+
None,
204+
'test-yourself-analysis-of-results'),
189205
('Types of machine learning',
190206
2,
191207
None,
@@ -631,6 +647,44 @@ <h2 id="good-books-with-hands-on-material-and-codes">Good books with hands-on ma
631647
from Goodfellow, Bengio and Courville's text <a href="https://www.deeplearningbook.org/" target="_blank">Deep Learning</a>
632648
</p>
633649

650+
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
651+
<h2 id="twst-yourself-deep-learning-1">Twst yourself: Deep learning 1 </h2>
652+
653+
<ol>
654+
<li> Describe the architecture of a typical feed forward Neural Network (NN).</li>
655+
<li> What is an activation function and discuss the use of an activation function.</li>
656+
<li> Can you name and explain three different types of activation functions?</li>
657+
<li> You are using a deep neural network for a prediction task. After training your model, you notice that it is strongly overfitting the training set and that the performance on the test isn&#8217;t good. What can you do to reduce overfitting?</li>
658+
<li> How would you know if your model is suffering from the problem of exploding Gradients?</li>
659+
<li> Can you name and explain a few hyperparameters used for training a neural network?</li>
660+
</ol>
661+
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
662+
<h2 id="test-yourself-deep-learning-2">Test yourself: Deep learning 2 </h2>
663+
664+
<ol>
665+
<li> Describe the architecture of a typical Convolutional Neural Network (CNN)</li>
666+
<li> What is the vanishing gradient problem in Neural Networks and how to fix it?</li>
667+
<li> When it comes to training an artificial neural network, what could the reason be for why the cost/loss doesn't decrease in a few epochs?</li>
668+
<li> How does L1/L2 regularization affect a neural network?</li>
669+
<li> What is(are) the advantage(s) of deep learning over traditional methods like linear regression or logistic regression?</li>
670+
</ol>
671+
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
672+
<h2 id="test-yourself-optimization-part">Test yourself: Optimization part </h2>
673+
674+
<ol>
675+
<li> Which is the basic mathematical root-finding method behind essentially all gradient descent approaches(stochastic and non-stochastic)?</li>
676+
<li> And why don't we use it? Or stated differently, why do we introduce the learning rate as a parameter?</li>
677+
<li> What might happen if you set the momentum hyperparameter too close to 1 (e.g., 0.9999) when using an optimizer for the learning rate?</li>
678+
<li> Why should we use stochastic gradient descent instead of plain gradient descent?</li>
679+
<li> Which parameters would you need to tune when use a stochastic gradient descent approach?</li>
680+
</ol>
681+
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
682+
<h2 id="test-yourself-analysis-of-results">Test yourself: Analysis of results </h2>
683+
<ol>
684+
<li> How do you assess overfitting and underfitting?</li>
685+
<li> Why do we divide the data in test and train and/or eventually validation sets?</li>
686+
<li> Why would you use resampling methods in the data analysis? Mention some widely popular resampling methods.</li>
687+
</ol>
634688
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
635689
<h2 id="types-of-machine-learning">Types of machine learning </h2>
636690

0 Bytes
Binary file not shown.

0 commit comments

Comments
 (0)