Deep Learning – 001: Introduction to Deep Learning.
Deep Learning (DL): can you answer these introductory questions on DL? Target: Starters in DL
Can you define AI, ML, DL?
Can you draw a diagram to show the relations of AI, ML, DL?
What is Symbolic AI?
Is Symbolic AI good for Image Classification, Speech recognition, and Language Translation? and Why?
How are AI, ML, DL for Image Classification, Speech recognition, and Language Translation? and Why? Can you compare these among them?
What is the motivation behind inventing Machine Learning?
How does programming differ for ML than classical programming?
Which one uses data intensively? Classical Programming or ML?
Which one uses Rules intensively? Classical Programming or ML? and why?
When was ML invented i.e. the idea? When it flourished? and why?
Why did not ML flourish earlier?
How recent is Deep Learning (DL)? When did it start to flourish?
How prominent is DL in Kaggle contests?
What is more extensively use in DL? Math or Engineering? and why? Is it for good or for bad? Can you think of any limitations?
Among ML, and DL where math is more extensively used? and why? where in AI, ML< DL such Math focused study might not be the greatest option? and why?
How ML differ from Math and Statistics esp. Statistics and why?
Can you name some Prediction approaches in ML?
How are data represented in ML and DL?
What is important in ML? Deep understanding or Representation of the layers of the process and data?
What is important in Dl? Deep understanding or Representation of the layers of the process and data?
What is the link among Brain, DL, and NN (Neural Network?)
What are other possible names of Deep Learning?
Does deep learning identify digits (picture recognition, image classification) in one step or multiple steps?
What is Deep NN?
What are some parameters for NN and DL? or just name what you use when writing (representing) a DL solution.
What are Weights, Input, Layers, Predictions, loss function, objective function in DL?
Can you define/explain (not memorized) and give examples of Weights, Input, Layers, Predictions, loss function, objective function in DL?
What is the optimizer in DL?
What is a Back-propagation Algorithm in DL? Explain, Example, Draw, write code and show
What applications have used DL successfully?
Will you apply DL for all Learning, classification, prediction problems? Why or why not? When DL can do the best?
What are other alternate approaches than DL?
Why did not DL florished/used much in the past?
Is handwritten digit classification be an application where DL can be used? What companies might get benefit with such applications?
–
Misc.
What are Kernel methods? Give examples. How are they related to DL?
Is SVM a kernel method? What does SVM stand for?
How does SVM work?
Is SVM great for image classification?
What is Gini Index? What does it mean?
What is Information gain in decision trees? what is the name?
What is random forest?
What is Gradient Boosting?
–
What DL approach worked best for image classification?
What is CNN? What is RNN? What is RNN-LSTM? What is LSTM? What is gated RNN?
How does CNN work? Can CNN be used for Image Classification? What level of accuracy you can achieve with CNN? If used, how will CNN work to classify images?
Do you know what Gradient Boosting is?
What are XGBoost Library?
What is Keras library?
What are the libraries available for DL implementation? What are the languages that are best for DL? and why?
CNN got invented on 1989; LSTM on 1997. Why they got popularity or in use today? Any similarities why DL flourished now?
what kind of CPU is geared towards DL?
What is CUDA?
What are the contributions of NVIDIA for DL?
What is common in NN, and DL? Addition, multiplication, matrix, matrix-multiplication?
What are the potential future applications of DL?
What are the potential improvements potential or important for DL?
Can you do DL in Python Library, Theano, TensorFlow, Keras?
What is the most used for DL? Python Library, Theano, TensorFlow, Keras
What is Deep Mind? What does the project do?
What is RMSProp?
What is Adam?
What are some optimization approaches in DL?
What is the role of Weight in DL? and in DL code?
What is TPU? Who invented it?
Give examples of some Loss functions used in DL
–
Some Answers
What is Symbolic AI?
"Symbolic artificial intelligence is the term for the collection of all methods in artificial intelligence research that are based on high-level "symbolic" (human-readable) representations of problems, logic and search. Symbolic AI was the dominant paradigm of AI research from the mid-1950s until the late 1980s."
From: https://en.wikipedia.org/wiki/Symbolic_artificial_intelligence
Example: Prolog
I did teach a very introductory practical course on Prolog on around 2002 – 2003. For sure, I do not remember the content that I taught.
"Symbolic Reasoning. A reasoning is an operation of cognition that allows – following implicit links (rules, definitions, axioms, etc.) – to produce new knowledge from already existing knowledge. The reasoning is said to be automated when done by an algorithm.
Symbolic Reasoning – Sem Spirit"
From: www.semspirit.com › artificial-intelligence › symbolic-reasoning
What are other possible names of Deep Learning?
Ans: Layered representations learning?
Hierarchical representations learning?
What are some parameters for NN and DL?
Ans: Weights, Input, Layer, Predictions, loss function, objective function
What applications have used DL successfully?
Ans: Human level image classification, Speech recognition, Hand-writing, machine translation, text to speech, Digital Assistant Google Now, Amazon Alexa, Autonomous Cars, Ads Targeting, Natural Language Questions, superhuman go Playing
What are other alternate approaches than DL?
Ans: Statistics for data analysis, Naive Bayes Algorithm, logistic Regression, hello world algorithm.
Why did not DL flourished/used much in the past?
Ans: Missing efficient way of training large neural networks
What is XGBoost?
"XGBoost is an implementation of gradient boosted decision trees designed for speed and performance."
https://machinelearningmastery.com/gentle-introduction-xgboost-applied-machine-learning/
Give examples of some Loss functions used in DL
"This tutorial is divided into three parts; they are:
Regression Loss Functions. Mean Squared Error Loss. Mean Squared Logarithmic Error Loss. …
Binary Classification Loss Functions. Binary Cross-Entropy. Hinge Loss. …
Multi-Class Classification Loss Functions. Multi-Class Cross-Entropy Loss. Sparse Multiclass Cross-Entropy Loss."
https://machinelearningmastery.com/how-to-choose-loss-functions-when-training-deep-learning-neural-networks/
Sayed Ahmed
Linkedin: https://ca.linkedin.com/in/sayedjustetc
Blog: http://Bangla.SaLearningSchool.com, http://SitesTree.com
Online and Offline Training: http://Training.SitesTree.com
Affiliate Links:
Hottest Deals on Amazon USA: http://tiny.cc/38lddz
Hottest Deals on Amazon CA: http://tiny.cc/bgnddz
Hottest Deals on Amazon Europe: http://tiny.cc/w4nddz
