"100 Days of Machine Learning Code Challenge" proposed by Siraj Raval
Progress: It starts and progresses in the course of Machine Learning Applied with Python by Platzi. The following chapters are finalized:
- Introduction
- How to define a Machine Learning problem
- The Machine Learning engineering cycle
- Set up a Pydata enviroment
- Data preparation
- Modeling and evaluation
Thoughts:
- The concepts of preprocessing and dataset features visualization are more clear.
- Functionalities of the libraries numpy, pandas, matplotlib, seaborn and scikit-learn are known.
- The Lasso regression model is known and applied.
Link(s) to work: Platzi-ML-IMDB
Progress:
- I finished Feature Engineering chapter from Machine Learning Aplicado con Python course by Platzi.
- I watched ML.NET sessions from Microsoft DotNetConf 2018 conference.
Thoughts:
- I learned feature engineering techniques to apply to datasets like: reduction, transformation, scaling and binary encoding..
- I can look to ML.NET like a development framework growing fast and in a midterm it can help me develop business apps combined with ML. I like this idea...I have to start trying it.
Link(s) to work:
- Platzi-ML-IMDB
- Machine Learning in .NET (ML.NET)
- Artificial Intelligence and Machine Learning for Every .NET Developer
Progress:
- I finished Machine Learning Aplicado con Python course by Platzi.
- The basic concepts of ML are reviewed again by reading blogs.
Details:
- Validation methods (Cross validation, validation curves and learning curves) and model evaluations for Decision trees with con Ensembles and Grid Search are studied.
- When studying again the basic concepts of ML like its applications and algorithms, it is always good to help to understand in a better way the different concepts in which it has been deepening.
Link(s) to work:
- Commit: Platzi-ML-IMDB
- Blog Aprende Machine Learning: http://www.aprendemachinelearning.com
- Blog Aprende sobre Machine Learning: http://ligdigonzalez.com
Progress:
I assisted to a local event in my city about Data Science & Artificial Neural Networks hosted by Python Cali
Progress:
- I did exercises to practice ML algorithms: Exercises of Iris flowers classification and survival predictions on the Titanic are done using Regression Logistic, KNN, SVM and Decision trees models.
- The first 2 chapters of the Curso de Redes Neuronales y Backpropagation by Platzi are studied.
Link(s) to work:
Commit: ML practices.
Progress:
Working on a introductory presentation about Machine Learning to share with a group of friends.
Link(s) to work:
Commit: ML Intro
Progress:
- Done introductory presentation about Machine Learning to share with a group of friends.
- I hosted a "mini meetup" with a group of friends to share my knowledge about ML.
Link(s) al trabajo:
Commit: ML Intro
Progress:
- I finished Redes Neuronales y Backpropagation course by Platzi.
- I started Move 37 course by School of AI.
Progress:
- I watched Artificial Intelligence for the impatient developer video from .NET Conf CO v2017.
- Making progress with Move 37 course by School of AI.
Progress:
Practice OpenAI Gym with the environment CartPole to implement and learn the basic RL methods of Random Search and Hill-Climbing.
Source: http://kvfrans.com/simple-algoritms-for-solving-cartpole/
Link(s) to work:
Commit: OpenAI Gym-CartPole
Progress:
- I watched Machine Learning 101 video from .NET Conf CO v2017.
- I watched Google Dopamine video from Move 37 course by School of AI.
Progress:
- I watched Welcome to the age of conversational interfaces video from .NET Conf CO v2017.
- I started Introducción a Deep Learning course by Platzi.
Progress:
I watched the next videos about AI:
- How Smart Agents will shape the future from .NET Conf CO v2017.
- Inteligencia artificial y SQL Server 2017 from .NET Conf CO v2017.
- Servicios Cognitivos en la nube (visión, voz, traducción) from .NET Conf CO v2017.
- Conversational UI for Bots from Xamarin Show.
Progress:
- I watched AI for Every Developer video from .NET Conf 2018.
- I finished Introducción a Deep Learning course by Platzi.
Progress:
- I did an exercise to predict the prices of Boston houses using Linear Regression model.
- I did an executive introductory presentation to ML.
- I hosted a hangout with a group of friends to explain basics ML exercises of Classification and Regression.
Link(s) to work:
- Commit: ML practices.
- Commit: Presentation.
Progress:
I've started Chapter 01 of Building Machine Learning Systems with Python book.
Progress:
- I watched Getting Started with Visual Studio Tools for AI video from Microsoft Build 2018.
- I watched Cognitive Services in Xamarin Applications video from Microsoft Build 2018.
- I've finished chapter 01 and started chapter 02 of Building Machine Learning Systems with Python book.
Progress:
- I watched Demystifying Machine and Deep Learning for Developers video from Microsoft Build 2018.
- I've finished Chapter 02 of Building Machine Learning Systems with Python book.
Progress:
- I've finished chapter 03 of Building Machine Learning Systems with Python book.
- I've read some articles from DZone AI Zone.
Link(s) to work:
- 11 Deep Learning With Python Libraries and Frameworks.
- Top Machine Learning Algorithms You Should Know to Become a Data Scientist.
- How to Get Started With Conversational AI.
- Introduction to AI for Enterprises.
Progress:
- I wrote a story on Medium about my personal motivation with AI (Spanish only).
- I read some articles about comparisons between Fast.ai and Deeplearning.ai courses.
Link(s) to work:
- Motivación personal para iniciar con Inteligencia Artificial.
- Launching fast.ai.
- Meet These Incredible Women Advancing A.I. Research.
- FAST.AI: UP TO SPEED WITH THE BEST OF DEEP LEARNING.
- Fast.ai tips from a complete newbie.
- Ten Techniques Learned From fast.ai.
- Learning Deep Learning — fast.ai vs. deeplearning.ai.
Progress:
I watched the next videos about AI:
- How to Learn Deep Learning (when you’re not a computer science PhD).
- Road to Deep Learning.
- Road to Deep Learning II.
Progress:
- Video Conceptos básicos de Machine Learning.
- Video Introducción a los entornos virtuales en Python.
- Video Road to Deep Learning III Redes convolucionales.
- I've review chapters 1 and 2 from Building Machine Learning Systems with Python book.
- Video Error cuadrático en regresión lineal de Khan Academy.
Progress:
- I've review chapter 3 from Building Machine Learning Systems with Python book.
- I've review day 1 of Crash Course from PyImageSearch.
Link(s) to work:
Progress:
I've review day 2 of Crash Course from PyImageSearch.
Link(s) to work:
Progress:
Done: Day 3 and 4 of Crash Course from PyImageSearch.
Link(s) to work:
Progress:
- I've listened Computer Vision Explained with PyImageSearch's Adrian Rosebrock podcast from Hanselminutes.
- I've watched Machine Learning en ArcGIS webinar from Esri Colombia.
Progress:
- Techniques like Cross validation, Accuracy score, Confusion matrix and Classification report are applied using a RandomForestClassifier model for Iris y Titanic exercises.
- Techniques like Root Mean Square Error (RMSE), Coefficient of determination (r2_score), L1 and L2 penalties are applied using a ElasticNet model for Boston houses exercise.
Link(s) to work:
Progress:
- I've watched Netflix documentary AlphaGo.
- I've watched the next videos:
- Qué necesitas para hacer Inteligencia Artificial from AMP Tech.
- Drone Flight Controller from Siraj Raval.
- 7 Ways to Make Money with Machine Learning from Siraj Raval.
- I've assisted to the Google's virtual event Let's Talk AI.
- Done: Day 5 and 6 of Crash Course from PyImageSearch.
Link(s) to work:
Progress:
Done: Day 7 and 8 of Crash Course from PyImageSearch.
Link(s) to work:
Progress:
I've watched the next videos:
- Time Series Prediction from Siraj Raval.
- Train Machine Learning Models with Azure ML in VS Code from AI Show-Channel 9.
- Taking a Look at Computer Vision’s Object Detection from AI Show-Channel 9.
- Build a Bot in Minutes with QnA Maker from AI Show-Channel 9.
Progress:
- I've watched What's new with Speech Services video from AI Show-Channel 9.
- Done: Day 9 and 10 of Crash Course from PyImageSearch.
Link(s) to work:
Progress:
- Presentations about classification in Machine Learning are adjusted.
- I've watched the following videos from AI Show-Channel 9:
- Done: Day 11, 12 and 13 of Crash Course from PyImageSearch.
Link(s) to work:
- PPT-Clasificación I.
- PPT-Clasificación II.
- PyImageSearch-CrashCourse-Day12.
- PyImageSearch-CrashCourse-Day13.
Progress:
- Done: Day 14 and 15 of Crash Course from PyImageSearch.
Link(s) to work:
Progress:
- Crash Course from PyImageSearch DONE.
Link(s) to work:
Progress:
- Done: Lab - Introduction to Custom Vision Service of Learning Path: Custom Vision Service from Microsoft AI School program.
Link(s) to work:
Progress:
- Microsoft AI School Program - Learning Path: Custom Vision Service: Lab Exporting a Custom Vision model and deploy it to an Android device: In progress.
- PyImageSearch book - DL4CV: Chapter 1: Done and Chapter 2: In progress.
Progress:
- I've read the following articles:
- I've watched the following videos:
- Microsoft AI School Program - Learning Path: Custom Vision Service: Lab Exporting a Custom Vision model and deploy it to an Android device: Done.
Link(s) to work:
Progress:
- I've read the following articles:
- PyImageSearch book - DL4CV: Chapter 2: Done.
Progress:
PyImageSearch book - DL4CV: Chapter 3: Done.
Progress:
Microsoft AI School Program - Learning Path: Custom Vision Service: Lab Exporting a Custom Vision model and deploying on iOS: Done.
Link(s) to work:
Progress:
PyImageSearch book - DL4CV: Chapter 4: Done.
Progress:
PyImageSearch book - DL4CV: Chapter 5 and 6: Done.
Progress:
- PyImageSearch book - DL4CV: Chapter 7: Done.
- Presentation about Azure Cognitive Services: Done.
Link(s) to work:
Progress:
PyImageSearch book - DL4CV: Chapter 8: Done + Chapter 9: In progress.
Progress:
PyImageSearch book - DL4CV: Chapter 9: Done.
Progress:
I've read the following articles:
- Getting Started with Windows Machine Learning.
- Machine Learning - Accelerate AI Solutions with Automated Machine Learning.
- Machine Learning - ML.NET: The Machine Learning Framework for .NET Developers.
- The 6 most useful Machine Learning projects of the past year (2018).
- Deploying Deep Learning Models.
Progress:
PyImageSearch book - DL4CV: Chapters 1 to 6 are reviewed again.
Progress:
- I've read the following articles:
- I've watch the following video:
- PyImageSearch book - DL4CV: Chapters 7 to 9 are reviewed again.
Progress:
PyImageSearch book - DL4CV: Chapter 10: In progress.
Progress:
PyImageSearch book - DL4CV: Chapter 10: Perceptron Algorithm for AND, OR, and XOR Datasets.
- I've watch the following articles:
- I've read the following articles:
- PyImageSearch book - DL4CV: Chapter 10: Backpropagation with Python from scratch - Bitwise XOR.
- I've studied the maths behind the Backpropagation algorithm following the examples: A Step by Step Backpropagation Example and Backpropagation Step by Step.
- PyImageSearch book - DL4CV: Chapter 10:
- Backpropagation with Python - MNIST Dataset (Subset).
- Multi-layer Networks with Keras - MNIST Dataset (Full).
- Multi-layer Networks with Keras - CIFAR-10 dataset.
- "Multi-layer Networks with Keras - CIFAR-10 dataset" exercise was run on Google Colab (CPU).
- PyImageSearch book - DL4CV: Chapter 10: Done.
- "Multi-layer Networks with Keras - CIFAR-10 dataset" exercise was run on Google Colab (GPU).
- PyImageSearch book - DL4CV: Chapter 11: Convolutional Neural Networks - Done.
- "Multi-layer Networks with Keras - CIFAR-10 dataset" exercise was run on Microsoft Azure Notebooks.
- PyImageSearch book - DL4CV: Chapter 12: Training Your First CNN - Done.
- PyImageSearch book - DL4CV: Chapter 13: Saving and Loading Your Models - Done.
- Tutorial PythonProgramming.net - Deep Learning with Python, TensorFlow, and Keras Parts 1 and 2: Done.
Link(s) to work:
- Tutorial PythonProgramming.net - Deep Learning with Python, TensorFlow, and Keras Parts 3 to 7: Done.
Link(s) to work:
- PyImageSearch book - DL4CV: Chapter 14: LeNet: Recognizing Handwritten Digits + Chapter 15: MiniVGGNet: Going Deeper with CNNs - Done.
- Tutorial PythonProgramming.net - Deep Learning with Python, TensorFlow, and Keras Parts 8 to 11: Done.
- Exploring TensorFlow Object Detection API with the following recourses:
Link(s) to work:
- Studying TensorFlow Object Detection API with the following resources:
Link(s) to work:
- I watched the video A Closer Look at Intelligent Retail.
- Keep studying TensorFlow Object Detection API with the following resources:
- Exploring Tensorflow Object Counting API.
Link(s) to work:
- Studying TensorFlow Object Counting API with the following resources:
Link(s) to work:
- Reviewing Deep Learning concepts: loss functions, learning rate, decay, momentum and architectures with the following resources:
- Understand the Impact of Learning Rate on Model Performance With Deep Learning Neural Networks.
- Loss and Loss Functions for Training Deep Learning Neural Networks.
- Gradient descent with momentum.
- 10 Advanced Deep Learning Architectures Data Scientists Should Know!.
- The 9 Deep Learning Papers You Need To Know About (Understanding CNNs Part 3).
- A Beginner's Guide to Object Detection.
- I've watched TensorFlow course on YouTube of Jesús Conde. Course Link.
- PyImageSearch book - DL4CV: Chapter 16: MiniVGGNet: Learning Rate Schedulers - Done.
- PyImageSearch book - DL4CV: Chapter 17: Spotting Underfitting and Overfitting + Chapter 18: Checkpointing Models - Done.
- Configuring and testing Azure Data Science Virtual Machine (DSVM) for Deep Learning.
- I developed an Airplanes Classifier using Azure Custom Vision Service and Xamarin.Forms (iOS/Android).
- PyImageSearch book - DL4CV: Chapter 19: Visualizing Network Architectures + Chapter 20: Out-of-the-box CNNs for Classification - Done.
- deeplearning.ai - DL Specialization: Course 1 - NN and DL - Week 1: Done.
- deeplearning.ai - DL Specialization: Course 1 - NN and DL - Week 2 - Logistic Regression as a Neural Network: Done.
- PyImageSearch book - DL4CV: Chapter 21: Case Study: Breaking Captchas + Chapter 22: Case Study: Smile Detection - Done. Starter Bundle: Done.
- deeplearning.ai - DL Specialization: Course 1 - NN and DL - Week 2: Done.
- Top 5 takeaways from TensorFlow Dev Summit 2019 video: https://www.youtube.com/watch?v=YzLnnGiLNRE.
- Studying TensorFlow Object Detection API with the following resources:
- TensorFlow Object Detection API tutorial.
- TensorFlow Object Detection API in 5 clicks from Colaboratory.
- AMP Tech - Detección de objetos con tensorflow - Parte 1.
- AMP Tech - Detección de objetos con tensorflow - Parte 2.
- TensorFlow Object Detection API: basics of detection - Part 1.
- TensorFlow Object Detection API: basics of detection - Part 2.
- I watched the following videos from Channel 9 - AI Show:
- I've studied Section 6 (Object Detection with OpenCV and Python) from Python for Computer Vision with OpenCV and Deep Learning course by Udemy.
Still studying TensorFlow Object Detection API. Proof of Concepts are done with Webcam, Images and Videos. For videos, multiprocessing and threading was implemented. The following resources are used:
- Increasing webcam FPS with Python and OpenCV.
- Building a Real-Time Object Recognition App with Tensorflow and OpenCV
- Real-time and video processing object detection using Tensorflow, OpenCV and Docker.
Keep studying and applying TensorFlow Object Detection API, focusing mainly on performing re-training of models with custom datasets. Also begins to work and learn datasets such as Open Images, COCO and ImageNet for classification, detection and segmentation tasks. For now I will continue working and deepening the development of applications with TensorFlow, TensorFlow Object Detection API and TensorFlow Hub.