AnshuTrivedi/sg_ml_projects

Repository containing all projects,activities,initiatives completed or going on in study group.Microsoft Scholarship Challenge Phase.

sg_ml_projects

Repository containing all projects,activities,initiatives completed or going on in study group.Microsoft Scholarship Challenge Phase.

Week 1 [20/07/20 - 27/07/20]

Week 1 activities details

Introductions of team members

We had a introductory session with all the group member in which every introduced themselves and shared interesting facts about them.

Poll objective :To identify profeciency in Machine Learning,it helped to begin with project work. Below is poll result.

Suggested platforms from where we picked up projects

Projects selection and team members

Projects	Team Members
1: Identify the apparels (Fashion MNIST)	Nil
2. Emergency vs Non-Emergency Vehicle Classification	@mars, @Pasupuleti Chandana, @Joyce Annie George, @Rajasekhar
3. iMaterialist (Fashion) 2019 at FGVC6 Fine-grained segmentation task for fashion and apparel	@Anshu Trivedi @Balkrushn Hirani
4. House Prices: Advanced Regression Techniques	@Todimu Isewon, @Pragna Das, @Antanas, @Shanaya Mehta, @Vishal Bansal, @Varshini Ravi, @Nuria A Malet Quintar
5. Big Mart Sales Prediction	@Rene Essonba @Fatima @Sabah @Quanitah Shaikh
6. Movie Recommendations	@christine_s, @Purvaja Durga Barnala, @Akanksha Singh, @Sunanda Somwase, @Surya, @shaini soni, @Marianna, @Sreenithi Sridharan, @Yolanda Zhang, @AyushiChoudhary, @Aishwarya Das, @Megha @lary @Sabyh,@Shubhra Sachdev
7. Windmill Turbines Fault Detection	@Shudipto Trafder, @Grace.D, @Sabyh, @Yash Kulkarni @Aarthi Alagammai
8. Stock Prices predictor using TimeSeries	@Akhil Dsouza

week 2 [27/07/20 - 02/08/20]

Week 2 plan

• Please decide in your team a representative or team leader who have to update here about team progress,organizing team meetups. I request to post everything related to projects work here instead of DMs .
• This week we will have meetup where we will discuss about roadblocks and project progress.
• We can organize session for beginners . Looking for speakers and session topics from beginners.
• Start project work and if feel any problem let us know .

Project : Windmill Turbines Fault Detection updates.

• @Yash Kulkarni and @Shudipto Trafder guiding and leading project team.

• @Shudipto Trafder shared resoures helpful for project work

  • Image classification https://www.kaggle.com/iamsdt/image-classification-cats-and-dogs-basic-to-end/
  • Tensorflow comes with two versions. TF1 and TF2. In 2020 you should start with TF2.Here is an introductory video from TF world 19. https://www.youtube.com/watch?v=5ECD8J3dvDQ
  • Udacity also offered a Free course on TF2. https://www.udacity.com/course/intro-to-tensorflow-for-deep-learning--ud187 This course covered ->
    1. Basic ML
    2. CNN
    3. RNN
    4. Time series
    5. TF Lite model (TF lite model is the light version of your main TF model)
  • Coursera: Tensorflow Specialization offered by deeplearning.ai https://www.coursera.org/specializations/tensorflow-in-practice It's a collection of 4 Courses
    1. Introduction to TensorFlow for Artificial Intelligence, Machine Learning, and Deep Learning
    2. Convolutional Neural Networks in TensorFlow
    3. Natural Language Processing in TensorFlow
    4. Sequences, Time Series and Prediction
  • And Last on from youtube. Freecodecamp, they are non-profit organization. You can check this course. it 7 hours long. It covers all the topics covered from udacity and extra one Reinforcement Learning. https://www.youtube.com/watch?v=tPYj3fFJGjk&t=91s

• 1st Project meetup keypoints

1. We have planned to cover all the basic concepts related to image data and start with smaller tasks like building a classifier. We are planning to train our model manually by using Google Colaboratory and not using the platforms like Microsoft Azure or AWS.
2. Majority of the team members want to have hands-on experience by building a model by actual coding the Neural Network layers in python. Most of the guys including me are new to the image data and we have to still look for more relevant datasets if possible. Maybe it will take little longer for us to complete the decided task and all of us are totally happy with it 🙂
3. Although few team members were not able to connect to the conference meeting, we have recorded our meeting so that everyone could go through a meeting discussion. Cheers !! to everyone in the team @gul jain @Grace.D @Sabyh @Aarthi Alagammai @LauraT @Yash Bhawsar. :star-struck:
4. A special mention to one of our team member @Shudipto Trafder who guided us with the approach of how to handle image data as he has got experience working with deep learning projects. Thanks to him for letting us all with the important links.:thank_you:

Project : Emergency vs Non-Emergency Vehicle Classification

• Polled for weekly project meetup and decided Saturday for meeting

Project: Movie Recommendations

• @Sreenithi Sridharan organised poll for weekly project meeting but couldn't decide due to voting proportion
• @Sreenithi Sridharan organised another poll to decide project meetup day
• First meetup organised
  Topic: Movie Recommendations Project - First Step Discussions
  Time: Jul 31, 2020, 10AM PDT | 5PM GMT | 10.30PM IST
  
• 1st Project meeting keypoints
  1. 👉 First, we discussed what a Movie Recommendation system is. It can either be a simple score based sorting of top-rated movies or personalised movie recommendations based on characteristics like Movie Genre, Director, Actor, Rating, etc. In our case, we will be working on a personalised Movie Recommendation System.
  2. 👉 So as with any ML project, our first step will be to find the right dataset for this project. Some known datasets are MovieLens, IMDb, and TMDb. So the first task we decided is that we will research these 3 datasets and the features they have and come up with observations that we can discuss upon in the next meeting. Based on the discussions, we will decide on which dataset to use for this project.
  3. 👉 And in the next meeting along with choosing the dataset, we'll discuss and split up the work for data cleaning, preprocessing, and feature engineering.
  4. 👉 Once the dataset is ready, we can have different team members try different ML models and compare the results. So for discussing about the dataset, we'll have another quick meeting day after tomorrow. Will setup a time poll for it soon!
• @Sreenithi Sridharan organised poll for quick meetupto to finalise the dataset.

Project:Big Mart Sales Prediction

• Members started thread for project collaboration
• Members are looking for project leader

Project: iMaterialist (Fashion) 2019 at FGVC6 Fine-grained segmentation task for fashion and apparel

• @Anshu Trivedi @Balkrushn Hirani discussed about project but couldn't make much progress.

New members joining projects

• @Shubhra Sachdev requested to join Movie Recommendation project
• @Akinwale Oshodi willing to join project work but didn't join yet.

week 3 [03/08/20]

week 4