No Code / Less Code Machine Learning/Deep Learning

~ Vimal 

As a novice, one will feel alienated with the buzz words and vocab used in the context of ML/DL world. To bring more people into the AI scenario, several of the tools are being developed by researchers and companies worldwide. They frame this with the phrase “Democratizing AI”. The corporates of course focus mainly on the subscription business model for enabling one to make their custom machine learning / deep learning model. Whereas few of the academia geeks focus on developing tools to bring the common man into the AI world. Some of the examples include

  • Ktrain
  • Ludwig
  • Fastai

In this blog post I wish to introduce a simple and easy to use library ktrain. This is a high level wrapper library for the Tensorflow keras(a low level API like library for the TF). Using this library we can  build, train & deploy a deep learning model at ease. It also provides interfaces and includes functionalities from other SOA deep learning library PyTorch(from FB Research).It abstracts the various processes one will encounter while building a deep learning model for his/her use case. It provides support to create models for text & image processing workflows. Text processing or widely referred as Natural Language processing is the toughest job every machine learning engineer will encounter. There will be a lot of bias in selecting the different pretrained models for different tasks(sentiment analysis, text classification, text generation, Q/A systems etc.). This ktrain library comforts the user by doing all the heavy lifting. A text classification pipeline will only require 3 lines of python code to train and build a model. It also possesses the methods to save the model for predicting the unknown datasets. One catch is the zeroshot classifier pipeline which supports the user to classify the text/sentences/documents without training. It has support to import all the transformer models like BERT, distilBERT, XLNet etc. Image classification task is easily achievable with the pretrained model like ResNet50, Inception etc. It supports graph & tabular dataset also to enable the user for custom building a model.

A quick tutorial to execute a zeroshot sentiment analysis classification(binary classification – Positive/Negative)

  1. Install the ktrain through pip 
  2. Execute the code

With the introduction of transfer learning capacities(through transformer architecture) to the natural language related tasks, the NLI/NLU/NLG(Natural Language Inference/Understanding/Generation) field is speeding up in the last couple of years paving the way to surpass the human capabilities in all the language related tasks.

Electronics Session – LDR + Arduino – Street Light Concept

~Vimal , Jenifa & Abilash

As a part of the “People Counter Project”, the children were given hands on learning about the LDR(light dependent resistor) and the interfacing of the same wit the arduino platform. children built the circuit, wrote the program with support and understood the concept LDR , differentiation of analog and digital values, serial communication between arduino and pc. They visualized the value of the ldr on the serial monitor tool. Further a task was given to them to build the concept of automatic street light, with little help from us they understood and wrote the code and demonstrated it. Also threshold concept was explained to them and asked to write a code for switching on two leds based on the threshold limit of the sensed value of the LDR connected to the arduino. They did that also.

Hands on Session on Math Concepts

~ Vimal & Praba

An interactive and engaging session was held for the STEM Land  mathematics teacher for showing Off Math concepts with Mr.Ravi Aluganthi. We as a team of Two worked on building a model to demonstrate the Pythagoras Theorem. We learnt the tools of trade of model building. Also learnt about the techniques for building the model. Through this session, we understand the mathematical concepts in a better manner than through the textbooks. We shall work on building the same kind of models to demonstrate the difficult to show math topics to enable the kids learn easier. We are very grateful to Mr.Ravi for his valuable support and help in our learning path.



Arduino Extension for Scratch 3.0

Scratch 3.0 doesnot have the standard extension to interface with the popular open source physical computing platforms like Arduino. An extension was build to communicate with the arduino platform. Arduino Uno needs to be flashed with the firmata firmware from the examples. The scratch communicates with the arduino through a intermediate tool developed using the pyfirmata library of the python. This tool helps the kids to learn physical computing concepts easily. This same framework can be used for research on Human Computer Interaction. Also for building tangible devices. The customized version of the Scratch 3.0 and the Scratch linking software called ScratchBridge is released in the github page:


Session on Brain with Inspiration Team(React, Respond, Realise)

~ Vimal, Abilash, Ranjith

An interactive session on the concepts React, Respond & Realise was held with kids to understand the various types of brain parts associated with these emotional reactions. Kids learnt about the types of brain and various functions these brain will do in our normal day to day life. Also they were able to relate these concepts with the story we have taken up(Maharaj Parikshith). All the kids are asked to draw the types of brain on their hand and on paper for better understanding.The kids came up with a lot of question on the topic, which were brainstormed and understood by all the kids. Every kid shared their learning from the session.