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.

Puppet Frog

~Sharat Kumar

In EVS session, we learned about amphibians and their characteristic features. We discussed what the unique features of amphibians are. We learned to model a Puppet frog to engage children with the lessons by storytelling and to explain the characteristic features of amphibians.

Amphibians are small organisms which come under vertebrates (animal with a spinal cord surrounded by cartilage or bone) that need water, or a moist environment, to survive. The species in this group include frogs, toads, salamanders, and newts. All can breathe and absorb water through their fragile and moist skin. Amphibians also have specialized skin glands that produce useful proteins.

We shall now see the equipment’s and steps required to make Puppet Frog.

Materials Required:

  1. Color Paper
  2. Compass
  3. Pencil
  4. Glue

Step 1:

Take two color papers and draw a circle using compass. Cut the remaining part

Step 2:

Make a hole in one circular sheet and cut it into half.

Step 3:

Make 4 legs, eyes and tongue with the remaining paper.

Step 4:

Apply glue to the edge of the circle and stick the semi-circle as shown below.

Step 5:

Stick the 4 legs and eyes as shown below using glue.

Step 6:

Now stick the tongue of the frog using glue as shown

Step 7:

Allow the glue to dry and the puppet is ready.

The House on Mango Street

~Hari & Praveen

This week we discussed the book “The House on Mango Street,” written by Sandra Cisneros, a Mexican-American author and a person of color. She grew up in a Hispanic community. Her memoir consisting of several short stories, brings out several issues around race, sexuality, culture, economic inequality, and gender inequality. This book helped us draw parallels between the critical issues as depicted in the stories and the cause centric initiatives that Quilt supports with various foundations.

Esperanza’s memoir shows her life in a neighborhood where female continuously becomes a victim of molestation, regardless of any age. The chapters “Rafaela Who Drinks Coconut & Papaya Juice on Tuesdays” and “Linoleum Roses” highlights Gender inequality and Domestic Abuse – by locking girls at home and forcing them to do house chores.

In the chapter “First job,” Esperanza confronted her first harassment – When her co-worker, an old adult, misbehaved with her, she could not do defend herself. She never brought up the abusive incident and continued to do her duty as she was in dire straits. She hid her age to join the company. Poverty suppresses her power of ‘Right to Freedom of Expression’ in one way or another.

We took the Quilt case study on ‘How might we empower low-income college girls to change their future?’ which focuses on gender equality and women empowerment. In Quilt AI, we analyze 107000 digital content pieces and 68000 search impressions. If this memoir got posted in any social media, Esperanza and other (possible) victims could have been made aware and ‘nudged’ towards a more reliable and safer path of support.

CROW

~ Saranya

Please refer this link  to know about “what for this blog is”

CROW

 

CROWS CHARACTERISTICS:

  • Males and females are almost identical.
  • Males are bigger than females.
  • Crows live in almost all parts of the world except Antarctica, the bottom part of South America, and New Zealand.

HABITAT:

  • Crows live in open spaces.
  • Agriculture fields.
  • Crows will not live in a forest or desert.

PARENTAL CARE:

  • Crows take 6 days to lay the eggs and 19 days of incubation.
  • The Male protects and gathers food. The female watches the baby birds and does not leave the nest unless to get water.
  • Both males and females work together to take care of their young.

LONGEVITY:

  • Crows will live from 6 to 7 years also crows can live up to 20 years of age in ideal condition.
  • Males and females live for the same amount of years.

SEASONAL PATTERNS:

  • When winder comes crows fly down to warmer climates.
  • During different seasons, Crows do not change their behavior

FACTS ABOUT CROW:

  • Groups of crows are called “murders”. The reason for this is that when a crow is dying of sickness, old age, or injury, the rest of the murder will often kill that crow in order to end it’s suffering.
  • Crows have the biggest brain based on body size out of all birds.
  • Crows have the ability to judge people by reading their faces and expressions.
  • Crows can imitate human voices like parrots.
  • Crows are a lot smarter than other birds.

HUMAN RELATIONSHIPS:

  • Many people think that crows are their Ancestors (people provide food to the crow).

Horned Lark

~Abilash

During the EVS session on Animal kindom, our guide Ravi Alunganthi suggested us to do a poster on any specific bird we were interested in. Each one of us chose a bird to be presented. Taking the  Horned Lurk, I googled its different characteristics such as name, habitat, diet, nesting, lifespan etc. Each individual presented their bird to the team, where we shared additional knowledge. The aim of the poster was for the students to get a better innate understanding of birds. As an activity given to the students during the lock down, a  student had shown interest and made her poster on Peacock.

  

Name: Horned Lark

Scientific Name: Eremophia alpestris

Family: Lark

Habitat: Prairies, fields, airports, shores, tundra. Inhabitants generally in open grounds, avoiding areas with trees or even bushes.

Diet: Feeds on small seeds from a great variety of grasses, weeds. Many insects are also eaten, especially in summer, when they may make up half of the total diet.

Eggs: Lays 2 to 5 eggs. Pale grey to greenish-white, blotched and spotted with brown. Incubation is by female, about 10-12 days.

Nesting: Often nests quite early in spring. Male defends nesting territory by singing, either on the ground or in flight. In-flight song display, male flies up steeply in silence, often to several hundred feet above the ground, then hovers and circles for several minutes while singing; finally dives steeply toward the ground. Nest site is on open ground, often next to grass clump, piece of dried cow manure, or other objects. Nest (built by female) is a slight depression in the ground, lined with grass, weeds, rootlets, with the inner lining of fine grass or plant down. One side of the nest often has a flat “doorstep” of pebbles.

 

 

 

UNDERSTANDING ALIASING AND SAMPLING USING PYTHON

~Bakyalakshmi

Sampling theorem:

fs≥2fm

  • A continuous time signal can be represented as samples and can be recovered back when sampling frequency fis greater than or equal to twice the highest frequency component of message signal.
  • If this condition does not satisfy, it leads to aliasing.
  • Aliasing is an effect   that causes different signals to become indistinguishable when sampled.

Visualizing using python:

import matplotlib.pyplot as plt #  to plot

import numpy as np

#numerical python to get array of float values and for sine operation

t = np.arange(0, 2e-3, 10e-6) # x axis time period

# sampling at fs =10kHz in time domain ts=1/fs (0.1ms)

ts = np.arange(0,2e-3,0.1e-3)

f = 1000 # message signal fm

b = np.sin(2*np.pi*f*t) #phase for sinewave

c = np.sin(2*np.pi*f*ts)

plt.plot(t,b,”g”) # plot of message signal (1kHz)

plt.plot(ts,c,”k*”) # plot of sampled message signal (1kHz)

fs>=2fm:      Input frequency= 1kHz          sampling frequency = 10kHz

f=9000

b = -np.sin(2*np.pi*f*t)

c = -np.sin(2*np.pi*f*ts)

plt.plot(t,b) # plot of message signal (9kHz)

plt.plot(ts,c,’r+’)  # plot of sampled message signal (9kHz)

fs<2fm:      Input frequency = 9kHz        sampling frequency = 10kHz

Sampled output of 1kHz and 9kHz :

Aliasing of 1kHz and 9kHz

part1: https://youtu.be/og-Pn2oOqP4

How to build 3-D Flexagon

~Saranya

1. Take a printed sheet shown in the below 
2. Mark numbers in each row so that we won’t miss the order. Follow the below image


3. Marked 1 as the 1st row 2nd as the 2nd row 3rd as the 3rd row and finally 4th as the 4th
row.
4. Cut shape that printed out in the paper(wherever it has bend you have to bend it).
Finally, you should get only that shape of the sheet there shouldn’t be any excess paper.
5. Take empty ballpoint pen to draw the all the diagonal to fold it easily(just draw a line on
top of the all the diagonal)
6. Draw what you want to convey for example(sample image shown in the below):

7. Wherever it says glue apply the glue and fold it.

8. Here is the link for how it works. https://youtu.be/x4aqoaT5zGw

Relief work with migrant laborers

During this lock-down period, the migrant laborers struggled a lot for food and shelter. At STEM land we had an opportunity to work with an NGO called Coast India for helping migrant people. Ten of us volunteered for this NGO. The NGO had collected a database from the state government on migrant laborers who had earlier called in for help and put this information in an app. Our responsibility was to call the concerned people and verify the data and update their current specific requirements of migrant laborers. We spent about 2hrs per day for about a month. We worked with the migrants who were from Jharkhand in Tamil Nadu.

We called around 600 contacts and through them we reached nearly 2000 others. It was a difficult task for us to communicate since most of us do not know Hindi and most migrants could communicate with limited Tamil, but we still managed to understand and help them get food , shelter and transportation to get home by working in tandem with the NGOs on the ground in their areas.

Initially the government had promised that with updated data the migrants will have money transferred to their accounts. However, this did not happen and many of them were frustrated when they didn’t get the support committed by their government and additionally the situation at hand was also hard. The calls were hard to take as they were emotional and it even affected some of us as we were expected to continue to follow up for updated information. Nonetheless, our team members consoled them and helped them by sending many emails to the NGOs for follow up. All this struggle and emotional stress was wiped away when many of the migrant people sent us messages that they have safely reached home or got the rations they needed. We would like to thank Coast India NGO for creating this opportunity to help in relief for those who were stranded during the Carona.

Some insights

Working on COVID-19 relief work with Bindu and team was one such experience which made me realize one of the most pressing issues in our society – migrant workers. This issue was invisible to many of us till we got the intensity of it – the number of migrants, many unregistered, working in most hostile conditions thousands of kilometres from their homes, away from family.

As Bindu planned, we were given contact details and basic data from initial calls (mostly state government call centres). We had to reach out to the people on our list and get the latest status regarding their food, stay, health and any other essential things. There were mainly two phases:

1) initial phase when our migrant friends were having difficulties getting food, dry ration etc. Though we used to take note of their needs, many times it wasn’t possible to reach out to them with the help considering strict lockdown and resources at hand, both financial and human. Bindu was working hard to manage the ration/food for those who are in urgent needs, even with the scarce resources at hand.

2) In the second phase, there was a shift from getting food to getting transportation to their home states. This was the time when I realized that though many don’t have enough food to twice a day, they were not asking for it but for transportation.

Many times, it happened that they were frustrated by filling up different forms of the two states, with changed rules over every week. I experienced that through phone calls but wasn’t able to do anything about their travel. Eventually some of them started walking to their home with no hope of governments and administration managing their travel.

In all this chaos and helpless situation, I just started to listen to them. Some of them couldn’t control their tears, some just having hopes in their eyes. Some just needed to know the status of travel, and some of them were hopeful that someone is at least listening to what they are facing. Though I couldn’t contact all the people from cases assigned to me, I had contacted many of other migrants and could share some important updates with them as my number was being circulated by many of them among their other friends.

I couldn’t work from the last two weeks on this issue as it was very time consuming (I used to get calls all day) and I had other responsibilities. I still feel I could at least help some of them if not as many as I would have liked.

It moved me to see these many friends of ours have to go through these difficult times but there are people like Bindu who are working tirelessly to help them. Glad to be part of this initiative.

– Ganesh Shelke

When I called them, I came to know most of them are facing problems with food, shelter and many have to go back to Jharkhand. Hearing their problems and helping them to go back to their hometown gives me satisfaction. Thanks for giving me this opportunity.

– Sharat Kumar.N

I felt connected when I talked to people. Language was an issue for me. I was not able to speak in Hindi and was not able to reach people easily. They were asking for support and used so many “please” it was hard for me at that time to listen to their queries. Some of them were physically not well, didn’t get ration items, had to pay rent, wanted to go home and so on. Most of the time I had this question within me that, is what I do really useful and does it reach the one in need. I had this question when many of them said that it has been 20 days since I have requested, but still we didn’t receive anything. But I processed it within myself and took it as an opportunity of my growth and others.

-Poovizhi

I feel content and happy that I was able to support others during their difficult situation. It was great that their government (Jarkhand) took so much effort to give their support to the migrant people through this NGO (coast India) and Jarkhand Sahayta App. When people informed me that they received support from NGOs I was very happy. When some people did not receive any help at that time, we sent an email to NGO people to give support to the migrant people. The NGO people took an effort and gave their support to the migrant people. It was a great and new experience

– Saranya

Initially I had a willingness to help people, so I called them and asked about their needs, and I was happy by doing this. For two weeks, I was supporting people by registering their basic needs in our portal and bringing some urgent issues to the team. But after two weeks most people asked me to help them to reach their hometown.

I shared the information both in Hindi and English regarding special trains and e-pass. Later I realized that I was creating expectations in people’s mind that they can travel to their hometown when the government allowed only one or two trains. So, I stopped calling new cases, and only supported the old cases which have been assigned to me.

-Ranjith

Even though I have been very busy in my work, I separately allocated time for calling migrants. When I was calling them, they were at the stage of losing hope, my words consoled them. I felt very happy. Working with the unknown people was new for me and I hope I have given my best.

-Vasantharaj gandhi

It was a different and challenging experience, because of the language which plays a medium for communication. I did this for my inner satisfaction. But because of some other works and self-development activities, I was not able to continue further. I appreciate and acknowledge all the others who did this service for a good cause.

-Vasanth

 

Algebraic kit

-Logeshwari, Tamizharasan

During the session with Ravi Alungati, we made a lot of math materials.

By using the cardboard boxes Gurumoorthy started doing Algebraic kit. Initially, I supported him. Then after that, I worked with Tamizharasan. We made x power 2, x and ones using the cardboard in inches.

To visually see a quadratic equation we can use this kit. The yellow-colored part is positive and the black colored part is negative. The squares are x power 2. The rectangle pieces are x and the small squares are the ones.

The example of a quadratic equation is as follows.

x^2 + 4x + 4 which can be factorized and we have (x+2) and (x+2)

The above equation can be represented as follows.

Quadratic equation with negative terms.

x^2 +0x -4 which can be factorized and we have (x-2) and (x+2)

The above equation can be represented as follows.