A Foray Into The World of Python: MUPy Conference
The second edition of the MUPy Conference, hosted at Manipal Institute of Technology on the 21st and 22nd October, was a great platform to introduce Python to the Manipal populace.
Python is a programming language known for its readability. The conference featured a wide range of speakers ranging from industry specialists to first-year students. They spoke about the different ways in which Python can be used and why it could be a good choice for them. They introduced the audience to the versatility, ease of use, and power of Python. The very fact that the largest room in NLH was completely full bears testimony to the success of the event.
Athulya Mohandas Menon
Mr Vivek Kalyanarangan, a member of the United Network of Data Science Professionals (UNP), opened the seminar on Machine Learning and the importance of Dockers to a packed room of eagerly awaiting participants. He was very efficient and concise as he talked about how Dockers can be used to replicate and reproduce various system environments and the various advantages they provide over Virtual Machines, such as using up much less of the computer’s physical resources. Using the real-life example of a search engine, he explained how simple bit-wise multiplication of search vectors is used to carry out searching. The example helped the listeners grasp the concept of how search engines use machine learning techniques like matrix multiplication to produce an output. He concluded with a demo of how a search engine can be built using a Docker system.
The next speaker, Dheeraj Rajagopal, started off by explaining the basic differences between a CPU and GPU. The CPU has fewer cores- each individual core is faster and has a greater capacity to perform sequential tasks. The GPU, on the other hand, has a higher number of duller cores, which makes it better at parallel tasks. He went on to explain the various problems faced while trying to transfer data into a GPU and how one could overcome them by using multiple threads. He stressed the importance of deep learning frameworks and how one could easily build computational graphs of large size using them.
Daniel J Evans, a student at Carnegie Mellon University, talked about the evolution of science and technology and how it affects our lives. He also talked about the history of computing and compared the ages of various programming languages like C, C++, and JAVA. The talk was a refreshing change and had a lot of information on the cultural and social impact that computing and data management has had on our lives. He discussed various publications and papers that have come out in the field of digital electronics and computing. He went on to talk about the approach taken by the authors and also commented on the quality and content of these papers. By doing so, he gave valuable insight into the historical background of Python to the ‘techies’, who weren’t really sure about the humble beginnings of computing.
Twenty-two year old Abhishek Kumar, who works as a deep learning scientist at Pedible Wealth, was the last speaker for the day and was easily the most engaging speaker. His informal yet informative style of presentation won over the audience fairly quickly as he explained topics such as Deep Learning and Neural Networking with ease. He spoke about segmentation, which is basically the process of partitioning images into multiple segments. It finds practical use in Tesla motors, self-driving cars, and medical imaging. He talked about the various techniques used for image segmentation such as the histogram matching method, before moving onto Deep Learning. Next, a bit of biology came into play as he discussed Neural Networking and how it contains an input layer, a hidden layer, and finally an output layer. The explanation of the algorithm involved was done creatively, using the picture of a small child and his toy, which garnered laughter from the crowd.
Aayushi Agarwal and Anisha Das
On the second day, the conference was divided, with interactions taking place simultaneously in different rooms. The talks attracted a sizable crowd consisting of amateurs and experienced coders alike.
Kartik Mandaville, a co-founder of SpringRole Inc. and an alumnus of MIT, spoke about Blockchain and Cryptocurrency. After giving a brief introduction about the process of verification that we generally use today, he went on to explain how Blockchain could revolutionize the way we record transactions. He also demonstrated how you can create your own Cryptocurrency using Python in just a matter of minutes.
Anurag Sharma, a first-year MIT student and an open-source and freelance software developer, talked about Open Source and Revision Control System (RCS). The talk included topics such as modern web and app development methods using Python and Model-View-Controller Architecture, and also gave attendees a brief introduction to Django, Web2py, and Flask which are all Python-based web frameworks. Ayodhyankit Paul, a machine learning enthusiast, demystified ‘decorators’ in his talk. Decorators are a very useful function in Python but many find it difficult to grasp. This talk aimed to introduce decorators by explaining how and why they work.
Naren Ravi has worked as a Trainer, Embedded Engineer, and Backend Engineer in his four years of industry experience. His talk aimed to clarify the difference between concurrency and parallelism, two seemingly similar concepts. He also talked about how you can use Python to achieve parallelism. Aneesh Joshi, a Deep Learning enthusiast, brought his knowledge to the table through his talk on Neural Style Transfer. Neural Style Transfer involves building a neural network that can transfer the style of one image to another. He tried to simplify the concept by explaining that style transfer can be concentrated to a problem of optimizing our result.
Lakshay Kalbhor, a first-year student, has created many open source projects in Python and Go. He, along with a few other students, created MIT Hodor, a Facebook messenger bot that tells you anything you ask it through the SLCM portal. It can tell you your attendance, timetable, and any other information that is available on the portal. He decided to create this bot as he noticed various people were having problems accessing SLCM on their cell phones. He explained the main process of how the bot works and talked about the problems they faced. Python was useful in creating the database and logging.
The afternoon session of the conference had six talks lined up, of which three were student speakers and the other three were speakers from the corporate field. The objective of the talks was to create awareness of the widespread use of the programming language of Python, with the goal of increasing its popularity.
A talk on the subject “Python in The Field of Image Processing and Computer Vision” by Siddharth Kotiyal, a current 4th year CSE student, focused on the use of Python in various fields of image processing such as object recognition, object detection, and motion analysis. The audience was taught some basic concepts of image detection, such as methods of detecting edges and corners, and the concept of how filters work. Ankush Anshuman, a third-year student and co-founder at Alexander Computing Pvt. Ltd., spoke about how we can use open source to make storage and manipulation of data much easier by accessing servers directly, such as with the use of Google sheets for automatic data backup.
“Neural embedding for Polysemous words” by Yash Kumar Lal, a fourth year CSE student and one of the founders of PyPals and MUPy, showed the more linguistic side of computer programming. They are currently researching on how to handle polysemy, the multiple sense of a single word when used on the internet. The talk delved into the field of distributional semantics. “When we write a sentence on the internet, a particular word can hold many meanings depending on the context of the sentence. We are trying to create a program that helps distinguish between the two different senses”, said Yash.
The next speaker, Mohammad Sajid, who works as a data scientist and teaches Data Sciences online at StepUp Analytics, spoke about the different aspects of data analysis. The talk had a very serious environment and was quite technical. He gave the audience a practical approach, with the example of returning all words related to the one entered in the search field. “What I learnt about Python in one week”, was a talk delivered by a Manipal alumnus and one of the founders of MUPy, Sailesh Sriram, who is currently working at Amazon. He gave a detailed description of all the things he learnt about Python in one week. The last talk was on “Server-less Apps using Python” by Ronald Das, a current fifth-semester student who is also a co-founder of Alexander Computing. He spoke at length about cloud computing and the concept of server-less architecture.
While the talks were strictly technical, they saw a decent participation level with eager and curious audiences. MUPy was successful in displaying the variety of ways in which Python can be used to create projects. The way in which the topics were presented made them quite easy to understand, even for a novice. The conference provided a launching pad from which those interested can enter the world of Python.