Analytics India Magazine has just concluded its day two of its inaugural virtual conference Plugin. Organised by AIM in collaboration with ADaSci, Altair, AnalytticaDatalab, SAS and Kyvos Insights, this 2-day online conference connects developers & leaders in the data science & machine learning space from across the world and offers and experience that is rich in knowledge and on par with in-person events.
Here are a few highlights from the action packed day two of Plugin:
Julia Computing Workshop By Viral Shah
Day 2 of Plugin was kicked-off by Julia co-founder Viral Shah. He introduced the audience to why Julia computing is gaining a strong foothold in the ML community. He briefed on how large scale organisations such as the NY Federal Bank have implemented Julia programming language for one of their applications. Julia has found its use in personalised medicine and even mapping of the universe. “Think of getting 100x more performance from your laptop. That is what Julia is all about, ” emphasised Viral. This was followed by a hands on 60 minute blitz developers by Dhairya Gandhi of Julia, which was later continued by Deepak. Abhijith Chandraprabhu, data scientist at Julia had a great presence throughout the workshop with his consistent answering of the numerous questions that were being asked by the excited crowd.
Identifying Model Drift Before It Is Too Late By Jacqueline Long Of SAS
Jacqueline Long, Principal Solutions Architect at SAS, spoke about how business decisions that rely on analytical models could be suffering from model drift. The real impacts are likely to be flawed decisions that affect the bottom line. Jacqueline pressed on the fact that contactless commerce is reshaping consumer behavior to adjust to the global COVID-19 situation. The seamless functioning of any ML pipeline requires selecting the right model and robust model management later. Jacqueline gave a demonstration on how tools of SAS model manager can help in assessing and comparing candidate models for champion model selection.
Talk on AI Driven Blueprint For Enterprises By Johnson Poh
In this talk, Johnson Poh, Head, Group Enterprise AI at UOB Bank discussed the essentials of artificial intelligence and its role in this new paradigm. Poh spoke about the relevance of emerging technologies, how to implement a data-driven strategy and how AI changed the financial industry, among others.
With his presentation, Poh addressed the following fundamental questions:
1. What is AI, Big Data and Analytics? Why are they relevant?
2. How to implement a data driven strategy taking on the lens of people, process and technology?
3. More specifically, what are the technical components required to build an end to end data driven pipeline?
Data Analytics in the Smart Factory By Sudhir Padaki Of Altair
In this talk, Sudhir Padaki, Director of Business Development at Altair, spoke about how to access data from the shop floor through OT/IT convergence solutions and build useful analytics to tackle Asset Monitoring, Predictive Maintenance. He touched upon the emergence of the fourth Industrial Revolution and why there is a growing interest in Analytics and how this is revolutionizing decision making in Smart Factories. He also discussed how one can leverage analytics on a shop-floor in real time to identify hidden indicators for future downtimes or identify anomalies. Sudhir also gave a demo on how Altair Panopticon assists in seamless integration of data pipelines.
Approaching (Almost) Any NLP Problem By Abhishek Thakur
This session by 4x Kaggle Grand Master Abhishek Thakur delivered a very informative talk where he covered all the corners of NLP. From preprocessing operations such as tokenisation to running BERT model to identify duplicate questions on Quora, Abhishek presented the audience with many state of the art approaches that are currently in use at the industry level. The talk was very informative and serves as a right start to beginners who would like to tackle NLP based problems.
Workshop By Dr Raul Rodriguez
In the last workshop of the Plugin conference, Dr Raul Rodriguez delivered a very informative session that covered the very tricky yet challenging application of AI for hiring and tracking employee performance. Dr Raul didn’t hide the fact that countries like China have taken this application a bit too far with their social credit system. However, he took a more pragmatic and less political approach and explained how challenging it is to build a model in the Chinese scenario. He later went on to explain how the Big 4(Google, Microsoft etc) are increasingly using custom built AI to hire candidates. In a follow up to this, Dr Raul gave a hands on session on how to build such models using state of the art techniques.
Data Literacy By Kirk Borne
In one of the last talks of this conference, Kirk Borne, a very well known name in ML community, spoke about the need for data literacy and why organisations need to pay attention. He covered five major themes: data awareness (what is it?), data relevance (why me?), data literacy (show me how), data science (where’s the science?), and the data imperative (create and do something with data). He has also discussed why it is important for data scientists to lead the efforts to build data literacy in society, in schools, and in professional development activities for organizations.
The day 2 of Plugin witnessed more than 20 sessions with talks from experts across the world. While day one has been a great success, day 2 too, has witnessed exceptional talks from the pioneers of the field. Pulling off a conference on a global scale during a pandemic is a challenging affair but thanks to the speakers, sponsors and most importantly the audience, Plugin received a tremendous response.