The Association of Data Scientists (ADaSci), in partnership with Analytics India Magazine (AIM), recently concluded the 2021 edition of the Deep Learning Developers Conference (DLDC), held virtually on 23-24 September. The two-day long influential conference on deep learning featured 30 keynote speakers and over 500 attendees, where some of the leading professionals and researchers from 200 organisations presented their feature talks and research papers. A full-day workshop on different deep learning technologies was also held on Day 2 of the event, with certificates being provided to all the attendees. 

The conference was virtually held, yet allowing fellow attendees to network with each other, providing opportunities to meet companies and the chance to talk with the speakers. It covered both technical and business aspects of trending developments in the domain of deep learning and attracted developers and executives from all over India. The sessions held at DLDC helped uncover a wide range of different approaches to the problems that are currently faced by the tech industry in the areas of deep learning, artificial intelligence and machine learning. The workshops and research paper presentations also garnered a large number of attendees and focused on insights to the latest in deep learning with hands-on experience. 

DLDC 2021 was sponsored by some of the most highly known tech conglomerates. The presenting sponsor being TSW, Gold sponsors MIQ, Ugam, A Merkle Company, and Platinum sponsors American Express and Genpact. 

How It Happened

The exciting first day started with a highly anticipated and insightful session by Mohan Silaparasetty, Head of Technology Programs at Times Professional Learning, where he spoke on “The State of AI and Deep Learning”. This keynote talked about the latest advances in deep learning and its current applications in artificial intelligence worldwide. Mohan dived deep into the details of emerging trends and technologies in deep learning, which featured textless NLP and generative networks, to name a few. 

The morning session was followed by tech talks from Manoj Kumar Rajendran, Principal Data Scientist at MiQ Digital India, and another from Vikas Behrani, Vice President, Data Science, at Genpact. Manoj talked about understanding and leveraging differential data privacy and how the tech world can be prepared for a differentially private world of data. The following talk from Vikas Behrani elaborated details on the Lap Estimate Optimizer and how race-day strategies for Formula E can be transformed using AI. 

In a subsequent talk, Radhakrishnan G, VP and Global Head of Commercial Risk Decision Science at American Express guided the attendees on helping small businesses with real-time credit decisioning using ML and AI, where he also described how Amex is leveraging machine learning techniques to enhance marketing. This talk wrapped up the morning sessions at DevCon. 

The post-lunch sessions started with a talk on how to deal with data imbalance in classification problems by Raghavendra Nagaraja Rao, Data Science, Academic Lead at Times Professional Learning, where he enlightened the attendees on how different techniques can be used to deal with imbalanced data. 

Swagata Maiti, Technology Architect, IP & Data Products, and Shaji Thomas, Vice President, Cloud & Data Engineering at Ugam, A Merkle Company, later continued the post-lunch session with their presentation on seven techniques that help create a scalable data platform, answering the question on “To data prep or data science?”. 

Data Scientists at CRED, Ravi Kumar and Samiran Roy, later explained the essence of using graph neural networks and how the emerging technology is being utilised by CRED in their post-lunch presentation on deep learning applications with graph neural networks.

As a huge volume of information in an enterprise flows through documents, Rahul Ghosh, VP of AI Research and Services at American Express AI Labs, helped understand how AI-powered document intelligence for enterprises can drive innovation and efficiency at scale during his time at the DLDC 2021. Hitesh Prakash Nahata, Senior Manager of Advanced Analytics at MiQ, presented a detailed talk on driving incremental outcomes through hyper-relevance that can aid analytics and how to gain deeper insights using such methodologies. 

DLDC 2021 also touched upon a brief history and key trends in deep learning for Computer Vision by speaker Angshuman Ghosh, Head of Data Science at Sony Research India, where he shed light on Computer Vision technologies and some breakthrough models that led to massive progress in the field.

Other prominent names included Ram Seshadri, Machine Learning Program Manager at Google, who gave an insightful discussion on his Deep Learning AutoML library, Deep AutoViML. The agenda of the talk included top features of the library and how the library can be used in the context of MLOps. Kaggle Grandmaster and Senior Data Scientist Luca Massaron talked about his recent work on using deep learning techniques to predict credit ratings for global corporate entities in his tech talk on Deep Learning for Credit Rating. 

Exciting Workshops and Paper Presentations

The second day was filled with workshops and paper presentations, starting with a workshop on text classification with vectorisers and pre-trained neural net models by Raghavendra Nagaraja Rao. A line up of key paper presentations included: Time Expression Extraction and Normalisation in Industrial Setting by Piyush Arora, Senior AI researcher at American Express AI Labs, Classification of Quasars, Galaxies and Stars using Multi-modal deep learning by Bharath Kumar Bolla, Senior Data Scientist at Verizon, Hyper localisation of leaks in piping and cabling systems using reinforcement learning by Indrajit Kar, Head of AI at Siemens Advanta, Global-Local Scalable Explanations Using Linear Model Tree by Narayanan Unny E., Head of Machine Learning Research, American Express AI Lab, Predicting Custom Ad Performance Metric using Contextual Features by Prateek Kulkarni, Data Science Team Lead at MiQ Digital and Divyaprabha M, Data Scientist at MiQ Digital and  Analysis of Sectoral Profitability of the Indian Stock Market Using an LSTM Regression Model by Jaydip Sen, Professor of Data Science and Artificial Intelligence at Praxis Business School.

Dipyaman Sanyal, Head of Academics and Learning at Hero Vired, conducted a tech talk on “Explainable and Interpretable Deep Learning” and Jatindra Singh Deo, Senior Technical Architect at Genpact along with Abhilash NVS, Data Scientist for Genpact, conducted a two-hour detailed workshop on industrialising AI/ML: Hands-on Model Deployment, post-lunch at day two. 

During the two days of DevCon 2021, the speakers also interacted with the attendees by answering their questions and doubts regarding the presentation through a live interactive chatbox and post-presentation discussion being conducted after each talk. This helped the attendees not only connect better with the speakers but also gain deeper knowledge and understanding of the topic. 

Summing up 

The Deep Learning Developers Conference 2021 set an example of how even during times of a global pandemic, an insightful yet engaging symposium can be held virtually. The conference included some of the biggest names in the tech industry who covered almost every aspect of Deep learning. Speaking on the success of the event, Bhasker Gupta, Founder & CEO, Analytics India Magazine, said, “It was great to see such an overwhelming response knowing about tough times the whole world is going through. DLDC 2021 would not have been possible without the tireless efforts from the entire Analytics India Magazine team and the sheer support of our sponsors”.