Analytics India Magazine is hosting one of India’s biggest Machine Learning conferences, scheduled to occur on January 30th and 31st. The Machine Learning Developer Summit aims to bring together data scientists and machine learning developers from all across the country. With the presence of top ML innovators from leading tech companies, MLDS is all set to talk about the most recent developments in machine learning and its accompanying use cases.
While attendees will not only learn a lot from these ML experts, they can inculcate important skills referring to data sciences and interact with like-minded individuals to add value. With that being said, here are 10 sure shot takeaways for every professional attending MLDS 2019.
1| Foundation Of ML – Building Your Data Lake On AWS
In a talk by Raghuraman Balachandran, the Solutions Architect at Amazon Web Services, attendees will learn about how to build clean data lakes on the software. Since they are one of the most common architectures required for data-driven organisations, this will provide an opportunity to have a hands-on experience on the concept, thereby increasing the chances of hiring.
Over the course of the talk, Mr Balachandran will speak about how to store unstructured, semi-structured, or fully-structured raw data as well as processed data for different types of analytics. He will also provide his experienced perspective on common challenges and patterns for designing an effective data lake on the AWS Cloud.
Key Takeaways: Learn how to harness AWS’ power for the creation of a dependable and useful data lake, straight from the horse’s mouth based on experience with clients.
2| Accelerate Inference For Deep Learning Models
Deep Learning models are extremely resourced intensive, with requirements for high-end GPUs and CPUs for optimal training. However, this still creates opportunities for clashes, as the target deployment environment introduces various challenges that are typically not present in the training environment.
In this talk, Sunil Kumar Vuppala, the Principal Scientist at Philips Research, will speak about methods to increase the performance of trained DL model during the inference phase. He will also present Intel’s inference engine as one option to run DL models in Caffe/ Tensorflow/ Keras using CPUs.
Key Takeaways: Learn to use various methods to improve DL models’ performance during the inference phase using Intel-powered tech to ensure the quality of output in the final product.
3| With The Explosive Growth In Demand For AI Skills, How To Be More Employable
Due to the highly disruptive nature of AI, many analysts expect explosive growth in demand for AI skills. However, this is sorely lacking, as it is not occurring at the anticipated rate. Amitabh Mishra, the CTO at Emcure Pharmaceuticals, will speak upon his reflections in the AI field and why it is not occurring. He will also offer tips for preparation for AI professionals, and elaborate upon the requirements in AI and related fields.
Key Takeaways: Learn how to be more employable as an AI professional and how to evolve in relation to evolving market conditions.
4| Walmart AI Journey – Use Case For Building Technologies For Merchants
This talk by Prakhar Mehrotra, Senior Director at Walmart Lab, will offer a perspective into Walmart’s movement in India. Walmart Labs is a subsidiary of Walmart which is aimed at providing a better experience for customers of the chain worldwide. The organisation also builds tech for use with Merchants to increase Walmart’s range.
Mr Mehrotra will speak about Walmart’s journey into the field of AI and offer insights into the advancements that the chain has made thus far.
Key Takeaways: How to build AI and implement it at scale for use in low-infrastructure environments.
5| How Important Is Data Cleansing And How ML Can Help In It?
A clean data lake is important for the production of a good ML model. Around 40 percent of the time and resources spent on creating a new ML model comprise of data cleansing. Somu Vadali, Chief of Data and Products, CnD Labs at Future Group, will speak about how data cleansing can be done more efficiently with ML.
He will also elaborate upon the need for the method, and streamlined, thought-through processes that can enable organisations to go quickly and reliably from raw data to features. His talk will also help organisations cut down their time-to-market with every subsequent new model or version of the existing model.
Key Takeaways: Learn how to establish methods for clean data, thus reducing time to market and increasing the efficiency of model creation, and other industry-best practices like data quality, validation, guardrails and more
6| The Use Of Augmented Genetic Algorithm
As every data scientist is aware, feature selection in large dimensional spaces is a staple of machine learning. Therefore, Sidharth Kumar, the Staff Decision Scientist at e-tailer Flipkart, will speak about a proposal to use genetic algorithms to solve this problem.
He will also offer transparency into his efforts to significantly reduce the number of generations required to convergence.
Key Takeaways: Demonstration by Kumar on the use of this algorithm in conjunction with SVMs, solving two problems with one solution.
7| A Primer On Building And Deploying Scalable Analytical Models In Production
Subramanian MS, the Head of Analytics at BigBasket.com, will reveal his methods to plan a scalable set of tools for ML models. This includes planning a lifecycle for analytical models, a scalable software and hardware layer, along with industry best practices for scalable DevOps.
Key Takeaways: Learn how to scale your startup/entrepreneurial idea effectively at every level from a company undergoing growth at scale.
8| How To Explain ML Models
ML models have been historically difficult to explain or ‘pitch’, owing to the fact that they are black box algorithms. Owing to the increasingly obsolescent state of linear models to solve predictive modelling problems, we are now adopting ML models. This makes it an important part of the evolving corporate setting.
Indranath Mukherjee, the Head of Strategic Analytics at AXA XL, will speak about how the analytics industry has begun to move to ML models for said problems because of their low bias and higher predictive power.
Key Takeaways: Learn how to pitch an idea for an ML model by explaining what is effectively a black box algorithm for use in a business setting from one of the keynote speakers.
9| How AI Is Shaping The Future Of Video Streaming
Over the top streaming has effectively disrupted existing technologies such as cable television, as seen by the meteoric growth of companies such as Netflix. Anubhav Shrivastava, the Head of Product Data Sciences at Voot, will speak about how AI techniques are disrupting this business itself. He will also offer examples of such implementations, such as image and video recognition, recommendation engines, media databases, and metadata generation.
He will also reveal information on plans for intelligent adaptive bitrate streaming, self-reliant marketing automation plans and predicting catastrophic failures.
Key Takeaways: Take a look into how a quickly growing OTT media platform is utilising AI, and how disruptions can occur in media, along with technical details.
10| Named Entity Recogniser – Practical Challenges And Strategies
This talk by Mathangi Sri, the Head of Data Science at PhonePe, will reveal how supervised algorithms are solving efficiently most of the classification problems in the text. However, this also comes with its shortcomings, such as entity recognition. Entity recognition is still an area of craft and domain knowledge, thus making Sri’s talk more valuable.
The speech will look at how to handle entity recognition, including the stack of techniques, their applications, strategies and real-life challenges while applying those.
Key Takeaways: An opportunity to gain a skill set regarding entity recognition, a closed field. Examples might be provided in an indigenous setting, seeing as the talk is by PhonePe.