Clustering is an unsupervised learning algorithm that has a lot of value to add when it comes to the grouping of data in a dataset. Entities in each group are comparatively more similar to entities of that group than those of the other groups. This talk explores network clustering as an approach for insights. We’ll explore real-world movie datasets — and show which actors are close to each other, and which film industries are most isolated. We’ll then see how network clustering can be applied to product networks, patent networks, and sports networks, and give you a comprehensive, industry-level understanding of the same.

Key Takeaways for the Audience:

• How to get insights and patterns from the data even when there are no numbers
• Understand network analysis and community structure detection
• Understand applications of Network Analysis in the industry

A common notion among starting data scientists is that the production deployment of the model is the end of the journey. In reality, there are never-ending ways in which the model can fail in production due to issues with the data, model, engineering, or human error. As business decisions are increasingly pivoting to being model-driven, the impact of these errors can have adverse real-world consequences. It is imperative to have a framework that continually monitors performance and technical debt and alerts the data scientist ahead/at the moment of potential failures.
This session is a practical what, why and how – of building such monitoring frameworks.

Key Takeaways from the session:

• Example-based walkthrough on the best practices in model monitoring – covering both engineering and data aspects
• How to set up a monitoring ecosystem, than encompasses sampling, setting up a tagging infrastructure, model versioning, and correct model interpretation
• Practical techniques and algorithms to detect data and concept drift that allow us to track and keep the model healthy over time

Data Pipelines have enabled us to unlock the potential of utilizing big data. Companies are now able to optimize data transfer while securing it at the same time. With data collection, transformation, and transfer reaching new highs it had become necessary for companies to optimize costs and thus, shift this process to the cloud.
In this deep dive, we will build an end-to-end data pipeline using Python and SQL on AWS. You will be learning about several key services such as AWS s3, DynamoDB (for bookmarking), Cloudwatch (for monitoring), Lambda, ECR, Glue Catalog, Athena, etc along with Docker
Key Takeaways:

• Architectural Overview of building end-to-end data pipeline using AWS Services
• Deploy Complex Python applications as Lambda Functions
• Build and Push custom Docker Images to ECR so that they can be used to configure Lambda Functions
• Relevance of Bookmarking and Job Run Details using tables in Dynamodb
• Configure s3-based Lambda Triggers

In this deep dive, we will create 2 web apps using TensorFlow JavaScript. The first will be a Deep learning web app to play PacMan live and the other will be an NLP app for Sentiment Analysis.

We will begin with TF.js initially, and introduce its APIs. Then we proceed to the introduction of computer vision, the basics of Deep learning, MobileNet model. In the next part, we will learn the basics of NLP with Deep learning. We will talk about the basics of NLP such as One hot encoding, multi hot, tf-idf, and advanced topics such as Word Embedding, and LSTM. And then we will proceed with creating the apps.

Key Takeaways:

• Deep learning with TensorFlow Javascript
• Computer vision with Deep learning
• NLP with Deep learning
• Advanced Deep learning concepts such as Word Embedding, Depthwise convolutions, Transfer learning, and more

AI continues to influence our lives. From the trivial task of typing on our smartphones to personalized recommendations on shopping websites, intelligent machines are everywhere. Our interactions with technology are increasingly pushing the boundaries of ethics in AI. We will introduce you to the concerns the industry foresees and how we as practitioners and organizations can tackle the questions about fairness and equality, privacy, exploitation, environment, sustainability, and more.

Key Takeaways

• Inspiring stories of AI
• The dark side of AI today
• What we can do to tackle some of these challenges
• What perspective should we consider?
• What can you do to activate responsible AI?