Surya Prakash Pathak

There are over 5 trillion megabytes of data on the internet, and private information and data from phones and laptops are all over the internet. We often tend to accept the privacy policies of various websites without even looking and hence causing a transfer of information to the world. However, some websites and platforms allow you to anonymize your personal information and still allow these websites to make inferences and analyze the data via Data anonymization. Using this capability of securing and ensuring almost encrypting personally identifiable data in a dataset, we can make the data live in the open source world. Such is the concept of Homomorphic Encryption, it allows us to eliminate the tradeoff between data usability and privacy, and keep it safe, secure, and private even in the most untrusted environments, like public clouds or external parties. In this session, we will cover what is Homomorphic Encryption and how this can change the outlook on Open Source Data. We will also demonstrate the intersection of AI and how holomorphic encryption can enable multi-party data sharing.

Other people: AAKANKSHA DUGGAL

We are living in a digital era where vast amounts of data is constantly being generated, evaluated, and updated. As a result, the need for enterprises to keep up with this pace has grown and we are rapidly moving towards a more data-driven society. With the help of AI/ML technology, we have the power to make knowledgeable data driven decisions and effectively identify new trends and patterns, leading to more creative solutions and innovative approaches to problem-solving. In light of the recent advancements in AI, particularly in predictive modeling, we now have a powerful tool at our disposal to quickly consume and analyze vast amounts of data. By using open source time series forecasting ML models like ARIMA and Prophet, we can provide more accurate predictions and insights in real-time, enabling organizations and teams to streamline processes and increase efficiency, improve and manage customer risk, and adapt to changing market conditions. In this talk we will discuss: 1. Open Source tooling for building predictive ML models (Python, Jupyter, MLFLow) 2. Time series forecasting techniques 3. Tips for managing ML workflows and model interpretations Attendees will leave this talk with a deeper understanding of predictive ML models and how open source can empower us to be more data driven.

Other people: Hema Veeradhi