ML Monthly Meetup #21 (image processing and data analytics)

Welcome to the ML monthly meetup. We are excited to resume the in-person event and bring our members together to learn AI, ML, DL and Data technologies.

Food and refreshment will be provided and also win swags/prizes.

Agenda:
* 10:30am~11:00am: Checkin and Food/Refreshment
* 11:00am~11:40am: Tech talk 1: Time-state analytics with Timeline Framework
* 11:45am~12:25pm: Tech talk 2: A Primer into Image Restoration using Tensorflow
* 12:30pm~1:10pm: Tech talk 3: GADE:Generic Anomaly Detection Engine
* 1:10pm~1:30pm: Networking and closing

Tech Talk 1: Time-state analytics with Timeline Framework, by Devjyoti Patra, Principal Engineer from Conviva.
Across many domains, we observe a growing need for more complex time-state analytics, which entails context-sensitive stateful computations over continuously-evolving systems. For instance, in-video distribution, we want to analyze the total time video sessions spend in a buffering state.

Computing metrics such as connection-induced rebuffering require us to model the state machine of the player and the user to ignore buffering during initialization and after user seeks. Unfortunately, existing data processing systems, including streaming systems, time-series databases, and batch systems are ill-equipped to address time-state analytics. These systems do not provide native abstractions for modeling processes evolving continuously over time. Consequently, while in theory such systems can tackle time-state analytics, in practice they entail poor cost-performance tradeoffs and involve significant development complexity.

We have created a novel State machine abstraction for continuously evolving states and implemented a streaming system based on Akka streams. The system receives raw event data from video players and uses Timelines to derive metrics about each video session. I will be presenting this streaming solution for time-state analytics of telemetry data from Video players.

Tech Talk 2: A Primer into Image Restoration using Tensorflow, by Soumik Rakshit, Google Dev Expert and ML engineer from Weights & Biases
In this session, we will take deep dive into several image restoration models and how to solve them using Tensorflow.

Tech Talk 3: GADE: Generic Anomaly Detection Engine, by Iqbal Ahmad, Staff Data Scientist

Sponsors:

• Altimetrik sponsors the venue and food.
• Packt sponsors 5 Machine Learning books