Live playback including PPT download | Build Online Deep Learning based on Flink & DeepRec

Follow Ali Lingjie's official account and continue to obtain exclusive technical information!Reply "0723" with a private message in the backend of WeChat, and you can get the special PPT of Alibaba Cloud's "Building Online Deep Learning Based on Flink & DeepRec".

Event Background

The new social model and economic form in the post-epidemic era will inevitably give birth to new business models. The traffic of online business and related application scenarios is developing rapidly. Conventional offline systems and offline machine learning platforms can no longer meet business development requirements..

Compared with offline machine learning, online machine learning has better performance in the timeliness of model update, model iteration cycle, and business experiment effect.Therefore, migrating machine learning from offline to online has become an effective means to improve business indicators.

The following is the relevant sharing of building Online Deep Learning based on Flink & DeepRec. Interested partners are welcome to pay attention to Ali Lingjie and get a special PPT.

Introduction to the topic

Topic: Flink ML: DataStream-based Iterative Engine and Machine Learning Algorithm Library

Video:

Click me to watch live replay

Instructor:

Gao Yun, Alibaba Cloud technical expert

Zhao Weibo Alibaba Cloud Algorithm Expert

Speech Outline:

1. Iteration engine based on DataStream

Why develop an iterative engine

Introduction to the principle of iterative engine

Introduction to iterative engine using API

Future development plan

2. Flink ML machine learning algorithm library

Introduction to Machine Learning Algorithm Library

Introduction to the function and performance of the algorithm library

Future development plan

Topic: Flink ML: Design, Construction and Planning of Real-Time Machine Learning Scenario Solutions

Video:

Click me to watch live replay

Instructor:

Zhou Yunfeng Alibaba Cloud Development Engineer

Huang Xingbo Alibaba Cloud Senior Development Engineer

Speech Outline:

1. What is Flink ML and why build Flink ML

2. Use Flink ML to build machine learning scenario solutions

API design of Flink ML

The flow of Flink ML training and deployment model for inference stream processing

Build end-to-end solutions using the Pipeline/Graph API

3. Introduction to Flink ML Python

Introduction to Flink ML Python API

Use PyFlink + Flink ML to build machine learning applications

4. Ecological construction of Flink ML

Independent codebase and documentation site

flink-extended ecological project

5. The future development direction of Flink ML

Support online algorithm, support stream batch integrated data

Algorithm performance optimization

Rich feature engineering algorithm library

Topic: DeepRec: Large-Scale Sparse Model Training/Inference Engine

Video:

Click me to watch live replay

Instructor:

Ding Chen Alibaba Cloud Technical Expert

Speech Outline:

1. Background of DeepRec

2. Key Features of DeepRec

3. DeepRec future plans

Topic: Online Deep Learning PAI-ODL

Video:

Click me to watch live replay

Instructor:

Peng Tao Alibaba Cloud Technical Expert

Speech Outline:

1. PAI-ODL Architecture

2. Key Technologies of PAI-ODL