Network Intelligence/Communication AI Beginner's Guide 002--The Correct Way to Open the Treasures of Open Innovation Platform--Task Ranking

On August 30, China Mobile released the "Smart Network National New Generation Artificial Intelligence Open Innovation Platform". The platform serves the research and development of network intelligent applications and provides the necessary research and development for typical tasksEnvironment, data, code and evaluation. As a senior R&D in the industry, you must go up and experience it.

Log in to the websiteChina Mobile - a new generation of artificial intelligence open innovation platform for smart networks or search on Bing BingSmart Network Open Innovation Platform or Open Innovation Platform Data Set Just look for the official website of China Mobile.

front page

The home page introduces the overall architecture, capabilities, cases, and ecological status of the platform.

Click the top button "Task" to enter the ranking task introduction page.

Task

You can see several current typical tasks. It covers network element intelligence and operation and maintenance intelligence. Here we take wireless fault root cause location as an example for trial use.

click to enter

1. Task introduction (locating the root cause of wireless faults)

A selection of mission descriptions can be seen. I learned it from two 3-field data tables.

  • Alarm
  • Trouble ticket

Determine the type of failure.

2. Start the instance and enter learning

Click to start the task and enter login registration when not logged in. If you are logged in, enter theinstance configuration page to supplementinformation

After filling in the required items according to the prompts, click Security Verification in the lower right corner. After completing the security verification by dragging the image, click New Instance. Enter the instance page (it takes a few seconds to configure resources), click Start to enter the instance.

The name is very important and must be highly identifiable. If you participate in the ranking list later, it will be easy to see your ranking.

3. Enter Jupyter and quickly learn to run it

After everything is successfully verified, about 5 seconds later, create and start an instance, and enter Jupiter

After entering Jupiter, drag the right laucher to the bottom and double-click to open Terminal (1 in the picture). See the folder with the same name as the task on the left, click to enter (2 in the picture), all are taught in a fool-proof way.

All kinds of tutorial manuals are available. No need to explain, you also know that you can double-click to open the text file as shown in the figure, copy a line of content to the previously opened Terminal, and press Enter to run it. According to the prompts, all environment preparation, data processing, training and inference can be completed in about 10 minutes.

During operation, we can enter the code folder on the left, and the complete fully annotated code can be viewed and studied here.

After the operation is completed, you can submit to the ranking list according to the prompts.

4. Submit reviews and obtain rankings

On the new page, click Submit Results, Upload, Select Results, and then OK. After about 5 minutes, your nickname will appear in the rankings.

Summarize:

Proceed as follows:

  1. Open the web page to register and log in.
  2. Select the task, start the instance, and fill in the necessary information (The focus is on the identifiable school and nickname)
  3. Fool's tutorial, copy and paste, run baseline code. It takes about 10 minutes to learn the code instructions and understand the project process.
  4. Submit a review to get ranked.

In just 4 steps, you can gain project experience in half an hour and learn project code and materials quickly. If you are unsure, you can contact the WeChat customer service assistant (cmri_open) for consultation and guidance.

Related reading:

Beginner's Guide 001--The correct way to open the treasure of the open innovation platform--Using data sets [Technical articles in the previous issue]

Next time I will introduce how to use CSI compressed feedback data for task ranking.

Guess you like

Origin blog.csdn.net/TELCOM17AI4NET/article/details/134374136