Keywords in this issue: EMU shower, hot standby train, train driver's cab, seamless line, image segmentation, automatic labeling
1 Organize the list of public accounts involved
1.1 Industry category
- RT Rail Transit
- People's Railway
- World Rail Transit Information Network
- Railway Signal Technology Exchange
- Beijing Railway
- Rail transit network
- on the list
- railway view
- ITS World
- Rail Transit Alliance VSTR
- Railway and Urban Rail Transit RailMetro
- track world
- things about the railway
- Railway Technology Innovation
- Smart transportation
- RTAI Smart Urban Rail Network
- Rail Transit Intelligent Equipment
- NE Rail Transit
- rail power supply
- Shanghai Railway
- Lanzhou Railway
- China Local Railway Association
- Railway Editorial
- Orbit Tribe
- modern rail transit
- 1435 Factory
- city subway
1.2 Artificial Intelligence
- Jiang Dabai
- rough technology
- Shinchigen
- AI Technology Review
- Smart stuff
- qubit
- i love computer vision
- Extreme Vision Platform
- artificial intelligence scientist
- AI Frontline
- When traffic meets machine learning
- computer vision life
- new machine vision
- Lei Feng Net
- Artificial Intelligence Industry Development Alliance AIIA
- Heart of Autopilot
- note man
2 digital tracks
(1) EMU "bath artifact", learn about it
Guide: There is a lot of wind and sand in the northwest region. After the EMU completes the transportation task, the outside of the car body is like wearing a layer of earth-colored armor. It is strong and thick. How to clean it quickly and easily ? Today, the editor takes everyone into the Lanzhou West EMU Station of Lanzhou Depot to see how the EMU "baths".
(2) Have you ever taken a "super train" that can replace any train?
Guide: Friends may not know that trains also have substitute players, just like substitute players on the sports field, substitute trains are also ready to play at any time, and they also have a unified name: "hot standby"! This "super train" that can replace any train is ready to leave at any time, but they never look forward to playing, because the safe operation of the train ahead is what they hope for.
3 track knowledge
(1) [Information] The main management structure of my country's railways
Guide: This article summarizes the management departments, operation departments, and construction departments of my country's railways, and sorts out the functions and jurisdictions of each department unit, which is helpful for quickly understanding the main management structure of my country's railways.
(2) Things only subway drivers know
Introduction: The subway is one of the most common and mainstream modes of transportation for our modern people to travel. When traveling by subway, many people may not be able to help but wonder: How is the subway put underground? Does the subway also have a driver? How does the subway turn around after reaching the terminal station? On May 15, CCTV popularized these "things only subway drivers know".
(3) Train fans | What "treasures" are hidden in the train driver's cab
Guide: "Dig, dig, dig in a small garden, plant small seeds and bloom small flowers..." This is an online nursery rhyme that has recently become popular on short video platforms. I believe many people know this nursery rhyme . Today, let us "dig and dig" the train driver's cab to see what treasures are hidden inside .
(4) How to divide subway and light rail? National standards tell you
Guide: Should a new rail transit in a city be called a subway or a light rail? This has long been a contentious topic. In order to unify the national urban rail transit classification, the General Office of the Ministry of Housing and Urban-Rural Development issued a notice to solicit public opinions on the national standard "Urban Rail Transit Classification (Draft for Comment)". The "Draft for Opinions" clarifies that subways and light rails are different in terms of capacity level, travel speed, optional models, and laying methods. The full text of the standard can be downloaded by replying to "Urban Rail Transit Classification" in the background of the official account "Zongeryanzhi".
(5) Stress release of seamless lines
Guide: This video introduces the principle of seamless line, temperature stress calculation, locked rail temperature, displacement observation method, stress relief method, daily maintenance and repair precautions, etc., which can quickly grasp the knowledge of seamless line.
4 Algorithm Research
(1) AI algorithm theory - image segmentation
Guide: This article gives a detailed introduction to image segmentation tasks, algorithm development history, application scenarios, algorithm classification, classic algorithms, evaluation indicators, open source data sets, popular papers, popular warehouses, etc., and the summary is relatively comprehensive.
Guide: The field of computer vision is also ushering in the trend of general models. With the improvement of the generalization ability of the model in the field of computer vision, it is expected to promote the development of general multi-modal AI systems. field is applied. This paper presents recent developments in segmenting large models.
(3) Farewell to the era of manual labeling | The latest SOTA-level semi-automatic labeling tool
Guide: This article will introduce the combination
Label-StudioandSAM (Segment Anything)provide a semi-automatic labeling solution to help you fully improve the efficiency of data labeling.
(4) Common distance algorithms and similarity calculation methods
Guide: This article sorts out common distance algorithms and similarity (coefficient) algorithms, and compares the differences between Euclidean distance and cosine distance.
5 AI Industry
Guide: The cost of using GPT-4 is 50 times that of GPT-3.5; and the time for a large language model to process 25 requests at the same time is only twice that of processing a single request... These data may sound a bit out of date Unexpected, but true. They are from a GitHub article called "Numbers Every Large Language Model (LLM) Developer Must Know". Only one day after the article was published, it received 1,200 stars.
Guide: Just now, according to the latest news from The Information, OpenAI is about to release a new open source large language model. OpenAI is open source again. Is it a "charity" from a major technology company or a "salvation" from an open source community? Most open source today still relies on giant models released by large companies with deep pockets. If OpenAI and Meta decide to shut down their businesses, the thriving open source community could become bust.
Guide: "If you don't go to Silicon Valley, you won't experience the feeling of the AI big model era." In November last year, the performance of ChatGPT surprised Li Zhifei, who once worked at Google headquarters and specialized in NLP (Natural Language Processing). Within a month, he set off twice to the frontiers of artificial intelligence, talking about large models with everyone. For AI entrepreneurs who have been frustrated for a long time, they are looking forward to a new stimulus too much. Li Zhifei, who gritted his teeth through the cold winter, said that he saw the future of AGI (General Artificial Intelligence) and became a firm believer.
