China's artificial intelligence will face unprecedented development opportunities, and it will go down in history as a real solution to mankind's steel needs. We also look forward to working together with all partners here in Tianjin to truly use artificial intelligence to build a better world.
API interface data collection Mainstream e-commerce data collection
1. What is data collection?
Establishing an algorithm model requires using a large amount of annotations and good data to train the machine, so that the machine can learn the features to achieve the purpose of "intelligence". Data annotation helps machines learn to recognize features in data. For example, if we want machine learning to recognize cars, we directly give the machine a picture of a car and it cannot recognize it. We must annotate the car picture and label it to indicate "this is a car". When the machine obtains a large number of labeled cars, After learning from the picture, we give the machine a picture of a car, and the machine will know that it is a car.
2. What are the types of data collection?
There are many types of data annotation, such as text collection, picture collection, voice collection, portrait collection, etc. Let's take the common annotation business of Ant Xiongbing Company as an example to briefly explain the categories and uses of data annotation.
1. Image collection
Image collection is the most common form of collection and has lower requirements for collectors. Common collections include human body collection, Internet e-commerce platform product detail picture collection , review picture collection, etc., and vehicle collection, which are mainly used in human body recognition, object recognition and other fields.
2. Face collection
This kind of collection is not only limited to face collection, but also includes human body contour collection. The requirements are more detailed and there will be requirements for the location of each point. Mainly used in face recognition, human body recognition and other fields.
, duration 00:44
3. Voice collection
Voice collection refers to listening to a piece of voice, and the collector transcribes the content of the voice heard. Mainly used in the field of speech recognition.
4. OCR transcription
OCR transcription generally requires that the text in the picture and other areas that need to be transcribed be selected and the text in the selected part be transcribed. Mainly used in the field of text recognition.
5. Text collection
This type of project is generally to identify the category of sentences in the text, or to identify the emotion (positive, neutral, negative) contained in the text. Mainly used in fields such as intelligent customer service.
6. Collection projects
Collection projects are generally not carried out through the platform. Most of them are carried out offline, and the tools used are also quite diverse. Common types include voice collection, video collection, and face collection. The collected data is generally cleaned before being put into use.
In addition, there are many types of data collection, so I won’t introduce them one by one here. If you want to know more about them, you can pay attention to other articles.
When people talk about big data, artificial intelligence and other words, they often think of advanced content such as AI algorithms, data mining, and machine learning. However, no matter how good the algorithm is, it cannot be realized without the support of a large amount of accurate data.