[monthly articles] May 2020 to June AWS essays on Artificial Intelligence

Time:2021-7-28

[monthly articles] May 2020 to June AWS essays on Artificial Intelligence

Unconsciously, we have shared with you Amazon Web services (AWS)   Relevant AI, machine learning, deep learning and other technical contents. Next, we might as well make a simple summary through this article, so that small partners with different needs can quickly find what they are interested in.

Amazon SageMaker

Traditional ml development is a complex, expensive and iterative process, and there is no integration tool available on the market for the whole machine learning workflow, which makes it even more difficult and discourages many developers. Amazon sagemaker was born to solve these problems. It is a fully hosted service that can help developers and data scientists quickly build, train and deploy machine learning models. How did you do it? Welcome to:How to efficiently build an ML model with Amazon sagemaker without spending a penny?

As a service that unifies all kinds of development tools required by ml, Amazon sagemaker helps developers write code, track experiments, visualize data, debug and monitor in an integrated visual interface, which greatly improves developers’ work efficiency. How can this service be started quickly? Welcome to:How to get started with Amazon sagemaker quickly?

Named entity recognition is a very important technology in the process of machine learning. It can be used to identify entities with specific meaning in text, such as person name, place name, organization name, proper noun and so on. For English, named entities in English have obvious formal marks, so entity boundary recognition is relatively easy; However, the task of Chinese named entity recognition is more complex, and the entity boundary recognition is more difficult than the entity category annotation subtask. How to solve this problem with Amazon sagemaker? Welcome to:Compared with English, what is the difficulty of Chinese named entity recognition? How to solve it?

Machine learning technology enables many cumbersome and time-consuming operations to be realized automatically, greatly improving work and life efficiency. However, in order to achieve this, we first need to have a sufficiently high-quality ML model, which is not easy to obtain. Around this problem, the recently released Amazon sagemaker autopilot can automatically create the best classification and regression machine learning model under complete control and visibility, helping us easily create high-quality ml models. Welcome to:Automation goes further, and ml models can be generated by themselves.

Deep learning

Because it can extract complex patterns from all kinds of complex data (such as free text, image and video), deep learning technology has been widely used. However, in practice, it will be found that many data sets are easier to express in the form of graph, such as the relationship between people on social networks. In fact, with the open source deep graph library, anyone can conduct in-depth learning on graphic data through Amazon sagemaker. What exactly should I do? Welcome to:What should graphics based deep learning do?

Data received at different times and used to describe the change of one or more characteristics over time can usually be called time series data. With this data, we can predict the future through the time series data generated in the past. What are the types of forecasts? Which algorithms can be implemented? Welcome to:With the help of sagemaker to build a graphical neural network, the deep learning ability reaches a new high.

Deep learning requires a lot of data and computing resources, and takes a lot of time to train the model, but it is difficult to meet these needs in practice. The use of migration learning can effectively reduce the amount of data, computing and computing time, and customize the business requirements in new scenarios. In this article, we use our own data to fine tune a pre trained image classification model in Amazon sagemaker based on mxnet, and build a model number classifier with high accuracy. How exactly? Welcome to:It’s not difficult to read the pictures and write such a program yourself.

When shopping on the e-commerce platform, everyone has seen the personalized recommendation system. This kind of system is an advanced business intelligence platform based on massive data mining, which can help e-commerce websites provide fully personalized decision support and information services for their customers’ shopping. In fact, with the help of the open source deep learning library gluon launched by Microsoft and Amazon, we can easily develop such a personalized recommendation system. Please read the specific methods:“Guess you like it”? Find out the routine and build a DIY recommendation system yourself.

The explosion of deep learning makes python, a once small language, suddenly hot. In contrast, although Java is still one of the most popular languages, with a large number of developers, especially the most widely used foundation in the enterprise market, it is difficult to find a suitable Java tool or framework for in-depth learning. The emergence of deep Java library has changed this situation. Welcome to:After the waves, the heat is infinite, and before the waves, Java still has great prospects.

machine learning

Machine learning has developed into a multi domain interdisciplinary subject in recent 30 years, and is widely used in data mining, computer vision, natural language processing, biometric recognition, search engine, medical diagnosis, detection of credit card fraud, securities market analysis, DNA sequence sequencing, speech and handwriting recognition, strategic games and robots. In order to help you solve these practical problems more conveniently, AWS has launched a variety of machine learning tools, including ml service class, API class, AI service tool class, etc. What scenarios are these tools suitable for? Welcome to:Great AWS machine learning toolkit summary.

Generative adversarial networks (GAN) is a generative machine learning model, which can be used to create fictional characters and scenes, simulate face aging and image style transformation, and even generate chemical molecular formulas. In this article, we introduce an idea that can use this kind of technology to imitate handwritten fonts with machine learning methods. Welcome to read:The “portrait” without portrait right is generated by yourself. You can use it at will. You’re welcome!

Amazon quicksight is a powerful service that can fully integrate massive data from different sources and formats, and then reveal the business insights contained in the data to users through machine learning based analysis and graphical reports. What’s the use of Amazon quicksight? How to get started? Welcome to:Amazon quicksight users see here: ML insights is officially released, and more business insights are within reach.

The training of ML model requires a lot of time and computing power, which I believe we all deeply understand. Since cloud computing is so popular now and resources such as computing, storage and database in enterprise application environment can run in the cloud, can we find a way to throw the most time-consuming model training work into the cloud for machine learning projects? Just run distributed tensorflow training with Amazon sagemaker. Please read:Just focus on the ML algorithm and leave the training of the model to it.

In order to combine machine learning with data in relational database, it is usually necessary to develop a custom application to read the data in the database, and then apply the machine learning model. Developing such applications requires a variety of skills to interact with databases and machine learning. So can we more easily apply machine learning to data in relational databases? There are still some methods. Welcome to read:Do not do data “Porter”, also play machine learning.

Many times, managed services can help us save a lot of things. From application development to deployment and subsequent operation and maintenance; From the creation of machine learning model to training to final operation… In many steps, we can choose managed cloud services to help us reduce development costs and workload and focus more on really important tasks. In fact, the data processing and model evaluation in the process of machine learning can also be done with the help of managed services. Welcome to:Play with ml and never do it yourself if you can manage it. Today, it’s the turn of data processing and model evaluation.

Machine learning is a highly iterative process. In the course of a single project, data scientists and ml engineers often train thousands of different models for maximum accuracy. In this process, the number of combinations of algorithms, data sets and training parameters (also known as hyperparameters) involved is unlimited… We need to find the optimal solution just like looking for a needle in a haystack. The newly released Amazon sagemaker experiments can help users organize, track, compare and evaluate machine learning experiments and model versions, and simplify the creation and optimization of models. Welcome to:The big data explosion has big data. What about the ML model explosion.

Amazon sagemaker model monitor is a new function of Amazon sagemaker. It can automatically monitor the machine learning model in production and send an alarm when there is a data quality problem, so as to help us quickly understand the quality problems in ML model. How is this function used? Welcome to:This class representative, the important task of supervising everyone’s study in the future will be handed over to you.

artificial intelligence

AI has been widely used, but if asked, how can you use the most understandable language to make it clear to people who don’t know anything about technology? Welcome to:Tell the “little beast” around you what AI is in easy to understand language.

All along, AWS has been trying to provide users with pre trained artificial intelligence services through innovations in the fields of computer vision and language, so as to help users use it smoothly without having professional knowledge in the field of machine learning. On this basis, the amplify framework introduced in this article helps developers add and configure artificial intelligence / machine learning capabilities to any web or mobile application with just a few lines of code. What exactly should I do? Welcome to:By doing so, your application will learn the ability to “learn”.

No matter in work or life, AI can provide us with great help in many aspects. However, have you ever thought that with the help of advanced AI capabilities, each of us can become a composer and write beautiful music with the help of AI. Welcome to:When music meets artificial intelligence – singers are performing in “cloud competition”, composers don’t want to try “cloud composition”.

This paper constructs a solution to make full use of AWS services to help us easily create an automobile product user manual with interactive ability to enhance the presentation of display technology. How powerful is this technology and what innovative experience can it bring to consumers? Welcome to:When AI is mature, write the AR product manual yourself.

[monthly articles] May 2020 to June AWS essays on Artificial Intelligence