A month ago, we held a tribute to the biggest cloud computing event of the year: AWS re:Invent 2019. Together with AWS – Amazon Web Services, Micropole organized the AWS re:Cap session at Brussels to inform everyone of the biggest announcements, share the latest technologies and explain interesting use cases to those interested.
But now, a month later, who still remembers all that was said by experts Sébastien, Thomas, and Thierry? That’s why we’re launching a small series consisting out of 3 technical articles, focusing on all the insights and use cases that were highlighted at the AWS re:Cap in February. Are you ready?
Alright then! Let’s start with ‘Part 1: Artificial Intelligence & Machine Learning’.
Creating a notebook is complicated, so AWS has introduced the serverless notebooks. With that, AWS manages the integration for you. You can share it without having to set up or manage things by yourself. It’s a plugin on the left side of the studio. Find out more here
Amazon SageMaker Processing introduces a new Python SDK that lets data scientists and ML engineers easily run preprocessing, postprocessing and model evaluation workloads on Amazon SageMaker. Collect the data from different data sources and bring it directly to where the next step in your process can find this data. From there you can start to train the model based on processing. This service is fully managed. Read more here.
Amazon SageMaker Experiments lets you organize, track, compare and evaluate machine learning (ML) experiments and model versions. The goal of SageMaker Experiments is to make it as simple as possible to create experiments, populate them with trials, and run analytics across trials and experiments. As you would expect, SageMaker Experiments is nicely integrated into Amazon SageMaker Studio. You can run complex queries to quickly find the past trial you’re looking for. You can also visualize real-time model leaderboards and metric charts. Read more here.
AWS Sagemaker Debugger automatically identifies complex issues developing in machine learning (ML) training jobs. When you train a model, it can take days up to hours or even weeks. Like Alexa, for instance, can take up to five days to learn a new model. So when your model building isn’t working well, instead of waiting and at the end notice one parameter is not correct, we can instrument the training inside the framework, to be informed of the status of the model while it is being trained. Debugger allows us to view inside and see what happens when a machine is learning a new model.
Isn’t that just the coolest? Find out more here.
Amazon SageMaker Autopilot automatically trains and tunes the best machine learning models for classification or regression, based on your data while allowing to maintain full control and visibility.
You tell AWS SageMaker “Hey my data can be found there, this is the format.” and then you can tell pilot which column to optimize and based on the type of optimization you want to do, AWS SageMaker will then choose what the best experiment is. And so it goes, it will be put in production. You will also have a detailed decision tree on why SageMaker chose this specific experiment, so you’re not stuck with a ‘black box’ at all. You’ll receive a fully detailed notebook so you can go over every detail and understand why.
Sounds pretty awesome right? Read more here.
AWS DeepRacer gives you an interesting and fun way to get started with reinforcement learning (RL). RL is an advanced machine learning (ML) technique that takes a very different approach to train models than other machine learning methods. Its superpower is that it learns very complex behaviors without requiring any labeled training data, and can make short term decisions while optimizing for a longer-term goal. Simply put: DeepRacer is a tool/toy to help developers work on their machine learning skills.
A little historical lesson: 3 years ago you could build models to recognize pictures and then you’d connect it to the machine. In Seattle, they used this to create a system to get their machine to recognize if their cat was coming back to the house alone or if it was bringing back a dead animal with it. Based on the cat’s choice of bringing back a certain present or not, the door would open or remain closed. Handy right?
This year it was all about races. You had to write code, that delivered solutions that would receive ‘good’ points upon ‘good’ actions and ‘bad’ points for ‘bad’ actions. The system will then learn what a good or bad action entails, following the consequences: namely by positive reinforcement for good behavior and punishment for bad behavior. You then had to put all of that in a simulator and then you’d let it train. By the end of it, your car should be able to drive a race on itself. It was a big success. And for the first time in 2018, a competitive race was launched. A local race was held at each local summit and the winners were invited to Las Vegas for one final race to determine the winner.
For next year AWS chose to go with a new idea: the deep composer. AWS would like to help you to learn generative adversarial networks or GAN. Generative adversarial networks (GANs) are a class of neural networks that are used in unsupervised machine learning. A GAN has two parts in it: the generator that generates images and the discriminator that classifies real and fake images. Like deep-fake video’s, for example. This one, proposed by AWS, allows you to play music. You can train your model to create your style. Create your melody, choose your style and then the system will create an orchestra around you. Go and have fun with it!
Amazon Fraud Detector is a fully managed service that makes it easy to identify potentially fraudulent online activities such as online payment fraud and the creation of fake accounts. Simply put: Amazon Fraud Detector’s idea is to create models that detect fraud for your transactional applications. This way the system can detect if a certain application is fraudulent or not. You can create a fraud detection model with just a few clicks and no prior ML experience because Fraud Detector handles all of the ML heavy liftings for you.
Sounds like the right solution for your business? Find out more here.
Amazon Connect is a call-center in the cloud.
You can now create your very own call-center in the cloud. You even receive numbers that your customers can call. You can create your workflow. The operator can work from anywhere with a virtual call, they just need to connect. You pay only by the minutes of calls, which is beneficial for your wallet as well.
Amazon Connect is so simple to set-up and use, you can increase your speed of innovation. With only a few clicks, you can set up an omnichannel contact center and agents can begin talking and messaging with customers. Get started today! Read more here.
However many customers have been saying they would like some transcription of the call, to do more in-depth analysis. That’s why AWS created Contact Lens. Contact lens performs real-time sentiment analysis on the call itself using machine learning capabilities. It can help you return the sentiment to something more positive during the calls your agent is making. Providing a better customer experience is essential, so this is a crucial part of your call-center that you need up and running now. Find out more about Contact Lens here.
Code guru is a machine learning-based system that finds patterns inside of code that can be improved instead of a senior code developer sitting next to you. It’s like having a distinguished engineer on call 24/7 but instead this tool eps you find the most expensive lines of coding that hurt your application performance and have you up all night, troubleshooting. Find out more about this amazing tool here.
There are a bunch of premade artificial intelligence tools you can consume as you go and you only pay for what you consume. You can enrich those tools and bring your vocabulary onto it while consuming them. This allows you to not only pay for what you need, which is amazing, but use what you want and spend time on developing the things you need or want. It’s right there, ready to use, ready for you to implement into your business. You’d be a fool not to take a chance and work with what AWS offers you here.