6 Ways to Start Utilizing Machine Learning with Amazon Web Services and Talend
A common perspective that I see amongst software designers and developers is that Machine Learning and Artificial Intelligence (AI) are technologies which are only meant for an elite group. However, if a particular technology is to truly succeed and scale, it should be friendly with the common man (in this case a normal software developer).
In this blog, I would like to discuss how Amazon Web Services (AWS) has changed the landscape of Machine Learning and AI through its various services and how Talend is further evangelizing this idea of “Democratizing Machine-Learning” among the IT masses. We will also review some sample scenarios where Talend is integrated with different Amazon Machine Learning services. I have also added links for corresponding Talend KB articles with sample Talend job details for the audience who would love to do some real hands-on experience.
Before going further into the topic, I would like to invite the readers to go through three interesting blogs to get more idea about Machine Learning and Artificial Intelligence.
- What Everyone Should Know about Machine Learning
- AI vs Machine Learning vs Deep Learning
- How to Operationalize Machine Learning with Talend
Getting Started with Amazon Web Services and Talend
Now, let’s get back to our original discussion point i.e., how we can democratize the machine learning area further so that a common IT developer can make use of these functionalities without writing any complex programming.
The relevance of AWS Machine Learning and AI Services comes in this context as these services helps to encapsulate the complex underlying processes from the end user. Talend augments the usage of these wonderful services further by packaging these service calls neatly with multiple data flows. This methodology helps to build various everyday applications in a simple yet sophisticated manner.
Age of “SaaS”
Software as a Service (“SaaS”) has become the new normal for Information Technology. In the context of Machine Learning and Artificial Intelligence, Amazon Web Services has created a wonderful series of SaaS modules. These modules help users to perform tasks like real-time predictions based on machine learning models, sentimental analysis of messages, language detection and translation, image moderation, text analysis from images, speech synthesis and much more. Talend helps to integrate these machine learning and AI services of Amazon in a seamless manner to create end-to-end applications in a graphical format.
Prediction made easy
Predicting real-time response based on a machine learning model was a daunting task, even just a while ago. Talend helps to make this task easy by integrating with Amazon’s Machine Learning Real-time prediction service. For example, a Botanist can now predict the class of a flower by analyzing its petal and sepal features instantly, based on the Machine Learning model he or she has created.
In the diagram above, Talend helps in dataset creation (which will be used to build ML model) by fetching flower classification data from different source systems in multiple formats. Talend also helps in the real-time prediction of a flower category based on the input data provided by Botanist by making a request to Amazon Machine Learning Service. Instead of writing complex codes spread across thousands of lines, programmers can create real-time prediction services by getting the help of Amazon Web Services and Talend.
The Talend KB Article, which outlines the various steps involved in real-time prediction using Machine Learning, can be referred from below link.
Language processing – Key for becoming “Global Village”
The world has become a global village and interactions between people from different parts of the world are increasing day-by-day. Language was one of the major roadblocks in enabling free communication between people all over the world. The natural language processing services of Amazon like Amazon Comprehend and Amazon Translate help us to understand the dominant language any given text text, translate it and perform the sentiment analysis for the incoming textual information. Talend integrates these Amazon AI services to convert end to end applications like real-time sentimental analysis dashboard and multilingual customer care system.
A quick example is the sentimental analysis dashboard as shown below. Talend is integrated with Amazon’s Comprehend service to identify the customer sentiments in real time and to send the sentimental analysis details to downstream system dashboards.
Another example which showcases Talend’s integration capabilities with Amazon Comprehend and Amazon Translate services is the creation of a multi-lingual customer care system. The incoming messages are analyzed to understand the dominant language used in it and the text is translated from non-supported languages to supported language automatically.
The two Talend KB articles I would recommend getting a detailed overview and hands-on experience about Talend’s integration with above two Amazon services are as shown below.
Image Processing and analysis opens a lot of new opportunities to make our life easier. Amazon Rekognition provides one of the top-notch image processing services. Talend helps you to build end to end applications like image moderation system, celebrity image analysis, text recognition process from images and facial analysis.
A sample illustration of the various possibilities can be shown from below diagram where Talend is easily integrated with Amazon Rekognition to perform real-time vehicle number plate analysis. This helps in various avenues like parking monitoring, toll collection and to capture automobiles running red lights.
Another interesting use case is to setup image moderation in websites as shown below where Talend verifies the incoming image by making request to Amazon Rekognition. Based on the result, the images are allowed to be consumed by consumer systems or image will be quarantined with warning messages to necessary teams.
Talend and Amazon Rekognition can also be used for other use cases like Celebrity Recognition from the images which can be used by media houses, facial analysis systems for home security systems, crime scene analysis. I would recommend you go through below KB article to understand the details of integration and would highly recommend trying some of the above scenarios by creating sample jobs using Talend.
Another interesting Talend integration with Amazon AI service is the amalgamation of Talend with Amazon Polly. Speech synthesis is a remarkable game changer in many applications. Amazon Polly can do speech synthesis from multiple languages. Talend helps to integrate this interesting AI service with other areas of a complex enterprise application.
A simple use case where Talend is seamlessly integrated with Amazon Polly is the flight tracking and automatic audio notification systems. As shown in below diagram, Talend transfers the data from various sources as text messages to Amazon Polly. The converted voice messages will be fetched and transferred to downstream systems.
The KB article which explains the steps to do the Talend integration with Amazon Polly can be referred from the link below. It also has a sample program where Talend is supplying the input text and converting them to mp3 audio files.
Many Talend users have a understand that Talend can be used to connect with pre-built components in the Talend Palette, but a true lover of Talend will explore other possibilities like creating a custom component, usage of routines etc. to integrate with different services. Integration of Talend with Amazon’s Machine learning and AI services is one such fascinating experiment to showcase the true power of Talend. It also illustrates how Talend and Amazon web services are changing the complex world of Machine Learning to a more democratic one for software engineers. Hope you found this overview useful and please let me know of your machine learning use cases with Talend in the comments.