Search the Community

Amazon Personalize is excited to announce automatic training for solutions. With automatic training, developers can set a cadence for their Personalize solutions to automatically retrain using the latest data from their dataset group. This process creates a newly trained machine learning (ML) model, also known as a solution version, and maintains the relevance of Amazon Personalize recommendations for end users. View the full article

Amazon Personalize is excited to announce the new Next-Best-Action recipe to help you recommend actions that your users have a high probability to take in real-time. Expanding beyond the items or content that Personalize has allowed customers to recommend, this new recipe assists you in determining the next best actions to suggest to your individual users that will help in increasing brand loyalty and conversion. Similar to other Amazon Personalize recipes, Next Best Action does not require ML expertise and allows you to integrate personalization seamlessly into your applications. View the full article

Amazon Personalize is excited to launch a new integration with Amazon OpenSearch Service that enables customers to personalize search results for each user and improve the user engagement from their search. The Amazon Personalize Search Ranking plugin within Amazon OpenSearch Service helps customers leverage the deep learning capabilities offered by Amazon Personalize and add personalization to OpenSearch search results, without any ML expertise. View the full article

Amazon Personalize launches a new integration with self-managed OpenSearch that enables customers to personalize search results for each user and assists in predicting their search needs. The Amazon Personalize Search Ranking plugin within OpenSearch helps customers to leverage the deep learning capabilities offered by Amazon Personalize and add personalization to OpenSearch search results, without any ML expertise. View the full article

Amazon Personalize now makes it easier to modify datasets by allowing customers to add columns to an existing schema. Amazon Personalize uses datasets provided by customers to train custom personalization models on their behalf. Customers modify existing datasets to add new filtering columns for enhanced business logic and to add new columns that can improve model training. Previously, to add new columns customers would need to reproduce existing resources starting at the dataset level. With this feature, customers can quickly update their schema to append an additional column without having to reproduce resources. View the full article

Amazon Personalize now uses the latest streamed data for batch recommendations, improving recommendation quality by capturing recent user interactions. Batch recommendations now use newly recorded events streamed via Personalize’s event tracker to generate recommendations without requiring retraining of the model. Previously, batch recommendations would only consider the interactions up to the point of the last model retraining. By considering more recent interactions, Personalize’s recommendations can now better respond to shifts in user behavior. View the full article

Amazon Personalize announces the ability to filter (include or exclude) recommendations for Related Items recipes based on the properties of the input or query item (e.g. the CurrentItem). Prior to this launch, Personalize customers could filter items based on dataset properties such as category or price range. Customers could also filter recommendations based on properties of the user getting recommendations (e.g. the CurrentUser), such as age or whether or not the user had taken a specific action. View the full article

Amazon Personalize now supports Amazon Virtual Private Cloud (VPC) endpoints, allowing Amazon Personalize to communicate with your resources on your VPC without going through the open internet. Amazon VPC is a service that you use to launch AWS resources in a private virtual network that you define and manage. To connect your VPC to Amazon Personalize, you define a VPC endpoint for Amazon Personalize. An endpoint is an elastic network interface with a private IP address that serves as an entry point for traffic destined to a supported AWS service. The endpoint provides reliable, scalable connectivity to Amazon Personalize, and doesn’t require an internet gateway or VPN connection. For more information, see What is Amazon VPC in the Amazon VPC User Guide. View the full article

We are excited to announce that Amazon Personalize now offers more flexibility in model training by allowing customers to select which columns in their datasets are used for training. Amazon Personalize enables developers to improve customer engagement through personalized product and content recommendations – no ML expertise required. Amazon Personalize uses data provided by customers to train custom models on their behalf. Previously, all columns in a customer’s dataset were considered during model training. Amazon Personalize now allows customers to exclude columns from training, making it easier to experiment with the data that is used to train models. With this new capability, customers now have the flexibility to filter on any of their data, regardless of whether it is used for training. View the full article

Today, Amazon Personalize announces the launch of popularity tuning for its Similar-Items recipe (aws-similar-items). Similar-Items generates recommendations that are similar to the item that a user selects, helping users discover new items in a customer’s catalog based on the behavior of all users and item metadata. View the full article

Amazon Personalize has launched a new open source Amazon Personalize Kafka Sink connector that makes it easy to stream data in real-time for use in Amazon Personalize. Amazon Personalize enables developers to improve customer engagement through personalized product and content recommendations – no ML expertise required. Amazon Personalize uses data provided by customers to train custom models on their behalf. With this launch, customers can readily ingest their data from Apache Kafka clusters and call the Amazon Personalize-specific APIs to enable real-time data steaming, without the need for custom code. This makes it faster and easier to bring real-time data to Amazon Personalize for those using Apache Kafka. View the full article

Amazon Personalize is integrating with Amazon SageMaker Data Wrangler to make it easier for customers to import and prepare their data. Amazon Personalize enables developers to improve customer engagement through personalized product and content recommendations – no ML expertise required. The quality of data used for model training affects the quality of the recommendations, which makes data aggregation and preparation a critical step to get high-quality recommendations using Amazon Personalize. With this launch, Amazon Personalize gives you the ability to prepare your data through Amazon SageMaker Data Wrangler before using it in Amazon Personalize. Customers can use Amazon SageMaker Data Wrangler to import data from 40+ supported data sources and perform end-to-end data preparation (including data selection, cleansing, exploration, visualization, and processing at scale) in a single user interface using little to no code. This allows customers to rapidly prepare their users, items or interactions dataset using Amazon SageMaker Data Wrangler by leveraging over 300 built-in data transformations, retrieving data insights, and quickly iterating by fixing data issues. View the full article

We are excited to announce that Amazon Personalize now provides analysis on your data to make onboarding easier than ever. Amazon Personalize enables developers to improve customer engagement through personalized product and content recommendations – no ML expertise required. Amazon Personalize trains custom models for each customer using their unique data. With this launch, Amazon Personalize now analyzes the data you provide and offers suggestions to assist you in improving your data preparation. View the full article

Today we are announcing support for tags in IAM policies to allow granular control over access to Amazon Personalize resources and operations. Amazon Personalize enables developers to improve customer engagement through personalized product and content recommendations – no ML expertise required. Tags are labels in the form of key-value pairs that can be attached to individual Amazon Personalize resources to manage resources, or allocate costs. With this launch, customers can also perform tag based access control for Amazon Personalize resources and operations such as modify, update or delete. For example, you can limit access to delete or update operations to specific individuals to avoid any accidental impact to your production environment. This functionality also allows customers with multi-tenant deployments to partition access to resources across their end customers. This functionality is available for several Amazon Personalize resources such as dataset groups, solutions, campaigns, recommenders, import jobs, batch inference, batch segment jobs and other resources. View the full article

Today, Amazon Personalize is excited to announce a new Trending-Now recipe that will help customers recommend items gaining popularity at the fastest pace among their users. Amazon Personalize enables developers to improve customer engagement through personalized product and content recommendations – no ML expertise required. User interests can change based on a variety of factors, such as external events or the interests of other users. It is critical for customers to tailor their recommendations to these changing interests to improve user engagement. With Trending-Now, you can surface items from your catalogue that are rising in popularity faster than other items, such as breaking news articles, popular social content or newly released movies. Amazon Personalize looks for items that are rising in popularity at a faster rate than other catalogue items to help provide an engaging experience. Amazon Personalize also allows customers to define the frequency at which it identifies trending items, with options for refreshing recommendations every 30 mins, 1 hour, 3 hours or 1 day, based on the most recent interactions data from users. View the full article

We are excited to announce that Amazon Personalize has extended limits to support datasets with up to 100 million users and 3 billion interactions. Amazon Personalize enables developers to improve customer engagement through personalized product and content recommendations – no ML expertise required. Amazon Personalize trains custom models for each customer using their unique data. Previously, these models could consider up to 50 million users in training. By doubling this limit, Amazon Personalize improves model performance for large customers by allowing them to train on a more diverse set of data. For customers with larger user bases that may exceed this limit, Amazon Personalize samples an optimal set of users before training. Previously, models trained by Amazon Personalize would also consider a maximum of 500 million of the latest interactions between users and items. Customers now have the option to increase their training time to consider up to 3 billion interactions. This can improve model performance by capturing more historical data for customers with a large user base or a high velocity of interactions. To increase the number of interactions considered by your model, simply request a service quota increase via the Service Quota console. View the full article

Today, Amazon Personalize launches the ability for customers to measure the business impact of their Personalize recommendations. Amazon Personalize is a fully managed machine learning service that allows customers to deliver personalized experiences to their users, no ML experience required. Customers can now evaluate the sum or count of any user interaction with a Personalize recommendation such as click, view, purchase, video start, add to cart, or download. Prior to this launch, customers could not evaluate the results of Personalize recommendations unless they built their own pipelines or workflows. Personalize has removed the need for this operational overhead by allowing customers to define the metrics they want to track, and automatically sending event data to the customer’s Amazon CloudWatch account for visualization and monitoring. The data can also be sent to an S3 bucket for download and integration into another business intelligence tool. View the full article

Amazon Personalize now allows you to apply business rules to your recommendations on the fly; without any extra cost. Dynamic filters save you time by removing the need to define all possible permutations of your business rules in advance and enable you to use your most recent information to filter recommendations. You control the recommendations for your users in real-time while responding to their individual needs, preferences, and changing behavior to improve engagement and conversion. Based on over 20 years of personalization experience, Amazon Personalize enables you to improve customer engagement by powering personalized product and content recommendations, and targeted marketing promotions. View the full article