Recently, Google Cloud partner Groovenauts, Inc. published a live demo of MatchIt Fast. As the demo shows, you can find images and text similar to a selected sample from a collection of millions in a matter of milliseconds:

Give it a try — and either select a preset image or upload one of your own. Once you make your choice, you will get the top 25 similar images from two million images on Wikimedia images in an instant, as you can see in the video above. No caching involved.

The demo also lets you perform the similarity search with news articles. Just copy and paste some paragraphs from any news article, and get similar articles from 2.7 million articles on the GDELT project within a second.

Vector Search: the technology behind Google Search, YouTube, Play, and more

How can it find matches that fast? The trick is that the MatchIt Fast demo uses the vector similarity search (or nearest neighbor search or simply vector search) capabilities of the Vertex AI Matching Engine, which shares the same backend as Google Image Search, YouTube, Google Play, and more, for billions of recommendations and information retrievals for Google users worldwide. The technology is one of the most important components of Google’s core services, and not just for Google: it is becoming a vital component of many popular web services that rely on content search and information retrieval accelerated by the power of deep neural networks.

So what’s the difference between traditional keyword-based search and vector similarity search? For many years, relational databases and full-text search engines have been the foundation of information retrieval in modern IT systems. For example, you would add tags or category keywords such as “movie”, “music”, or “actor” to each piece of content (image or text) or each entity (a product, user, IoT device, or anything really). You’d then add those records to a database, so you could perform searches with those tags or keywords.

In contrast, vector search uses vectors (where each vector is a list of numbers) for representing and searching content. The combination of the numbers defines similarity to specific topics. For example, if an image (or any content) includes 10% of “movie”, 2% of “music”, and 30% of “actor”-related content, then you could define a vector [0.1, 0.02, 0.3] to represent it. (Note: this is an overly simplified explanation of the concept; the actual vectors have much more complex vector spaces). You can find similar content by comparing the distances and similarities between vectors. This is how Google services find valuable content for a wide variety of users worldwide in milliseconds.

With keyword search, you can only specify a binary choice as an attribute of each piece of content; it’s either about a movie or not, either music or not, and so on. Also, you cannot express the actual “meaning” of the content to search. If you specify a keyword “films”, for example, you would not see any content related to “movies” unless there was a synonyms dictionary that explicitly linked these two terms in the database or search engine. 

Vector search provides a much more refined way to find content, with subtle nuances and meanings. Vectors can represent a subset of content that contains “much about actors, some about movies, and a little about music”. Vectors can represent the meaning of content where “films”, “movies”, and “cinema” are all collected together. Also, vectors have the flexibility to represent categories  previously unknown to or undefined by service providers. For example, emerging categories of content primarily attractive to kids, such as ASMR or slime, are really hard for adults or marketing professionals to predict beforehand, and going back through vast databases to manually update content with these new labels would be all but impossible to do quickly. But vectors can capture and represent never-before-seen categories instantly.

Vector search changes business

Vector search is not only applicable to image and text content. It can also be used for information retrieval for anything you have in your business when you can define a vector to represent each thing. Here are a few examples:

• Finding similar users: If you define a vector to represent each user in your business by combining the user’s activities, past purchase history, and other user attributes, then you can find all users similar to a specified user.  You can then see, for example, users who are purchasing similar products, users that are likely bots, or users who are potential premium customers and who should be targeted with digital marketing.
• Finding similar products or items: With a vector generated from product features such as description, price, sales location, and so on, you can find similar products to answer any number of questions; for example, “What other products do we have that are similar to this one and may work for the same use case?” or “What products sold in the last 24 hours in this area?” (based on time and proximity)
• Finding defective IoT devices: With a vector that captures the features of defective devices from their signals, vector search enables you to instantly find potentially defective devices for proactive maintenance.
• Finding ads: Well-defined vectors let you find the most relevant or appropriate ads for viewers in milliseconds at high throughput.
• Finding security threats: You can identify security threats by vectorizing the signatures of computer virus binaries or malicious attack behaviors against web services or network equipment. 
• …and many more: Thousands of different applications of vector search in all industries will likely emerge in the next few years, making the technology as important as relational databases.

OK, vector search sounds cool. But what are the major challenges to applying the technology to real business use cases? Actually there are two:

• Creating vectors that are meaningful for business use cases
• Building a fast and scalable vector search service

Embeddings: meaningful vectors for business use cases

The first challenge is creating vectors for representing various entities that are meaningful and useful for business use cases. This is where deep learning technology can really shine. In the case of the MatchIt Fast demo, the application simply uses a pre-trained MobileNet v2 model for extracting vectors from images, and the Universal Sentence Encoder (USE) for text. By applying such models to raw data, you can extract “embeddings” – vectors that map each row of data in a space of their “meanings”. MobileNet puts images that have similar patterns and textures closer to one another in the embedding space, and USE puts texts that have similar topics closer.

For example, a carefully designed and trained machine learning model could map movies into an embedding space like the following:

With the embedding space shown here, users could find recommended movies based on the two dimensions: is the movie for children or adults, and is it a blockbuster or arthouse movie? This is a very simple example, of course, but with an embedding space like this that fits your business requirements, you can deliver a better user experience on recommendation and information retrieval services with insights extracted from the model. 

For more about creating embeddings, the Machine Learning Crash Course on Recommendation Systems is a great way to get started. We will also discuss how to extract better embeddings from business data later in this post.

Building a fast and scalable vector search service

Suppose that you have successfully extracted useful vectors (embeddings) from your business data. Now the only thing you have to do is search for similar vectors. That sounds simple, but in practice it is not. Let’s see how the vector search works when you implement it with BigQuery in a naive way:https://www.youtube.com/embed/wHNJspvxj2w?enablejsapi=1&

It takes about 20 seconds to find similar items (fish images in this case) from a pool of one million items. That level of performance is not so impressive, especially when compared to the MatchIt Fast demo. BigQuery is one of the fastest data warehouse services in the industry, so why does the vector search take so long?

This illustrates the second challenge: building a fast and scalable vector search engine isn’t an easy task. The most widely used metrics for calculating the similarity between vectors are L2 distance (Euclidean distance), cosine similarity, and inner product (dot product).

But all require calculations proportional to the number of vectors multiplied by the number of dimensions if you implement them in a naive way. For example, if you compare a vector with 1024 elements to 1M vectors, the number of calculations will be proportional to 1024 x 1M = 1.02B. This is the computation required to look through all the entities for a single search, and the reason why the BigQuery demo above takes so long.

Instead of comparing vectors one by one, you could use the approximate nearest neighbor (ANN) approach to improve search times. Many ANN algorithms use vector quantization (VQ), in which you split the vector space into multiple groups, define “codewords” to represent each group, and search only for those codewords. This VQ technique dramatically enhances query speeds and is the essential part of many ANN algorithms, just like indexing is the essential part of relational databases and full-text search engines.

As you may be able to conclude from the diagram above, as the number of groups in the space increases the speed of the search decreases and the accuracy increases.  Managing this trade-off — getting higher accuracy at shorter latency — has been a key challenge with ANN algorithms. 

Last year, Google Research announced ScaNN, a new solution that provides state-of-the-art results for this challenge. With ScaNN, they introduced a new VQ algorithm called anisotropic vector quantization:

Anisotropic vector quantization uses a new loss function to train a model for VQ for an optimal grouping to capture farther data points (i.e. higher inner product) in a single group. With this idea, the new algorithm gives you higher accuracy at lower latency, as you can see in the benchmark result below (the violet line): 

This is the magic ingredient in the user experience you feel when you are using Google Image Search, YouTube, Google Play, and many other services that rely on recommendations and search. In short, Google’s ANN technology enables users to find valuable information in milliseconds, in the vast sea of web content.

How to use Vertex AI Matching Engine

Now you can use the same search technology that powers Google services with your own business data. Vertex AI Matching Engine is the product that shares the same ScaNN based backend with Google services for fast and scalable vector search, and recently it became GA and ready for production use. In addition to ScaNN, Matching Engine gives you additional features as a commercial product, including:

• Scalability and availability: The open source version of ScaNN is a good choice for evaluation purposes, but as with most new and advanced technologies, you can expect challenges when putting it into production on your own. For example, how do you operate it on multiple nodes with high scalability, availability, and maintainability? Matching Engine uses Google’s production backend for ScaNN, which provides auto-scaling and auto-failover with a large worker pool. It is capable of handling tens of thousands of requests per second, and returns search results in less than 10 ms for the 90th percentile with a recall rate of 95 – 98%.
• Fully managed: You don’t have to worry about building and maintaining the search service. Just create or update an index with your vectors, and you will have a production-ready ANN service deployed. No need to think about rebuilding and optimizing indexes, or other maintenance tasks.
• Filtering: Matching Engine provides filtering functionality that enables you to filter search results based on tags you specify on each vector. For example, you can assign “country” and “stocked” tags to each fashion item vector, and specify filters like “(US OR Canada) AND stocked”  or “not Japan AND stocked” on your searches.

Let’s see how to use Matching Engine with code examples from the MatchIt Fast demo.

Generating embeddings

Before starting the search, you need to generate embeddings for each item like this one:

  {
  "Id":"b5c65fea9b0b8a57bfa574ea",
  "Embedding": [
    0.16329009830951691,
    0.92436742782592773,
    0.00095699273515492678,
    0.011479727923870087,
    0.0089491046965122223,
    0.019959751516580582,
    0.031516745686531067,
    0.0066015380434691906,
    0.46404418349266052,
    ...

This is an embedding with 1280 dimensions for a single image, generated with a MobileNet v2 model. The MatchIt Fast demo generates embeddings for two million images with the following code:

  class Vectorizer:
    def __init__(self):
        self._model = tf.keras.Sequential([hub.KerasLayer("https://tfhub.dev/google/imagenet/mobilenet_v2_100_224/feature_vector/5", trainable=False)])
        self._model.build([None, 224, 224, 3])  # Batch input shape.

    def vectorize(self, jpeg_file):
        ...snip...
        embedding = self._model.predict({"inputs": input_tensor})[0].tolist()
        return embedding

After you generate the embeddings, you store them in a Google Cloud Storage bucket. 

Configuring an index

Then, define a JSON file for the index configuration:

  {
  "contentsDeltaUri": "gs://match-it-fast-us-central1/wikimedia_images/index-1",
  "config": {
    "dimensions": 1280,
    "approximateNeighborsCount": 150,
    "distanceMeasureType": "SQUARED_L2_DISTANCE",
    "algorithm_config": {
      "treeAhConfig": {
        "leafNodeEmbeddingCount": 1000,
        "leafNodesToSearchPercent": 5
      }
    }
  }
}

You can find a detailed description for each field in the documentation, but here are some important fields:

• contentsDeltaUri: the place where you have stored the embeddings
• dimensions: how many dimensions in the embeddings
• approximateNeighborsCount: the default number of neighbors to find via approximate search 
• distanceMeasureType: how the similarity between embeddings should be measured, either L1, L2, cosine or dot product (this page explains which one to choose for different embeddings)

To create an index on the Matching Engine, run the following gcloud command where the metadata-file option takes the JSON file name defined above.

  gcloud --project=gn-match-it-fast beta ai indexes create \
       --display-name=wikimedia-images \
       --description="Wikimedia Image Demo" \
       --metadata-file=metadata/wikimedia_images_index_metadata.json \
       --region=us-central1

Run the search

Now the Matching Engine is ready to run. The demo processes each search request in the following order:

1. First, the web UI takes an image (the one chosen or uploaded by the user) and encodes it into an embedding using the TensorFlow.js MobileNet v2 model running inside the browser. Note: this “client-side encoding” is an interesting option for reducing network traffic when you can run the encoding at the client. In many other cases, you would encode contents to embeddings with a server-side prediction service such as Vertex AI Prediction, or just retrieve pre-generated embeddings from a repository like Vertex AI Feature Store.
2. The App Engine frontend receives the embedding and submits a query to the Matching Engine. Note that you can also use any other compute services in Google Cloud for submitting queries to Matching Engine, such as Cloud Run, Compute Engine, or Kubernetes Engine, or whatever is most suitable for your applications.
3. Matching Engine executes its search. The connection between App Engine and Matching Engine is provided via a VPC private network for optimal latency.
4. Matching Engine returns the IDs of similar vectors in its index.

Step 3 is implemented with the following code:

  class MatchingQueryClient:
    ...snip...

    def query_embedding(self, embedding, num_neighbors=30):
        request = match_service_pb2.MatchRequest()
        request.deployed_index_id = self._deployed_index_id
        for v in embedding:
            request.float_val.append(v)
        request.num_neighbors = num_neighbors
        response = self._stub.Match(request)
        return response

The request to the Matching Engine is sent via gRPC as you can see in the code above. After it gets the request object, it specifies the index id, appends elements of the embedding, specifies the number of neighbors (similar embeddings) to retrieve, and calls the Match function to send the request. The response is received within milliseconds.

Next steps: Making changes for various use cases and better search quality

As we noted earlier, the major challenges in applying vector search on production use cases are:

• Creating vectors that are meaningful for business use cases
• Building a fast and scalable vector search service

From the example above, you can see that Vertex AI Matching Engine solves the second challenge. What about the first one? Matching Engine is a vector search service; it doesn’t include the creating vectors part.

The MatchIt Fast demo uses a simple way of extracting embeddings from images and contents; specifically it uses an existing pre-trained model (either MobileNet v2 or Universal Sentence Encoder). While those are easy to get started with, you may want to explore other options to generate embeddings for other use cases and better search quality, based on your business and user experience requirements.

For example, how do you generate embeddings for product recommendations?  The Recommendation Systems section of the Machine Learning Crash Course is a great resource for learning how to use collaborative filtering and DNN models (the two-tower model) to generate embeddings for recommendation. Also, TensorFlow Recommenders provides useful guides and tutorials for the topic, especially on the two-tower model and advanced topics. For integration with Matching Engine, you may also want to check out the Train embeddings by using the two-tower built-in algorithm page.

Another interesting solution is the Swivel model. Swivel is a method for generating item embeddings from an item co-occurrence matrix. For structured data, such as purchase orders, the co-occurrence matrix of items can be computed by counting the number of purchase orders that contain both product A and product B, for all products you want to generate embeddings for. To learn more, take a look at this tutorial on how to use the model with Matching Engine.

If you are looking for more ways to achieve better search quality, consider metric learning, which enables you to train a model for discrimination between entities in the embedding space, not only classification:

Popular pre-trained models such as the MobileNet v2 can classify each object in an image, but they are not explicitly trained to discriminate the objects from each other with a defined distance metric. With metric learning, you can expect better search quality by designing the embedding space optimized for various business use cases. TensorFlow Similarity could be an option for integrating metric learning with Matching Engine.

Interested? Today, we’re just beginning the migration from traditional search technology to new vector search. Over the next 5 to 10 years, many more best practices and tools will be developed in the industry and community. These tools and best practices will help answer many questions, like… How do you design your own embedding space for a specific business use case? How do you measure search quality? How do you debug and troubleshoot the vector search? How do you build a hybrid setup with existing search engines for meeting sophisticated requirements? There are many new challenges and opportunities ahead for introducing the technology to production. Now’s the time to get started delivering better user experiences and seizing new business opportunities with Matching Engine powered by vector search.

Acknowledgements

We would like to thank Anand Iyer, Phillip Sun, and Jeremy Wortz for their invaluable feedback to this post.

Generative AI is a rapidly expanding technology with a wide range of potential applications. Google Cloud Learning is thrilled to offer a new, no-cost Generative AI Fundamentals skill badge. This skill badge is designed for anyone eager to learn about the power of generative AI. No technical skills or prior knowledge required!

For those new to digital credentials, Google Cloud skill badges are digital credentials issued by Google Cloud in recognition of your knowledge of Google Cloud products and services. Individuals can earn skill badges on Google Cloud Skills Boost, and can share their skill badge to their social media profile and resume.

Watch the short videos in the generative AI courses and complete the final quiz to earn the Generative AI Fundamentals skill badge pictured below. In as little as 120 minutes, you will learn the basics of how generative AI works, how Google Cloud AI technology can be used by businesses and individuals, and how the principles of responsible AI lead to ethical decisions about the use of generative AI. 

By earning the skill badge, you will demonstrate your understanding of foundational concepts in generative AI.

The topics covered in the courses include:

1. Introduction to Generative AI 

• Explain how generative AI works
• Describe generative AI model types
• Describe generative AI applications

2. Introduction to Large Language Models

• Define large language models (LLMs)
• Describe LLM use cases
• Explain prompt tuning
• Describe Google’s generative AI development tools

3. Introduction to Responsible AI

• Identify the need for a responsible AI practice within an organization
• Recognize that decisions made at all stages of a project make an impact in Responsible AI
• Recognize that organizations can design an AI infrastructure to fit their own business needs and values

Earn the skill badge and show off your generative AI knowledge today! And for more content to help you stay up to date with generative AI, check out “The Prompt” and our generative AI primer for executives on Transform with Google Cloud.

Customers tell us that sharing and exchanging data with other organizations is a critical element of their analytics strategy, but it’s hamstrung by unreliable data and processes, and only getting harder with security threats and privacy regulations on the rise. 

Furthermore, traditional data sharing techniques use batch data pipelines that are expensive to run, create late arriving data, and can break with any changes to the source data. They also create multiple copies of data, which brings unnecessary costs and can bypass data governance processes. These techniques do not offer features for data monetization, such as managing subscriptions and entitlements. Altogether, these challenges mean that organizations are unable to realize the full potential of transforming their business with shared data.

To address these limitations, we are introducing Analytics Hub, a new fully managed service, available in Q3, in preview, that helps you unlock the value of data sharing, leading to new insights and increased business value. With Analytics Hub you get:

• A rich data ecosystem by publishing and subscribing to analytics-ready datasets. 
• Control and monitoring over how your data is being used, because data is shared in one place.
• A self-service way to access valuable and trusted data assets, including data provided by Google. For example, a unique dataset from Google Search Trends will be available, that you can query and combine with your own data.
• An easy way to monetize your data assets without the overhead of building and managing the infrastructure. 

Built on a decade of cross-organizational sharing

While Analytics Hub is a new service, it builds on BigQuery, Google’s petabyte-scale, serverless cloud data warehouse. BigQuery’s unique architecture provides separation between compute and storage, enabling data publishers to share data with as many subscribers as you want without having to make multiple copies of your data. With BigQuery, there are no servers to deploy or manage, which means that data consumers get immediate value from shared data. Data can be provided and consumed in real-time using the streaming capabilities of BigQuery and you can leverage the built in machine learning, geospatial, and natural language capabilities of BigQuery or take advantage of the native business intelligence support with tools like Looker, Google Sheets, and Data Studio.

BigQuery has had cross-organizational, in-place data sharing capabilities since it was introduced in 2010. We took a look at usage metrics in BigQuery and found that over a 7 day period in April, we had over 3,000 different organizations sharing over 200 petabytes of data. These numbers don’t include data sharing between departments within the same organization.

As you can see, data sharing in BigQuery is already popular. But we want to make it easier and even more scalable.

Raising the bar on data sharing 

To make data sharing easier and more scalable in BigQuery, Analytics Hub introduces the  concepts of shared datasets and exchanges. As a data publisher, you create shared datasets that contain the views of data that you want to deliver to your subscribers. Next, you create exchanges, which are used to organize and secure shared datasets. By default, exchanges are completely private, which means that only the users and groups that you give access to can view or subscribe to the data. You can also create internal exchanges or leverage public exchanges provided by Google. Finally, you publish shared datasets into an exchange to make them available to subscribers. 

Data subscribers search through the datasets that are available across all exchanges for which they have access and subscribe to relevant datasets. This creates a linked dataset in their project that they can query and join with their own data. Subscribers pay for the queries that they run against the data while the publisher pays for the storage of the data. Data providers can add new data, new tables, or new columns to the shared dataset and these will be immediately available to subscribers. In addition, the publisher can track subscribers, disable subscriptions, and see aggregated usage information for the shared data. 

Analytics Hub makes it easy for you to publish, discover, and subscribe to valuable datasets that you can combine with your own data to derive unique insights. Here are some types of data that will be available through Analytics Hub:

• Public datasets: Easy access to the existing repository of over 200 public datasets, including data about weather and climate, cryptocurrency, healthcare and life sciences, and transportation. 
• Google datasets: Unique, freely-available datasets from Google. One example of this is the COVID-19 community mobility dataset. Another example is the forthcoming Google Trends dataset, which will provide the top 25 search terms and top 25 rising search terms over a 5 year window in 210 distinct locations in the US. Trends data can be used by everyone in the organization to gain insights into what customers care about.
• Commercial (paid for) datasets: We are working with leading commercial data providers to bring their data products to Analytics Hub. If you are interested in delivering your data via Analytics Hub, we’re also introducing Data Gravity, an initiative that provides storage benefits and new distribution paths for data published through Analytics Hub. 
• Internal datasets: We know that data sharing can be challenging in larger organizations. Analytics Hub can be used for internal data, for example, to share standardized customer demographics with your sales engineering and data science teams.

Customers and partners using Analytics Hub

“Google Search Trends data has always been an important tool for our WPP agency data teams. At WPP we believe that data variety is a superpower which is why we are excited to use the new Trends dataset availability within BigQuery, plus the launch of Analytics Hub. The best creativity in the world is informed by data insights, and influenced by what people search for, so the operational efficiencies we’ll gain via the Analytics Hub and the insights we can drive with Trends data are just phenomenal.”
—Di Mayze Global Head of Data and AI, WPP

“Equifax Ignite is our shared data analytics environment within our Equifax data fabric. We are excited to partner with Google to leverage Analytics Hub and BigQuery to deliver data to over 400 statisticians and data modelers as well as securely sharing data with our partner financial institutions.”
—Kumar Menon, SVP Data Fabric and Decision Science, Equifax

“The flow of data and insights between our teams at Deloitte and our clients is paramount for building truly transformational data cultures. With its purpose-built architecture for secure data exchanges and sharing analytics resources, Google Cloud’s Analytics Hub can help provide significant operational efficiencies for how Deloitte teams support our clients’ data-driven initiatives within their industry ecosystems. It will also help minimize the worries about scale, privacy and security, or the administrative burden associated with each.”
—Navin Warerkar, Managing Director, Deloitte Consulting LLP, and US Google Cloud Data & Analytics GTM Lead

“Crux Informatics is proud to partner with Google to support the launch of Analytics Hub, removing friction for those who need access to analytics-ready data. With thousands of datasets from over 140 sources, Crux Informatics will accelerate access to data on Analytics Hub and together provide a more efficient and cost effective solution to deliver datasets in Google Cloud’s ecosystem.”
—Will Freiberg, CEO, Crux Informatics

Next steps for Analytics Hub

This is just the beginning for Analytics Hub. As we get to preview and general availability, we will be adding additional capabilities, including workflows for publishing and subscribing, publishing analytics assets (Looker Blocks, Data Studio reports, Connected Google Sheets) along with the shared data, the ability for data publishers to specify query restrictions on the usage of their data, and making it easy for data publishers to create sandbox environments for subscribers to work with their data, even if they are not yet on Google Cloud. We will provide features in Analytics Hub for monetization of data, including managing subscriptions, data entitlements, and billing.

Please sign up for the preview, which is scheduled to be available in the third quarter of 2021. In the meantime, you can learn more about BigQuery and how to leverage its built-in data sharing capabilities. Please go to g.co/cloud/analytics-hub to register your interest in Analytics Hub.