Editor’s note: Since launching its Spaces feature, Twitter has demonstrated that hearing people’s voices can bring conversations on Twitter to life in a completely new way. Next, it aimed to make it easier for customers to join and listen to live conversations they personally care about. In this blog, we learn how the Twitter Spaces Engineering team is bringing this vision to life with AutoML, powering a new ML heuristic which serves personalized recommendations to Twitter customers. The authors would like to thank Chuan Lu, Joe Balistreri, Chen-Rui Chou, Pablo Jablonski, Alberto Parrella, Pradip Thachile and Sam Lee from Twitter, as well as Helin Wang from Google, for contributions to this blog.

Since Twitter introduced Spaces in 2020 to enable live audio conversations on its platform, the Twitter Spaces Engineering team has been continually testing, building, and updating this feature in the open. Today, anyone can join, listen, and speak in a Space on Twitter, and the feature’s popularity has taken off. But this success also poses a challenge: with millions of people creating and joining Spaces at any time, how can they find the Spaces to engage with while they’re happening? Taking this as an opportunity to further improve the experience of its customers, Twitter has turned to machine learning (ML) and cloud technology for answers.

“ML fits into the natural progression of Twitter consumer and revenue product building, especially for a product feature such as Spaces,” explains Diem Nguyen, Senior Machine Learning Engineer and Data Scientist at Twitter. “We launched Spaces with a base-line algorithm using the ‘most popular’ heuristic which assumes that if a Space is popular, there’s a good chance you’d like it too. But our aim is to leverage ML to surface the most interesting and relevant Spaces to a particular Twitter customer, making it easier for them to find and join the conversations they personally care about. This is a complex functionality that Google Cloud ML capabilities help us to enable.”

Setting the stage for building new features with limited ML resources

While looking for the right tools to power this vision, Nguyen and her team started evaluating in December 2021 whether the Vertex AI platform and AutoML in particular could solve challenges observed when they first started building Spaces. These included a lack of dedicated ML resources to build and deploy the product feature, and the need to work on a multi-cloud environment.

“We had three key questions in mind during our assessment,” Nguyen explains. “Can we realistically deploy the AutoML model off-platform? Once deployed, can it solve for the request load that we get from the service we’re serving (in this case, the Spaces tab)? And finally, can we develop and maintain such a solution without a dedicated team of ML experts for this project?” The answer to all three questions was yes.

Positive answers motivated the Spaces Engineering team to take the solution to production in February 2022. “We started using AutoML Tables to train high-accuracy models with minimal ML expertise or effort, alleviating our resource constraint,” says Nguyen of the results. “Soon AutoML also stood out for its high performance and for supporting easy deployment beyond the Google Cloud Platform, making it ideal for this project hosted in a multi-cloud environment.”

Increasing customer engagement at speed with accurate ML predictions

With a classification model in place to predict the probability of user engagement in a particular Space, Twitter now aims to optimize its model with aggregated data around Twitter features that can help it better understand customer preferences. For example, if a customer has historically engaged with a particular topic and a new Space matches that topic, the ML model increases the score of that Space being served to that user on the Spaces tab.

Because Spaces are live audio conversations, the Spaces tab needs to be ranked to customers in near real time so they don’t miss out. With this in mind, Twitter’s model currently performs 900 queries per second on the Spaces tab, and evaluates 50,000 candidates per second. Meanwhile, 99% of these requests are faster than 100 milliseconds, and 90% of requests are faster than 50 milliseconds.

To measure the success of this project, Nguyen’s team conducted A/B experiments around key customer engagement metrics–A stands for the ‘most popular’ heuristic previously in production, and B is the new AutoML model which seeks to personalize Spaces recommendations to the interests of individual Twitter users. Three months into the project, the numbers were encouraging. “After deploying our AutoML Tables solution we saw an increase of 1.96% in Spaces daily active customers, which is one of our key metrics. We also noticed an increase of 1.99% in Spaces join in rates, and an increase of 8.42% in user clicks to explore a Space,” Nguyen shares. “These are positive signals that users are now engaging more with the Spaces tab service on the Twitter app, which is exactly what we set out to do with this project.”

Powering new use cases with hands-off ML frameworks

With this first solution running in production to improve the performance of the Spaces tab, Nguyen starts to ask how else it might support the experience of Twitter users moving forward. “The Spaces tab is a small surface on the Twitter app. With our current ML solution we’re some distance away from serving our home tab traffic, which is where a lot of our traffic happens and therefore would involve a much bigger-scale operation. Getting there will take some work but we’re evaluating the possibility of optimizing our model performance for this in collaboration with Google Cloud,” says Nguyen.

“As a product-led company, we focus on continually improving the customer experience and we want to iterate faster to get to that point. AutoML brings that value to our product teams because it is so hands-off. You don’t need to write any model code in order to reap the benefits from this machine learning framework; AutoML automatically experiments with many different model architectures and comes up with a state-of-the-art model that addresses your needs. So while it is not a one-size-fits-all solution, it is a great solution with the potential to power many more Twitter use cases,” she concludes.

Google Cloud is proud to support our customers with the cleanest cloud in the industry. For the past four years, we’ve matched 100% of our electricity use with renewable energy purchases, and we were the first company of our size to commit going even further by running on carbon-free energy 24/7 by 2030.  As we work to achieve 24/7 carbon-free energy, we help you take immediate action to decarbonize your digital applications and infrastructure. We’re also working with our customers across every industry to develop new solutions for the unique climate change challenges that organizations face. Today, we’re excited to expand our portfolio of carbon-free solutions and announce new partnerships that will help every company build a more sustainable future. 

First, we’re launching Carbon Footprint, a new product that provides customers with the gross carbon emissions associated with their Google Cloud Platform usage. Now available to every GCP user for free in the Cloud Console, this tool helps you measure, track and report on the gross carbon emissions associated with the electricity of your cloud usage. Of course, the net operational emissions associated with your Google Cloud usage is still zero. With growing requirements for Environmental Social and Governance (ESG) reporting, companies are looking for ways to show their employees, boards and customers their progress against climate targets. Using Carbon Footprint, you have access to the gross energy related emissions data you need for internal carbon inventories and external carbon disclosures, with one click. 

Built in collaboration with customers like Atos, Etsy, HSBC, L’Oréal, Salesforce, Thoughtworks and Twitter, our Carbon Footprint reporting introduces a new standard of transparency to support you in meeting your climate goals. You can monitor your gross cloud emissions over time, by project, by product and by region, giving IT teams and developers metrics that can help them reduce their carbon footprint. Our detailed calculation methodology is published so that auditors and reporting teams can verify that their cloud emissions data meets GHG Protocol guidance.

“The power of knowledge combined with the power of technology innovation plays a vital role in proactively responding to the climate crisis we are facing. With Google Carbon Footprint reporting, Atos feeds emissions data in our Decarbonization Data Platform, demonstrating potential emissions reductions from the Google Cloud Platform to our customers. This reporting opens up new levels of emissions transparency, trajectory planning, and data insight to support our customers in meeting, and potentially accelerating towards, their climate goals.”—Nourdine Bihmane, Head of Decarbonization Business Line, Atos

“The capability to measure and understand the environmental footprint of our Public Cloud usage is among the key axis of our sustainable tech roadmap. With Google Cloud Carbon Footprint, we are now able to directly follow the impact of our sustainable infrastructure approach and architecture principles.”—Hervé DUMAS, Sustainability IT Director, L’Oreal

While digital infrastructure emissions are just one part of your environmental footprint, accurately accounting for IT carbon emissions is necessary to measure progress against the carbon reduction targets required to avert the worst consequences of climate change. To help you account for emissions beyond our cloud and across your organization, we’re excited to partner with Salesforce Sustainability Cloud, integrating our Google Cloud Platform emissions data into their carbon accounting platform. 

“As we face unprecedented climate challenges, companies across the globe need to embed sustainability into the core of their business in order to meet growing customer and stakeholder expectations, and reduce their environmental impact. Together, Google Cloud and Salesforce Sustainability Cloud can help our joint customers accelerate their path to Net Zero, leveraging data-driven insights and visualizations to track and reduce their carbon emissions to drive sustainable change.”—Ari Alexander, GM of Salesforce Sustainability Cloud. 

From information to action

With the gross energy-related emissions footprint of data associated with your Google Cloud usage now available, we’re committed to providing tools to not only measure your carbon footprint, but help you reduce it. We recently launched low-carbon region icons to help you choose cleaner regions to locate your Google Cloud resources. New users who see the icons are over 50% more likely to choose clean regions over others, ensuring their applications emit less carbon over time. 

For current Google Cloud users, we’re pleased to announce that Active Assist Recommender will include a new sustainability impact category, extending its original core pillars of cost, performance, security, and manageability. Starting with the Unattended Project Recommender, you’ll soon be able to estimate the gross carbon emissions you’ll save by removing your idle resources. Unattended Project Recommender uses machine learning to identify, with a high degree of confidence, projects that are likely abandoned based on API and networking activity, billing, usage of cloud services, and other signals, and provides actionable recommendations on how to remediate those abandoned projects. By deleting these projects, not only can you reduce costs and mitigate security risks, but you can also reduce your carbon emissions. In August, Active Assist analyzed the aggregate data from all customers across our platform, and over 600,000 gross kgCo2e was associated with projects that it recommended for cleanup or reclamation. If customers deleted these projects they would significantly reduce future gross carbon emissions. Check out this blog to learn more about Active Assist.

Solutions for climate resilience 

Many of our customers face difficult questions about how their business impacts the natural environment today, and how it will be affected by climate change in the future. Answering these questions requires rich datasets about the planet, better analytics tools and smarter models to predict potential outcomes. For over a decade Google Earth Engine has supported scientists and developers with hyperscale computing power and the world’s largest catalog of satellite image data. Today, we are delighted to announce the preview of Earth Engine as part of Google Cloud Platform. Now, you can access Earth Engine and combine it with other geospatial-enabled products like BigQuery. By extending Earth Engine’s powerful platform to enterprises through Google Cloud, we are bringing the best of Google together. 

Over the past year we’ve worked with a number of organizations to use Earth Engine technology with tools like BigQuery and the Cloud AI Platform to develop new solutions for responsible commodity sourcing, sustainable land management and carbon emissions reduction. Earth Engine enables companies to track, monitor and predict changes in the Earth’s surface due to extreme weather events or human-caused activities, thus  helping them save on operational costs, mitigate and better manage risks, and become more resilient to climate change threats. This new offering will wrap the unique data, insights and functionality of Earth Engine with a fully-managed, enterprise-grade experience and reliability.

As we work with our customers to accelerate their sustainability initiatives, earth observation data is proving critical to effectively plan for the long-term impacts of climate change. To extend our geospatial and sustainability use cases we’re also expanding our partnerships with CARTO, Climate Engine, Geotab, NGIS, and Planet to bring their data and core applications to Google Cloud.

These partners will each make their existing platforms and datasets available globally on Google Cloud, giving you low-latency and reliable access to critical data and applications that will inform your sustainability initiatives. By integrating water availability, agricultural data, weather risks, and extensive daily satellite imagery into Earth Engine and BigQuery, you can achieve more ambitious goals for the sustainability of your business and our planet.

Committing to help you meet your climate goals

With each of these tools, we’re working to reduce the barriers you face in adopting more sustainable technology practices. We understand that building more sustainable applications and infrastructure is not easy. You face competing priorities, technical challenges, and the perception that climate action is costly. 

It doesn’t have to be this way. Today, we are making a sustainability pledge to you: teams across Google Cloud are committing to eliminating the barriers you face in building a more sustainable digital future for your organization, and will help you take action today to realize your climate goals. We’ll do this in a number of ways: 

1. In digital transformation projects and workshops, sustainability teams will always have a seat at the planning table, so we can work together on using cloud technology to build a more sustainable future. 
2. We’re putting low-carbon signals natively into our products to help developers choose more sustainable options early in their application development. 
3. We’ll ensure carbon impact is measured consistently with other key performance indicators. Leveraging the social cost of carbon, the ROI models and value assessments you conduct with Google Cloud will project your emissions impact too. 
4. We’ll be transparent about our carbon impact, by publishing third-party reviewed reports and methodologies, so you can trust the data for your own reports and disclosures. 
5. We’ll continue to work with the industry on best practices, including educational resources like Sustainable IT – Decoded, a new masterclass created in partnership with Intel, that shares the expertise of sustainability thought leaders. 

For the next decade we need to work together to avert the worst consequences of climate change. We’ve made tremendous progress in building technology that helps everyone do more for the planet, and we’re excited to see what you do with it. Visit this page to learn more about Google Cloud’s sustainability efforts.

Digitec Galaxus AG is the biggest online retailer in Switzerland, operating two online stores: Digitec, Switzerland’s online market leader for consumer electronics and media products, and Galaxus, the largest Swiss online shop with a steadily growing range of consistently low-priced products for almost all daily needs. 

Known for its efficient, personalized shopping experiences, it’s clear that Digitec Galaxus understands what it takes to deliver a platform that is interesting and relevant to customers every time they shop. 

The problem: Personalizing decisions for every situation

Digitec Galaxus already had established an engine to help them personalize experiences for shoppers when they reached out to Google Cloud. They had multiple recommendation systems in place and were also extensive early adopters of Recommendations AI, which already enabled them to offer personalized content in places like their homepages, product detail pages, and their newsletter. 

But those same systems sometimes made it difficult to understand how best to combine and optimize to create the most personalized experiences for their shoppers. Their requirements were threefold:

1. Personalization: They have over 12 recommenders they can display on the app, however they would like to contextualize this and choose different recommenders (which in turn select the items) for different users. Furthermore they would like to exploit existing trends as well as experiment with new ones.
2. Latency: They would like to ensure that the solution is architected so that the ranked list of recommenders can be retrieved with sub 50 ms latency.
3. End-to-end easy to maintain & generalizable/modular architecture: Digitec wanted the solution to be architected using an easy to maintain, open source stack, complete with all MLops capabilities required to train and use contextual bandits models. It was also important to them that it is built in a modular fashion such that it can be adapted easily to other use cases which have in mind such as recommendations on the homepage, Smartags and more . 

To improve, they asked us to help them implement a machine learning (ML) contextual bandit based recommender system on Google Cloud taking all the above factors into consideration to take their personalization to the next level. 

Contextual bandits algorithms are a simplified form of reinforcement learning and help aid real-world decision making by factoring in additional information about the visitor (context) to help learn what is most engaging for each individual. They also excel at exploiting trends which work well, as well as exploring new untested trends which can yield potentially even better results. For instance, imagine that you are personalizing a homepage image where you could show a comfy living room couch or pet supplies. 

Without a contextual bandit algorithm, one of these images would be shown to someone at random without considering information you may have observed about them during previous visits. Contextual bandits enable businesses to consider outside context, such as previously visited pages or other purchases, and then observe the final outcome (a click on the image) to help determine what works best. 

Creating a personalization system with contextual bandits

While Digitec Galaxus heavily personalizes their website homepages, they are very very sensitive and also require more cross-team collaboration to update and make changes. 

Together with the Digitec Galaxus team, we decided to narrow the scope and focus on building a contextual bandit personalization system for the newsletter first. The digitec Galaxus team has complete control over newsletter decisions and testing various ML experiments on a newsletter would have less chance of adverse revenue impact than a website homepage. 

The main goal was to architect a system that could be easily ported over to the homepage and other services offered by Digitec with minimal adaptations. It would also need to satisfy the functional and non-functional requirements of the homepage as well as other internal use cases.

Below is a diagram of how the newsletter’s personalization recommendation system works:

• The system is given some context features about the newsletter subscriber such as their purchase history and demographics. Features are sometimes referred to as variables or attributes, and can vary widely depending on what data is being analyzed. 
• The contextual bandit model trains recommendations using those context features and 12 available recommenders (potential actions). 
• The model then calculates which action is most likely to enhance the chance of reward (a user clicking in the newsletter) and also minimize the problem (an unsubscribe). 

Calculating whether a click was a newsletter or an unsubscribe enabled the system to optimize for increasing clicks and avoid showing non-relevant content to the user (click-bait). This enabled Digitec Galaxus to exploit popular trends while also exploring potentially better-performing trends. 

How Google Cloud helps

The newsletter context-driven personalization system was built on Google Cloud architecture using the ML recommendation training and prediction solutions available within our ecosystem. 

Below is a diagram of the high-level architecture used:

The architecture covers three phases of generating context-driven ML predictions, including: 

ML Development: Designing and building the ML models and pipeline 
Vertex Notebooks are used as data science environments for experimentation and prototyping. Notebooks are also used to implement model training, scoring components, and pipelines. The source code is version controlled in Github. A continuous integration (CI) pipeline is set up to automatically run unit tests, build pipeline components, and store the container images to Cloud Container Registry. 

ML Training: Large-scale training and storing of ML models 
The training pipeline is executed on Vertex Pipelines. In essence, the pipeline trains the model using new training data extracted from BigQuery and produces a trained, validated contextual bandit model stored in the model registry. In our system, the model registry is a curated Cloud Storage. 

The training pipeline uses Dataflow for large scale data extraction, validation, processing, and model evaluation, and Vertex Training for large-scale distributed training of the model. AI Platform Pipelines also stores artifacts, the output of training models, produced by the various pipeline steps to Cloud Storage. Information about these artifacts are then stored in an ML metadata database in Cloud SQL. To learn more about how to build a Continuous Training Pipeline, read the documentation guide.

ML Serving: Deploying new algorithms and experiments in production 
The training pipeline uses batch prediction to generate many predictions at once using AI Platform Pipelines, allowing Digitec Galaxus to score large data sets. Once the predictions are produced, they are stored in Cloud Datastore for consumption. The pipeline uses the most recent contextual bandit model in the model registry to evaluate the inference dataset in BigQuery and give a ranked list of the best newsletters for each user, and persist it in Datastore. A Cloud Function is provided as a REST/HTTP endpoint to retrieve the precomputed predictions from Datastore.

All components of the code and architecture are modular and easy to use, which means they can be adapted and tweaked to several other use cases within the company as well.

Better newsletter predictions for millions

The newsletter prediction system was first deployed in production in February, and Digitec Galaxus has been using it to personalize over 2 million newsletters a week for subscribers. The results have been impressive, 50% higher than our baseline. However, the collaboration is still ongoing to improve the results even more. 

“Working at this level in direct exchange with Google’s machine learning experts is a unique opportunity for us. The use of contextual bandits in the targeting of our recommendations enables us to pursue completely new approaches in personalization by also personalizing the delivery of the respective recommender to the user. We have already achieved good results in our newsletter in initial experiments and are now working on extending the approach to the entire newsletter by including more contextual data about the bandits arms. Furthermore, as a next step, we intend to apply the system to our online store as well, in order to provide our users with an even more personalized experience. To build this scalable solution, we are using Google’s open source tools such as TFX and TF Agents, as well as Google Cloud Services such as Compute Engine, Cloud Machine Learning Engine, Kubernetes Engine and Cloud Dataflow.”—Christian Sager, Product Owner, Personalization ( Digitec Galaxus)

Since the existing architecture and system is also dynamic, it will automatically adapt to new behaviours, trends, and users. As a result, Digitec Galaxus plans to re-use the same components and extend the existing system to help them improve the personalization of their homepage and other current use cases they have within the company. Beyond clicks and user engagement, the system’s flexibility also allows for future optimization of other criteria. It’s a very exciting time and we can’t wait to see what they build next!