Exclusive Content:

Haiper steps out of stealth mode, secures $13.8 million seed funding for video-generative AI

Haiper Emerges from Stealth Mode with $13.8 Million Seed...

“Revealing Weak Infosec Practices that Open the Door for Cyber Criminals in Your Organization” • The Register

Warning: Stolen ChatGPT Credentials a Hot Commodity on the...

VOXI UK Launches First AI Chatbot to Support Customers

VOXI Launches AI Chatbot to Revolutionize Customer Services in...

Getting Ready for the Unforeseen

Handling Unseen Values in Machine Learning: A Case Study on Taboola’s Recommender System

In the world of machine learning, one of the challenges that often arise is dealing with categorical features that represent real-world objects, such as words, items, and categories. However, what happens when we encounter new object values during the inference stage that have never been seen before? How can we ensure that our model can still make sense of these new inputs?

These unseen values, also known as Out of Vocabulary (OOV) values, must be handled appropriately. Different algorithms have different methods for dealing with OOV values, and it’s important to consider the assumptions made about the categorical features as well.

In this blog post, we’ll focus on the application of deep learning to dynamic data, using Taboola’s recommender system as an example. This system encounters new values regularly, such as unique item identifiers and advertiser IDs. These unseen values pose a challenge as they were not present during the model’s training phase.

One solution to handling OOV values is to replace all rare values with a special OOV token before training. By exposing the model to the OOV token during training, it can learn a meaningful embedding for all OOV values and mitigate the risk of overfitting to rare values.

However, simply using an OOV token may not be enough to ensure optimal performance. Rare items that are injected with the OOV token may not benefit from the model’s memorization capabilities, leading to poorer performance on these items. Furthermore, if the OOV embedding is learned using a distribution specific to rare items, it may not generalize well to the general population of items.

To address this issue, Taboola’s recommender system implemented a new approach. Instead of randomly injecting the OOV token before training, the model trained on all available values during each epoch. At the end of the epoch, a random set of examples were sampled, and the OOV token was injected for further training. This allowed the model to benefit from both OOV and non-OOV embeddings, improving performance significantly.

By continuously seeking improvements and considering unexpected challenges, the recommender system was able to enhance its performance in production. This case highlights the importance of continuously exploring new approaches and fine-tuning models to achieve optimal results in machine learning applications. To read more about this approach, you can find the original post on engineering.taboola.com.

Latest

I Asked ChatGPT About the Worst Money Mistakes You Can Make — Here’s What It Revealed

Insights from ChatGPT: The Worst Financial Mistakes You Can...

Can Arrow (ARW) Enhance Its Competitive Edge Through Robotics Partnerships?

Arrow Electronics Faces Growing Challenges Amid New Partnership with...

Could a $10,000 Investment in This Generative AI ETF Turn You into a Millionaire?

Investing in the Future: The Promising Potential of the...

Don't miss

Haiper steps out of stealth mode, secures $13.8 million seed funding for video-generative AI

Haiper Emerges from Stealth Mode with $13.8 Million Seed...

VOXI UK Launches First AI Chatbot to Support Customers

VOXI Launches AI Chatbot to Revolutionize Customer Services in...

Investing in digital infrastructure key to realizing generative AI’s potential for driving economic growth | articles

Challenges Hindering the Widescale Deployment of Generative AI: Legal,...

Microsoft launches new AI tool to assist finance teams with generative tasks

Microsoft Launches AI Copilot for Finance Teams in Microsoft...

Tailoring Text Content Moderation Using Amazon Nova

Enhancing Content Moderation with Customized AI Solutions: A Guide to Amazon Nova on SageMaker Understanding the Challenges of Content Moderation at Scale Key Advantages of Nova...

Building a Secure MLOps Platform Using Terraform and GitHub

Implementing a Robust MLOps Platform with Terraform and GitHub Actions Introduction to MLOps Understanding the Role of Machine Learning Operations in Production Solution Overview Building a Comprehensive MLOps...

Automate Monitoring for Batch Inference in Amazon Bedrock

Harnessing Amazon Bedrock for Batch Inference: A Comprehensive Guide to Automated Monitoring and Product Recommendations Overview of Amazon Bedrock and Batch Inference Implementing Automated Monitoring Solutions Deployment...