#Recommendation Systems
How VK Taught the Feed to Sell: 43-fold growth in orders through a separate recommendations contour
VK has created a specialized contour for product posts that can distinguish interest in content and the intention to buy. The result is 53% growth in the conversion of impressions to orders, and 43 times for the orders themselves.
Складной iPhone за $2000, битва за Китай и трекер мозга: главное из мира технологий
Apple готовится показать свой первый гибкий смартфон стоимостью свыше $2000, китайские техногиганты выкатывают новые складные флагманы, а Австралия объявляет войну навязчивым алгоритмическим лентам.
Why RecSys Metrics Lie: How Preprocessing Nuances Break Reproducibility and What to Do About It
Developers from Sber AI Lab and AIRI have introduced SplitLight, a tool for auditing datasets and splits in recommender systems. The experts proved that even basic sorting of events with identical timestamps can significantly change model quality without any modifications to its architecture.
Neural Ranking at VK: How Replacing Gradient Boosting with Neural Networks Boosted Recommendations
The VK AI team shared their experience transitioning from classic decision trees to a multi-task neural network architecture for content ranking, which increased watch time in VK Clips by 5.5%.
VK Unifies Service Recommendations Using a Single Neuroprofile
The VK AI team has implemented the "Heimdall" end-to-end transformer model, which connects user activity across VK Clips, VK Video, and the social network feed into a unified recommendation system.