On Wednesday, December 14, 2022 Google announced a new spam-prevention system called SpamBrain. This AI-based system leverages the power of artificial intelligence to detect sites buying links or passing out unnatural links as well as detecting spam directly. The update is referred to as the December 2022 link spam update and will take about two weeks to fully roll out. It will affect all languages and may cause changes in rankings as any credit gained from unnatural links is lost.
Google has always stressed that links primarily obtained for artificial manipulation of Search rankings are considered link spam and their algorithms and manual actions aim to nullify these links at scale. They ask that anyone who comes across sites engaging in inorganic link building report them via the ‘report-spam’ guidelines found here: https://goo.gle/sc-forum . Those with specific feedback regarding this update can post in the ‘help community’ found here: https://goo.gle/sc-forum . More information about SpamBrain can be found here: /search/blog/2022/04/webspam-report-2021#spambrain:-our-most-effective-solution-against-spam
Four authors, Duy Nguyen, Ildar Akhmedyanov, Jacob N Scott and Karthikgeyan Elangovan have posted an article discussing how techniques from reinforcement learning (RL) can be used to improve the performance of recommender systems. They focus on two particular RL techniques - Actor-Critic algorithms and Deep Reinforcement Learning (DRL) - and explain the benefits that each can bring to recommender systems. The authors also provide a detailed overview of recent deep reinforcement learning research for recommendation tasks and offer insight into potential challenges associated with its use. Finally, they discuss how RL methods can be combined with traditional collaborative filtering approaches to further enhance performance. This article is a useful resource for those looking to understand the role that reinforcement learning can play in improving their recommender system performance.