AI Briefing#recommendation-system
The latest AI and developer news about #recommendation-system, with the original source and a short summary.
The latest AI and developer news about #recommendation-system, with the original source and a short summary.
Pinterest improved ad relevance and ROAS with a sequential recommendation model that integrates real-time context.
Meta improved recommendation system inference efficiency by eliminating user embedding duplication with IKBO.
Pinterest redesigned its conversion-focused candidate generation model, boosting RoAS and conversion performance.
Bluesky's For You feed runs on a single Go server and SQLite sitting in a living room.
Pinterest optimized feed diversity and quality together, evolving from DPP to SSD and soft spacing.
Meta scaled ad ranking to LLM scale through request routing and hardware co-design.
By using an LLM and knowledge distillation to quantify the authenticity of reviews, the star rating inflation problem was addressed and the discriminative power of recommendations was improved.
Kurly introduced a Collaborative Embedding model with Knowledge Distillation applied, achieving approximately 140% ARPU increase on sale pages.
Azar used PolicyEngine to cut experiment setup and review time from 2 days to 30 minutes.
LINE VOOM built a Milvus-based real-time recommendation system to solve the problem of delayed content reflection.
To boost new user acquisition and trendy product appeal, Misikgwan was restructured around toggle navigation, personalization, and community.
To overcome the limitations of existing attribute-based similarity calculations, Oliveyoung developed a semantic product similarity model that leverages user search logs.
Kakao revealed a real-time ad serving bucket system and LLM-based spam classification with reasoning capabilities.
Kurly's Data Service Development Team built a zero-downtime real-time serving system for its cart recommendation model through TorchServe tuning and GracefulShutdown implementation.
Kurly introduced a complementary-product recommendation model for the shopping cart, raising conversion rate and purchase amount by about 100%.
The event will be held on December 12 at Kakao Agit in Pangyo, with online live streaming and recorded videos also provided.
At KSC2022 held in Jeju, Kakao's recommendation team and cloud platform team conducted technical sessions and operated a recruitment booth.
The Kakao Ad Recommendation Team integrates six optimization models and real-time data pipelines into an MLOps structure to build an advanced serving system for large-scale RTB environments.
KakaoWebtoon improved recommendation accuracy by combining text, image, and art style embeddings with a multi-armed bandit algorithm.
An account of working as an intern on the Kakao Recommendation Team for 2 months, improving recommendation services and conducting R&D projects, leading to a transition to a full-time crew member
The method employs a weighted average of topic-specific MABs to reflect real-time feedback and optimize long-term rewards.
When Melon's similar song recommendation is used, the weekly average number of streams increases by 24% and the number of unique songs by 53%.
Kakao held the second session of its internal technical seminar Techtalk, sharing experiences on TensorFlow serving optimization and large-scale data training.
Kakao unveiled its AI recommendation system technology and future vision, highlighting the importance of information filtering technology and outlining future development directions.