Tagged articles
online training
11 articles · Page 1 of 1
Vision-OPD introduces an online self‑distillation framework that lets a 9B multimodal LLM internalize fine‑grained visual evidence from whole images, achieving state‑of‑the‑art results on six detailed visual‑understanding benchmarks without extra inference tools.
Vision-OPDbenchmarkfine-grained visual understanding
0 likes · 13 min read

Alibaba’s Advertising Ranking team introduced the Historical Data Reuse (HDR) algorithm, which automatically selects similar past promotion days, fine‑tunes the production CVR model with a TransBlock layer and distribution‑correction weighting, delivering up to 10 % AUC gains and double‑digit RPM, CVR, and ROI improvements during the 2022 Double‑11 event and offering a reusable solution for other domains facing abrupt user‑behavior shifts.
conversion rate predictiondistribution shifthistorical data reuse
0 likes · 30 min read

To eliminate data fragmentation, feature inconsistencies, and multi‑language implementation challenges, Bilibili built a unified real‑time advertising feature platform that aligns offline, hourly, and online pipelines via a shared C++ library and JNI, boosting CTR prediction accuracy, cutting training costs, and increasing ad revenue by over 1 %.
AdvertisingCTR predictionFlink
0 likes · 11 min read

This article presents Ctrip's end‑to‑end solution for precise ad‑inventory forecasting using an LSTM model combined with entity embedding, covering data preprocessing, K‑means clustering, model architecture, offline‑online incremental training, early‑stop mechanisms, evaluation metrics, and Python service deployment.
EmbeddingLSTMPyTorch
0 likes · 19 min read

Xiaohongshu’s in‑house LarC platform powers real‑time, multimodal recommendation, life‑search, and generative‑AI commercial content for its 200 million‑user community by processing billions of daily feedback samples, employing conflict‑free parameter servers, diversified sequence modeling, and large‑scale representation learning to deliver personalized, fresh, and diverse user experiences.
AI infrastructureMachine Learning PlatformRecommendation System
0 likes · 13 min read

This article presents a comprehensive overview of 58's recruitment recommendation system, detailing business challenges, multi‑stage recall strategies, vector‑based deep retrieval, cost‑sensitive loss design, session optimization, online incremental training, extensive offline and online evaluations, and practical lessons for future improvements.
AIRecommendation Systemcost-sensitive loss
0 likes · 15 min read

The 2020 Tencent Advertising Algorithm Contest has launched a series of live training sessions to help participants master the TI-ONE and Angel platforms, featuring expert demonstrations on May 15th.
Algorithm CompetitionAngel deep learning engineTI-ONE platform
0 likes · 3 min read


58 TechNov 29, 2019 · Artificial Intelligence
This article details the end‑to‑end optimization of 58.com’s commercial traffic ranking system, covering data‑flow upgrades, advanced feature engineering, real‑time and multi‑task model improvements, and a multi‑factor ranking mechanism, while sharing practical results and future directions.
RankingReal-time Data Pipelinefeature engineering
0 likes · 17 min read


58 TechOct 28, 2019 · Artificial Intelligence
This article details the comprehensive optimization of 58's commercial traffic ranking system, covering data‑flow upgrades, advanced feature engineering, model enhancements—including online training, multi‑task and relevance models—and a multi‑factor ranking mechanism that together improve monetization efficiency and user experience.
RankingReal-time Data Pipelinee-commerce
0 likes · 16 min read

This article provides a comprehensive introduction to recommendation systems, covering their definition, overall offline and online architectures, feature engineering, collaborative filtering, latent semantic models, ranking algorithms, and evaluation methods including A/B testing and offline metrics.
A/B testingRankingRecommendation System
0 likes · 28 min read

Alibaba leveraged reinforcement learning, highlighted by MIT Technology Review’s 2017 breakthrough list, to transform its e‑commerce search and recommendation systems during Double 11, deploying large‑scale online and batch training pipelines, dynamic market segmentation, and real‑time decision models that boosted click‑through rates by up to 20 %.
e-commercemachine learningonline training
0 likes · 14 min read
