The article explains that AI‑generated prose often feels off because its structure—sentence flow, paragraph emphasis, and lack of a personal bias—doesn’t match human writing, and it offers concrete examples and practical steps for designers to identify the AI flavor.
We introduce an AIGC‑driven content analysis agent for Taobao that automatically builds a fine‑grained feature taxonomy, multimodally labels images, videos and text, mines causal quality features, and deploys auto‑prompt, memory and LoRA‑based training pipelines, achieving over 90% labeling accuracy, 80% labor savings and double‑digit CTR gains while cutting video generation cost to 25% of GPT‑4o.
This article details how Toutiao's senior algorithm architect designs a transparent recommendation system, covering system overview, three-dimensional feature modeling, real-time training pipelines, recall strategies, content analysis, user tagging, evaluation methods, and content safety measures.
The article details the end‑to‑end architecture of Toutiao's massive recommendation platform, covering system overview, content and user feature extraction, model training, recall strategies, evaluation methodology, and content safety mechanisms, while highlighting practical challenges and engineering solutions.
This article explains the architecture and principles of Toutiao’s recommendation system, covering its three‑dimensional model of content, user and environment features, content analysis techniques, user tagging, real‑time training pipelines, evaluation methods, and content safety measures that together drive personalized feeds.
The talk explains how iQIYI leverages AI—through video understanding, tag generation, graph construction, and scene‑based point selection—to automatically create and place diverse ad formats, optimizing ad inventory, relevance, and revenue while meeting brand and production constraints.
Youku’s AI‑powered Beidouxing system analyzed audience tags, attractiveness scores and performance data to select Lei Jiayin and Yi Yangqianxi for “The Longest Day in Chang’an”, guiding casting, episode frequency and other production choices while reducing subjective bias and expanding the talent pool.
This article presents a comprehensive overview of Toutiao's recommendation system, detailing its three‑dimensional modeling approach, real‑time training pipeline, feature engineering, content and user analysis techniques, evaluation methodology, and the extensive content‑safety mechanisms employed to ensure reliable and responsible information distribution.
