Tagged articles
large-scale AI
15 articles · Page 1 of 1
The TAAC × KDD Cup 2026 workshop in Jeju Island gathered researchers and engineers from academia and leading tech firms to dissect real‑world recommendation challenges, present the competition’s unified sequence‑modeling and feature‑interaction task, and showcase winning solutions that balance effectiveness, efficiency, and scalability.
KDD CupScaling LawTAAC 2026
0 likes · 7 min read

At WAIC, China’s first fully domestic 100,000‑GPU AI supercluster, the Shuguang 8000, was shown to run at full capacity within its first week—processing over 150,000 jobs daily and peaking above 500,000—thanks to intelligent scheduling, high‑density cabinets, proprietary RDMA interconnects, top‑ranked storage and immersion cooling, while offering open access programs for researchers and enterprises.
AI supercomputingHigh-Performance ComputingShuguang 8000
0 likes · 4 min read

JD’s Oxygen AIIC combines human‑in‑the‑loop ontology engineering, a semantic search‑then‑discrimination pipeline, and a self‑evolving multi‑task LLM/VLM model to produce high‑quality product knowledge for over a hundred thousand categories and billions of daily SKU updates, boosting search coverage to 80%, attribute auto‑fill to over 80%, cutting quality issues by 37% and raising click‑through by 9% while achieving 94.2% precision and 82.8% recall.
JD.comLLMSemantic Search
0 likes · 21 min read

xAI’s massive fleet of roughly 550,000 Nvidia H100 and H200 GPUs in its Memphis and Colossus data centers is operating at a mere 11% model FLOPs utilization, highlighting how scaling to hundreds of thousands of GPUs creates coordination, network, and scheduling bottlenecks that waste most of the hardware’s compute power.
AI infrastructureGPU utilizationNVIDIA H100
0 likes · 5 min read

Decoupled DiLoCo, a new distributed training framework introduced by Jeff Dean and colleagues, enables resilient large‑scale AI pre‑training across heterogeneous hardware by decoupling learners, using lightweight syncers, adaptive quorum, and balanced tensor fragmentation, dramatically improving goodput and reducing bandwidth while preserving model quality.
Bandwidth ReductionDecoupled DiLoCoGoodput
0 likes · 10 min read


JD TechJan 31, 2026 · Artificial Intelligence
This article details JD Retail's 9N‑LLM unified training framework that tackles the massive data, hardware heterogeneity, and algorithmic challenges of generative recommendation by integrating TensorFlow and PyTorch, supporting GPU/NPU, and delivering high‑throughput sample processing, sparse/dense optimization, and flexible reinforcement‑learning capabilities.
GPU/NPURaylarge-scale AI
0 likes · 26 min read

RecIS is an open‑source, PyTorch‑based unified framework designed for ultra‑large‑scale sparse‑dense computation in recommendation systems, offering a full solution for training models with massive samples, multimodal inputs, and large embeddings, and demonstrating significant performance gains over TensorFlow and TorchRec in production deployments.
PyTorchdeep learning frameworklarge-scale AI
0 likes · 24 min read

This article examines the key pain points of massive AI compute clusters—including heterogeneous hardware compatibility, efficient scheduling, training and inference acceleration, and fault‑tolerant operations—while presenting practical management and performance‑tuning strategies, a cloud‑native AI platform implementation, and future directions for the ecosystem.
AI computingOperationsPerformance Tuning
0 likes · 7 min read


NewBeeNLPJul 5, 2024 · Artificial Intelligence
Meta’s new paper introduces the Wukong model, demonstrating that expanding dense‑layer parameters and computational FLOPs in large‑scale recommendation systems follows a clear scaling law, yielding consistent performance gains across massive internal datasets, with detailed analysis of feature modules, parameter impacts, and experimental results.
CTR modelsMetaScaling Law
0 likes · 10 min read

This article presents Baidu's comprehensive approach to feed recommendation ranking, covering business and data background, feature engineering principles, core algorithmic strategies, system architecture design, and upcoming plans to integrate large language models for more intelligent and fair recommendations.
Baidufeature engineeringlarge-scale AI
0 likes · 19 min read

The paper introduces SHARK, a two‑component framework that uses a fast Taylor‑expanded permutation method to prune embedding tables and a frequency‑aware quantization scheme to apply mixed‑precision to embeddings, achieving up to 70% memory reduction and 30% QPS improvement in industrial short‑video and e‑commerce recommendation systems.
Efficiencyembedding compressionlarge-scale AI
0 likes · 8 min read


AntTechFeb 24, 2023 · Artificial Intelligence
The 2022 China Institute of Electronics (CIEE) Science and Technology Award recognized a collaborative project between Beijing University of Posts and Telecommunications and Ant Group for pioneering large-scale heterogeneous graph neural network models, a trillion‑scale dynamic graph learning system, and extensive industry applications, earning top honors, patents, papers, and standards.
Graph Neural NetworksTechnology Awardheterogeneous graphs
0 likes · 4 min read

Tencent engineers broke the world record for training ImageNet with 128 V100 GPUs in just 2 minutes 31 seconds, detailing a suite of optimizations—including a new Light distributed training framework, single‑machine speed boosts, multi‑machine communication enhancements, and advanced batch convergence techniques—that together dramatically cut training time while maintaining high accuracy.
GPUImageNetTencent Cloud
0 likes · 9 min read


21CTOFeb 18, 2020 · Artificial Intelligence
This article details Toutiao’s large‑scale recommendation system, explaining how it models content, user, and environment features, the variety of algorithms and real‑time training pipelines used, feature engineering categories, recall strategies, content analysis, user tagging, evaluation methods, and content‑safety mechanisms.
Real-time TrainingRecommendation Systemcontent safety
0 likes · 18 min read

The Alibaba iDST team clinched first place in the ACM MM2017 LSVC competition by leveraging Alibaba Cloud’s ODPS to extract eight multimodal features, achieving a 0.8485 mAP on the validation set, and demonstrating the critical role of rich modality fusion in large‑scale video classification.
AlibabaODPSlarge-scale AI
0 likes · 5 min read
