This talk introduces automated knowledge graph representation learning, covering background, key techniques such as triple‑based, path‑based and subgraph‑based models, AutoML‑driven model search (AutoSF, Interstellar, RED‑GNN), evaluation metrics, and future research directions in AI.
This article reviews three representative document intelligence models—StrucText, LayoutLMv3, and GraphDoc—detailing their input features, feature fusion strategies, self‑supervised tasks, and underlying architectures, and explains how they learn embeddings for segments, words, or regions to enable classification and key‑value extraction.
This article presents the DIRN model, which leverages product inclusion graphs and graph‑based embeddings to address the challenges of ranking both single‑item and complex travel products on Fliggy, demonstrating significant CTR, CVR, and GMV improvements through offline experiments and online A/B testing.
This talk introduces graph deep learning techniques for natural language processing, covering the motivation for graph representations, traditional graph-based NLP methods, fundamentals of graph neural networks, static and dynamic graph construction, representation learning, and showcases the open‑source Graph4NLP Python library with example applications.
This talk presents a deep‑learning framework that formulates financial index‑fund construction as a sparse portfolio optimization problem, solves the mixed‑integer programming via a two‑stage graph‑neural‑network pipeline, and demonstrates superior tracking performance and scalability on large‑scale index datasets.
This article explains how Kuaishou leverages graph neural networks for item‑to‑item (I2I) recall in short‑video recommendation, detailing the system background, pipeline architecture, optimization techniques such as similarity measurement, graph structure learning, edge‑weight learning, and future research directions.
This article traces the development of embedding methods—from the early word2vec model through item2vec, DeepWalk, Node2vec, EGES, HERec, GraphRT, and target‑fitting approaches like DSSM and YouTube recommendation—highlighting how sequence‑construction and target‑fitting paradigms have shaped modern recommendation systems and AI applications.
This article presents a comprehensive study of graph‑based recall techniques for content recommendation, detailing how knowledge‑graph‑augmented user‑behavior graphs and novel attention‑driven models such as GADM, SGGA, and SGGGA improve performance for users with sparse interaction histories.
The DataFun Summit 2022 Knowledge Graph Forum, held on March 12, presents cutting‑edge research on knowledge graph representation learning, multi‑hop reasoning, temporal KG question answering, and their applications in finance and retail, featuring talks by leading experts from JD, Fourth Paradigm, Stanford, and Meituan.
This article reviews the motivation, methods, and results of graph-based self‑supervised pretraining for molecular data, introduces the ChemRL‑GEM model that incorporates 3‑D structural information, and demonstrates its superior performance on ADMET, affinity prediction, and benchmark competitions using the PaddleHelix platform.
The paper proposes graph‑based deep recall models that enrich sparse user behavior sequences in video recommendation by integrating content knowledge graphs and adaptive attention mechanisms, demonstrating that variants such as GADM, SGGA, and SGGGA significantly boost click‑through rates in online experiments.
Recent advances in deep learning have enabled a novel two‑stage approach for designing financial index funds, where graph neural networks first select a sparse set of assets and then allocate weights, dramatically reducing computational complexity while achieving performance comparable to traditional mixed‑integer programming methods.
This article presents an end‑to‑end solution that leverages graph neural networks and dynamic bipartite graph construction to detect payment fraud in eBay's e‑commerce platform in real time, addressing traditional model limitations, graph latency challenges, and demonstrating superior performance over GBDT approaches.
This article introduces Baidu's PGL2.2 graph learning platform, explains graph modeling and message‑passing GNN techniques, details training strategies for small, medium and large graphs, showcases node classification and link‑prediction methods, and describes how the framework is applied in search, recommendation, risk control, and knowledge‑graph competitions.
The Alibaba Mama technical team secured five paper acceptances at The Web Conference 2023, presenting advances in unbiased delayed‑feedback conversion modeling, uncertainty‑regularized knowledge‑distilled CVR debiasing, feature‑aware probability calibration, coordinated two‑stage ad auctions, and scalable decoupled graph neural networks for large‑scale e‑commerce retrieval.
Galileo is an open‑source, distributed graph deep‑learning framework that supports ultra‑large heterogeneous graphs, dual TensorFlow/PyTorch back‑ends, and a flexible API, enabling fast prototyping of graph neural networks such as HeteSAGE for real‑world recommendation and other AI scenarios.
The 2021 review of Alibaba’s Mama Search Advertising estimation platform details advances in model deepening—such as hash‑based embedding compression, adaptive dynamic parameters and graph neural networks—model interaction via a multi‑stage cascade with ranking distillation and oracle bias, and system efficiency gains from HPC training, mixed‑precision, multi‑hash embeddings, and fp16 quantization that deliver roughly a thirty‑fold speed‑up.
This talk introduces the motivation, design, theoretical analysis, and extensive experimental results of Tencent Angel Graph's Graph Attention Multi‑Layer Perceptron (GAMLP) and Node‑Dependent Local Smoothing (NDLS), which address GNN scalability and flexibility by using node‑wise adaptive propagation, attention‑based feature fusion, and a lightweight training pipeline.
This article provides a comprehensive overview of fraud detection using graph neural networks, covering background definitions, fraud categories, GNN application steps, a timeline of key research papers, practical challenges, solutions, and a collection of open‑source resources and datasets.
This article presents a comprehensive study on detecting malicious webpages using heterogeneous graph structures and Graph Convolutional Networks, detailing background challenges, technical approaches, model iterations, optimization techniques for large‑scale deployment, experimental results, and directions for future research.
This article reviews the fundamentals of fraud, surveys the evolution of graph neural network research for fraud detection, outlines practical application steps, discusses key challenges such as disguise, scalability, and label scarcity, and provides representative papers, new research directions, industrial case studies, and open-source resources.
This article introduces graph representation learning, explains the Graph Information Bottleneck (GIB) framework for obtaining robust graph embeddings, and presents AD‑GCL, a contrastive learning method that leverages GIB principles to improve graph neural network performance without requiring task labels.
This article presents a comprehensive study on detecting malicious webpages by constructing heterogeneous graphs from URL redirection and textual features, applying Graph Convolutional Networks and Cluster‑Text‑GCN models, detailing optimization techniques for large‑scale deployment, and outlining future research directions.
This talk presents a series of graph‑based techniques for natural language processing, including the Story Forest system for hot event discovery, the GIANT framework for ontology creation and user interest modeling, and a divide‑and‑conquer approach to long‑text matching that leverages graph neural networks and community detection.
This article explains the fundamentals of graph neural networks, covering graph concepts, node classification via neighborhood aggregation, message‑passing mechanics, mathematical notation, a full DGL‑PyTorch implementation on the Reddit dataset, and training results showing accuracy improvements up to 91 %.
The CNCC 2021 forum brought together leading academics and industry experts to discuss advances in graph neural networks, graph computing for quantum chemistry, and practical applications of knowledge‑graph reasoning in sectors such as real‑estate and online video, showcasing both research breakthroughs and industrial deployment strategies.
This article introduces graph representation learning, explains the Graph Information Bottleneck (GIB) framework for obtaining robust graph embeddings, and presents AD‑GCL, a contrastive learning method that leverages GIB principles to improve robustness and performance without requiring task labels.
This article introduces knowledge‑enhanced graph semantic understanding techniques, detailing the ERNIE pre‑training models, GraphSAGE, and the ERNIESage family of node, edge, and multi‑neighbor architectures, and demonstrates their explicit and implicit knowledge integration in industrial search, recommendation, and map‑POI applications.
This article presents a curated collection of high‑star GitHub "awesome" repositories covering knowledge graph fundamentals, relation extraction, KG‑QA, graph construction, graph neural networks, dynamic graph learning, and multimodal knowledge graphs, providing links, summaries, and key resources for researchers and practitioners.
This presentation by Stanford PhD student Huang Kexin explores the challenges and innovations of applying graph machine learning to molecular and biomedical networks, introducing specialized GNN architectures, actionable hypothesis generation, domain‑scientist interfaces, few‑shot learning, and the Therapeutics Data Commons for accelerating drug discovery.
This talk by a Stanford PhD student explores how graph neural networks can be adapted for molecular and biomedical networks, discusses the limitations of standard GNNs, introduces novel methods such as SkipGNN and G‑Meta, and demonstrates their use for drug‑drug interaction prediction, hypothesis generation, and treatment discovery with few‑shot learning.
This article details how Shuhe Technology leveraged large‑scale graph neural networks, built with DGL and PyTorch, to improve financial fraud detection by preparing massive relationship graphs, pruning sparse nodes, extracting rich features, addressing class imbalance, and achieving a stable AUC gain of about four points.
This article describes how Shuhe Technology leveraged graph neural networks to improve financial risk assessment by preparing massive relational graph data, selecting DGL as the development framework, designing a GraphSage‑GAT model, addressing data sparsity and imbalance, and achieving notable AUC gains over traditional methods.
This issue presents deep dives into HarmonyOS ACE UI architecture, common infinite-loop pitfalls of React’s useEffect hook, a Three.js tutorial for building a 3D room, an introductory guide to graph neural networks, and a fundamentals‑focused machine‑learning primer, offering practical insights across front‑end, graphics, and AI domains.
This article explains how contrastive learning, originally popular in computer‑vision, can be interpreted and applied to recommendation‑system recall and coarse‑ranking models, covering its theoretical roots, typical architectures like SimCLR, MoCo and SwAV, and practical tricks such as in‑batch negatives, embedding normalization, temperature scaling, and graph‑based extensions.
This article reviews recent research on graph self‑supervised learning, focusing on contrastive learning fundamentals, the SimCLR‑style framework, representative models such as GRACE and its adaptive augmentation extension GCA, experimental evaluations, and future directions for graph contrastive methods.
This article describes how Baidu's account security team leverages large‑scale graph technology and graph neural networks to detect and characterize black‑industry cheating gangs, presents a customized GraphSAGE link‑prediction model, and evaluates its superiority over MLP and GCN embeddings for downstream risk‑control tasks.
To curb large‑scale, organized fraud on Baidu’s platform, the Account Security team built a scalable heterogeneous graph framework that links accounts, features, and devices, trains GraphSAGE‑based node embeddings via link‑prediction, and leverages these representations to uncover fraud gangs, boosting detection accuracy above 90% across billions of nodes.
The talk presents an overview of graph neural network development, explains 极验's graph modeling research and evolution, and details offline and real‑time GCN solutions, including self‑supervised training, large‑scale handling, and performance comparisons, highlighting practical applications in fraud detection and risk control.
Alibaba Mama announced the open‑source release of its Curvature Space Learning Framework and a Federated Learning Solution, promising up to 80% storage savings, 15% matching‑precision gains, enhanced privacy through federated computation, and broader applicability across search advertising, logistics, finance and healthcare.
This article explains how large‑scale UGC data is used to build a multi‑type interest point graph, describes the mining, hierarchical and associative relationship extraction methods, and demonstrates how the graph improves content understanding and recommendation accuracy while mitigating filter‑bubble effects.
This article introduces graph fundamentals, various application scenarios such as science, code logic, Spark workflows, social networks, and event graphs, then details graph data modeling, analysis, matrix computations, and the deployment of graph neural networks using frameworks like DGL, highlighting practical engineering considerations.
The article surveys six Alibaba Mama papers accepted at CIKM 2021, presenting novel AI methods—including a heterogeneous graph neural network for keyword matching, a star‑topology multi‑domain CTR model, a compact hash embedding technique, adaptive masked twins layers, automated hierarchical conversion prediction, and a scalable multi‑view ad retrieval system—each demonstrating substantial online performance improvements and large‑scale deployment.
This article reviews the fundamentals of entity‑relation extraction, surveys joint extraction techniques such as sequence labeling, table‑filling and seq2seq models, discusses document‑level graph‑based methods, highlights experimental findings, and outlines future research directions in knowledge‑graph construction.
The article explains how eBay protects its global payment system using a graph‑neural‑network driven risk management framework called xFraud, which combines heterogeneous graph sampling, node‑type encoding, attention mechanisms and dynamic‑graph extensions to detect and explain both individual and organized fraud patterns in real‑time.
NeuralSparse is a supervised graph sparsification framework that removes task-irrelevant edges to improve GNN generalization, combining a sparsification network with downstream GNN training, and demonstrates superior performance across multiple graph benchmarks compared to random edge dropping and other sparsification methods.
The SURGE framework leverages graph neural networks to construct and pool interest graphs from user interaction sequences, achieving stable and fast convergence, robust long‑sequence modeling, and significant performance gains over existing sequential recommendation methods on e‑commerce and short‑video datasets.
The article introduces DGL Operator, an open‑source Kubernetes‑based controller that automates the lifecycle of distributed graph neural network training with DGL, explains its terminology, challenges of native DGL distribution, and provides detailed architecture, workflow, and YAML/CLI examples for easy deployment.
DGL Operator is an open‑source Kubernetes controller that automates the lifecycle of distributed graph neural network training by handling configuration generation, graph partitioning, training execution, and resource cleanup, providing a cloud‑native solution for large‑scale GNN workloads.
This comprehensive survey reviews the latest research on graph neural networks applied to natural language processing, covering graph construction methods, graph representation learning techniques, encoder‑decoder models, static and dynamic graph building, and discusses challenges, benchmarks, and future directions in the field.
This article reviews recent academic and industrial advances in click‑through rate prediction, classifying optimization techniques for the three‑layer CTR architecture—Embedding, Hidden, and Output—while summarizing three SIGIR papers on graph‑based user behavior modeling, explicit semantic cross‑feature learning, and learnable feature selection for pre‑ranking.
The article presents a comprehensive anti‑fraud framework that analyzes black‑market fraud characteristics, reviews conventional fraud‑mitigation methods, and proposes a multimodal graph‑neural‑network approach—leveraging device, behavior, and content similarity—to accurately identify fraudulent users on large‑scale internet platforms.
This article explains how NetEase Cloud Music tackled cold‑start recommendation challenges in live streaming by leveraging Baidu's PGL distributed graph learning framework to train massive graph neural networks that transfer user behavior from music domains to live content, achieving significant performance gains.
This article introduces the fundamentals of graph neural networks, explains their expressive power for relational risk identification, and details Ant Group's innovations—including the GeniePath architecture and a bandit‑based sampling optimizer—that achieve superior performance on benchmark datasets.
This article introduces graph neural networks, explains their underlying sampling and aggregation mechanisms, and demonstrates how they are applied in large‑scale recommendation scenarios such as video and content feeds at Tencent, highlighting practical results and lessons learned.
GraphScope, an open‑source one‑stop platform from Alibaba DAMO Academy, unifies interactive queries, graph analytics, and graph learning on massive, rapidly evolving graphs, offering high‑performance distributed memory management, Gremlin optimization, and seamless Python integration to tackle real‑world AI and big‑data challenges.
This article systematically reviews mainstream embedding technologies—including matrix factorization, static and dynamic word embeddings, and graph‑based methods—explaining their principles, implementations, and practical applications in recommendation, advertising, and search systems.
This article describes how graph theory and graph convolutional neural networks are leveraged to model user‑IP relationships, detect fraudulent advertising clusters, and improve detection accuracy and recall through a combination of unsupervised graph algorithms and supervised GCN training in a large‑scale ad‑anti‑fraud system.
This week’s developer roundup covers JD’s open AI‑driven supply‑chain decision engine, LinkedIn’s Dagli Java machine‑learning library, PayPal’s crypto trading rollout, SpaceX’s Starlink service approval in Canada, AppleWatch’s NightWare nightmare‑interrupt feature, Intel’s new server GPU, and two cutting‑edge GNN research papers on multivariate time‑series forecasting and molecular generation.
The article introduces Galileo, JD Retail's distributed graph deep‑learning platform that supports heterogeneous and dynamic graphs, ultra‑large scale training, flexible model customization, and seamless integration with TensorFlow and PyTorch, highlighting its architecture, core challenges, built‑in algorithms, and upcoming open‑source release.
This article reviews how skip‑gram‑based graph embedding methods such as DeepWalk, LINE and node2vec can be interpreted as matrix factorization, explains the NetMF and NetSMF frameworks that use sparse matrix approximations and random SVD for large‑scale networks, and discusses extensions like GATNE and deep clustering approaches to address practical challenges in constructing and applying graph representations.
The team built an automated graph learning framework that preprocesses diverse graphs, employs four GNN architectures, conducts rapid hyper‑parameter tuning, and fuses models with density‑aware weighting, securing first place in KDD Cup 2020 AutoGraph and boosting Meituan’s ad recall and CTR prediction.
Alibaba Entertainment Group secured four ACM MM paper acceptances, presenting a probabilistic graphical model for crowdsourced visual quality assessment, an attention‑driven Siamese network with reinforcement learning for robust object tracking, a scene‑aware context‑graph method for unsupervised video anomaly detection, and a cross‑modal graph‑matching approach for visual grounding.
The article explains how the GEAR model uses graph neural networks and BERT representations to aggregate multiple pieces of evidence for fact verification, improving accuracy on datasets like FEVER and offering applications in misinformation detection, knowledge‑graph completion, and advertising analytics.
Based on academic and industry experience, iQIYI has designed a deep semantic representation learning framework that integrates multimodal side information and deep models such as Transformers and graph neural networks, improving recall, ranking, deduplication, diversity and semantic matching across recommendation and search scenarios.
iQIYI’s deep semantic representation learning framework integrates multimodal content, knowledge graphs, and user behavior through layered data, feature, strategy, and application components, employing early, late, and hybrid fusion with Transformers, GCNs, and other deep models to deliver high‑quality embeddings that boost recommendation, search, and streaming performance across dozens of business scenarios.
Alibaba’s open‑source Graph‑Learn framework brings industrial‑scale graph neural network capabilities to production by offering lightweight portability, modular extensibility, reusable interfaces, and seamless integration with major deep‑learning ecosystems, while showcasing real‑world security and e‑commerce applications and outlining future hardware and algorithmic directions.
This article surveys recent advances in recommendation system technology, covering the evolution from a two‑stage recall‑ranking pipeline to a four‑stage architecture, and detailing emerging trends in model‑based recall, user‑behavior sequence modeling, knowledge‑graph integration, graph neural networks, advanced ranking models, multi‑objective optimization, multimodal fusion, and listwise re‑ranking.
This article examines the challenges of deepening Graph Convolutional Networks (GCN), introduces ResGCN, DenseGCN, and skip‑neighbor designs to enable deeper architectures, presents experimental results showing improved performance with 28‑layer models, and outlines future research directions.
The article introduces AliGraph, Alibaba’s comprehensive Graph Neural Network platform showcased at NeurIPS 2019, explaining its layered architecture, scalable graph engine, extensible operators, and real‑world applications across e‑commerce, security and cloud services, while highlighting performance gains, supported algorithms, and the strategic focus on GNN research and development.
Professor Song Le presented at the Cloudwise Conference how Ant Financial leverages large‑scale graph neural networks, automated machine‑learning platforms, and multi‑agent reinforcement learning to model complex financial networks, improve risk control, and drive diverse fintech applications.
This article summarizes Alibaba’s twelve KDD 2019 papers, covering advances in long‑sequence CTR modeling, fashion recommendation, sponsored search, exact‑K recommendation, meta‑learning, transfer learning, scalable graph convolution, heterogeneous graph neural networks, knowledge‑driven product description, and related workshops, highlighting both algorithmic innovations and industrial deployments.
The article explains how the heat diffusion equation and the Laplacian operator on graphs provide a physical intuition for Graph Convolutional Networks, showing the equivalence between continuous‑space Fourier analysis and discrete‑space message passing, and linking these concepts to semi‑supervised learning and GraphSAGE implementations.
This article introduces the fundamental architecture of Graph Neural Networks, traces their evolution from early 2005 concepts through spatial and spectral models, explains why GNNs are gaining attention, and poses key questions about graph spectrum, Fourier transforms, and spectral‑domain graph convolutions.
This article explores how cognitive computing combined with graph neural networks and text generation enables large‑scale interest mining, interpretable embeddings, and multi‑modal recommendation in e‑commerce, outlining platform implementations, explainable methods, and future directions for AI‑driven consumer engagement.
In a recent interview, iQiyi’s Vice President Xie Danming predicts that 2019 will mark a golden harvest for artificial intelligence, driven by 5G‑enabled AR/VR expansion, mature voice AI, emerging graph neural networks, and growing application‑focused investment, while highlighting challenges such as model cost, multimodal analysis, and the need for lightweight algorithms.
Euler, Alibaba's newly open‑sourced graph deep‑learning framework, combines distributed graph processing with neural network training to handle billions of nodes and edges, supports heterogeneous graphs, offers built‑in algorithms, and has already boosted advertising, fraud detection, and other industry applications.
This article introduces the GEM method, the first heterogeneous graph neural network designed for malicious account detection, explains the nature and characteristics of malicious accounts, describes why graph neural networks are effective, and presents experimental results from the authors' CIKM 2018 study.
The article presents three Ant Financial research papers featured at KDD 2018—one on graph‑learning fraud detection for return‑freight insurance, another introducing the adaptive GeniePath graph neural network, and a third describing a distributed collaborative hashing system for large‑scale recommendation—highlighting their methodologies, experimental results, and practical impact on Ant Financial’s services.
