Tagged articles
Online Advertising
65 articles · Page 1 of 1
The paper introduces the Trajectory Auto‑Regressive (TAR) model, which uses multi‑scale trajectory generation, a VQ‑VAE latent compression, and a state‑action fusion architecture to address granularity mismatch between fine‑grained decision steps and coarse‑grained feedback in online advertising, achieving over 12% CTR lift, smoother budget pacing, and faster inference compared to prior baselines.
Budget PacingMulti-Scale GenerationOnline Advertising
0 likes · 18 min read

The DARA framework splits online advertising budget allocation into a few‑shot LLM reasoning stage and a fine‑grained optimizer stage, enhanced by a dynamically updated RL‑fine‑tuning algorithm (GRPO‑Adaptive), achieving significantly lower ROI variance than traditional baselines in both real and simulated environments.
Few-Shot LearningLLMOnline Advertising
0 likes · 16 min read

This paper introduces ABPlanner, a few‑shot, context‑aware budget planner that enhances budget‑constrained auto‑bidding in online advertising by hierarchically allocating budgets across short‑term stages and training a sequential decision‑maker with deep reinforcement learning, achieving significant gains in simulated and real‑world A/B tests.
Few-Shot LearningOnline Advertisingauto-bidding
0 likes · 13 min read

This paper introduces ComRecycle, an intelligent computation recycling framework for Taobao's display advertising system that caches and reuses ad candidates across recall, coarse‑ranking, and fine‑ranking stages, achieving up to 23% CPU and 22% GPU savings while maintaining recommendation quality.
Online AdvertisingOnline Convex OptimizationResource Optimization
0 likes · 17 min read

The paper introduces a model‑based reinforcement‑learning auto‑bidding framework that learns a neural‑network environment model from real logs, generates confidence‑aware virtual data fused with real data, and employs the COMBO+MICRO stabilizer and a Lagrange‑dual method for ROI‑constrained bidding, delivering up to 6.8 % higher consumption, 5 % GMV growth and 3.7 % ROI improvement on Alibaba’s platform.
Online Advertisingauto-biddingbudget constrained bidding
0 likes · 22 min read

The paper introduces Contextual Generative Auction (CGA), a generative framework that directly optimizes ad placements while modeling permutation‑level externalities, decouples allocation from payment learning, and achieves near‑optimal Myerson‑style outcomes, delivering up to 3.2% higher RPM, 1.4% more CTR, 6.4% GMV growth, and 3.5% increased advertiser ROI in large‑scale Taobao experiments.
ExternalitiesGenerative ModelsOnline Advertising
0 likes · 18 min read

AuctionNet is a newly introduced benchmark that recreates a massive, realistic online advertising auction environment using latent diffusion‑generated traffic data, provides an 80 GB dataset of 5 × 10⁸ logs from 48 bidding agents, and offers baseline evaluations—including an Online LP that outperforms others—supporting thousands of fair NeurIPS 2024 competition submissions and open‑source tools for large‑scale game decision‑making research.
Generative ModelsOnline Advertisingauto-bidding
0 likes · 15 min read


NewBeeNLPNov 14, 2024 · Artificial Intelligence
The 30th SIGKDD conference in Barcelona featured 2,046 research papers with a 20% acceptance rate, and this article compiles the 59 recommendation‑system papers—covering large‑model recommenders, graph‑based methods, sequential models, fairness, privacy, advertising, debiasing, reinforcement learning and more—for researchers to explore the latest academic advances.
Graph Neural NetworksKDD2024Online Advertising
0 likes · 15 min read

The paper presents AIGB, a generative auto‑bidding framework that replaces reinforcement‑learning with a conditional diffusion model to generate optimal bidding trajectories, and demonstrates through offline benchmarks and Alibaba’s online A/B tests that it consistently outperforms RL baselines, boosting buy count, GMV, and ROI while maintaining low latency.
Generative ModelsMarketing AIOnline Advertising
0 likes · 18 min read

Meituan’s five long papers accepted at KDD 2024 introduce a dual‑intent model for search‑recommendation, a joint auction mechanism for ads, a robust ATE estimator for heavy‑tailed metrics, a decision‑focused causal learning framework for marketing, and an efficient on‑demand order‑pooling system for real‑time courier assignments.
Controlled ExperimentsGraph Neural NetworksKDD 2024
0 likes · 12 min read

The paper introduces LS‑IMP, a two‑stage local‑search algorithm with four novel operators that efficiently solves guaranteed‑delivery advertising inventory allocation under non‑convex multilinear media‑preference constraints, consistently outperforming commercial solvers and heuristics in solution quality and speed on real‑world datasets.
AlgorithmOnline Advertisinginventory allocation
0 likes · 17 min read

Advances in automated bidding for online ads have progressed from classic control and linear programming to reinforcement‑learning pipelines, offline and sustainable online RL, and finally generative‑model approaches, each enhancing decision strength, adaptability, and fairness while addressing simulation gaps, multi‑objective constraints, and real‑time efficiency.
Auction DesignGenerative AIOnline Advertising
0 likes · 25 min read

The article explains the architecture of an online advertising system, detailing the roles of ad placement and playback, the design of coarse ranking modules, budget pacing mechanisms, and pricing algorithms that aim to maximize ROI while controlling costs.
AdvertisingBackendOnline Advertising
0 likes · 14 min read

The paper presents incentive‑compatible, individually rational auction mechanisms for automated ad bidding where advertisers report private budget and ROI constraints, characterizes feasible allocation and payment rules via monotone budget functions, introduces a personalized ranking‑score auction using a “key ROI,” and demonstrates through experiments that the design achieves near‑optimal welfare and revenue while ensuring truthful reporting.
Online AdvertisingROIauction theory
0 likes · 17 min read

This article explains how the Alibaba‑Mama team addresses constrained ROI bidding in a black‑box adversarial environment by introducing a Minimax Regret Optimization framework that aligns training and test distributions, builds a causal world model, and demonstrates robust performance on synthetic and real‑world ad auctions.
Online Advertisingadversarial biddingconstrained optimization
0 likes · 14 min read

The paper introduces RLTP, a reinforcement‑learning‑based pacing system that models delayed‑impression preloaded ads as an MDP, uses a dueling DQN to select traffic probabilities, and simultaneously meets exposure targets, ensures smooth delivery, and maximizes CTR, outperforming rule‑based and PID baselines while removing complex multi‑stage pipelines.
Online AdvertisingRLTPad pacing
0 likes · 16 min read

PerBid introduces a personalized automated bidding framework that creates context‑aware RL agents for advertiser clusters using a profiling network to embed static and dynamic campaign features, and experiments on Alibaba’s display‑ad platform show up to 10.85% performance gains while markedly improving fairness across heterogeneous advertisers.
Online AdvertisingPersonalizationautomated bidding
0 likes · 23 min read

This paper proposes a two‑stage Potential Generalized Second Price auction for augmented ads, ranking guide ads by expected welfare from their linked second‑step ads, shifting billing to the second click to eliminate free‑riding, and demonstrates via offline and online experiments on Taobao that it boosts click‑through, revenue, and GMV while lowering CPC.
Online Advertisinge-commercemachine learning
0 likes · 16 min read

The paper introduces two truthful auction mechanisms—MPU for public‑type and MPR for private‑type bidders—that combine VCG payments for utility‑maximizers with GSP payments for value‑maximizers, achieving incentive compatibility, individual rationality, robustness, and a social‑welfare approximation of up to 2 (optimal within a 1.25 factor) in mixed online advertising markets.
GSPOnline AdvertisingVCG
0 likes · 20 min read

AI‑Generated Bidding (AIGB) replaces reinforcement‑learning with a conditional generative model that learns the joint distribution of bids, objectives and constraints from historical trajectories, enabling interpretable, diverse, constraint‑aware bidding strategies that improve efficiency, scalability and explainability for large‑scale advertising platforms.
Generative AIOnline Advertisingautomated bidding
0 likes · 15 min read

The article surveys advertising auction mechanisms, explaining game‑theoretic foundations, Myerson’s lemma, welfare‑maximizing designs such as VCG and GSP, revenue‑focused extensions with reserve prices (mGSP, aGSP, rGSP, sGSP), and outlines future research on auto‑bidding, machine‑learning optimization, and generative‑AI impacts.
Online Advertisingauction theorygame theory
0 likes · 38 min read

The Hierarchically Constrained Adaptive Ad Exposure (HCA2E) framework treats each user request as a knapsack item and uses a hierarchical greedy‑plus‑beam‑search optimization with a preservation‑order strategy to jointly maximize platform revenue and user experience while respecting global and per‑request ad‑placement constraints, achieving near‑optimal performance and stable, scalable results in extensive offline and online feed‑advertising experiments.
Online Advertisingdynamic ad placementfeed recommendation
0 likes · 17 min read

The Sustainable Online Reinforcement Learning (SORL) framework tackles offline inconsistency in auto‑bidding by iteratively gathering safe online data from real ad systems with a Lipschitz‑based exploration method and training a variance‑suppressed conservative Q‑learning policy, achieving safer, more stable, and higher‑performing bids on Alibaba’s platform.
Online AdvertisingSafe Explorationauto-bidding
0 likes · 18 min read

Two Tencent IEG Growth Platform papers accepted at WSDM 2023 and AAAI 2023 introduce a feature‑missing‑aware routing‑fusion network (MarfNet) and a cross‑domain adaptive framework (CDAF) that significantly improve online advertising LTV prediction despite sparse features and labels.
AICross‑domain AdaptationLTV prediction
0 likes · 4 min read

Xiaohongshu’s commercial team modeled the real‑time bidding process for its “Fries” ad product, derived an optimal linear‑programming bid formula, and implemented a simple two‑parameter PID‑controlled scheme that meets client pacing, delivery guarantees, and platform profit goals while using practical heuristics.
Advertising OptimizationOnline Advertisingalgorithmic strategy
0 likes · 12 min read

Alibaba’s research team secured three NeurIPS 2022 papers—introducing an Adaptive Parameter Generation network that boosts click‑through rates and revenue, a tuning‑free Global Batch Gradient Aggregation method that speeds recommendation model training by 2.4×, and a Sustainable Online Reinforcement Learning framework that outperforms existing auto‑bidding strategies.
NeurIPSOnline Advertisinggradient aggregation
0 likes · 6 min read

This talk by JD Retail's Zhao Yu explains how reinforcement learning is modeled and deployed for large‑scale traffic control during major sales events, detailing system architecture, reward design, offline simulation, model upgrades, and future research directions.
JD.comOnline AdvertisingRL modeling
0 likes · 20 min read

The paper proposes an actor‑critic reinforcement‑learning model (ACRL) that leverages PPO and a deep structured semantic model to optimize real‑time bidding strategies for mobile game ads under CPM and budget constraints, addressing long user lifecycles and sparse conversion data while demonstrably improving ROI in both offline simulations and online A/B tests.
Mobile AdvertisingOnline AdvertisingROI
0 likes · 16 min read

The paper presents CONFLUX, a request-level fusion ranking framework that uses linear programming and cascade distillation to allocate ad impressions between contract and real-time bidding ads, improving platform revenue and ad effectiveness while addressing offline training, latency, and model drift challenges.
CONFLUXKDD 2022Linear Programming
0 likes · 14 min read

AdaCalib is a posterior‑guided feature‑adaptive calibration model for online advertising that learns per‑feature piecewise‑linear calibration functions via a deep neural network with adaptive bucketing, improving probability estimates and ranking, achieving lower Field‑RCE, higher AUC, and a 5% CVR lift in live tests.
CalibrationOnline Advertisingdeep neural networks
0 likes · 19 min read

Alibaba Mama boosts external-media advertising ROI by deploying AI-driven models—privacy-preserving federated learning, hierarchical representation integration, uncertainty-regularized knowledge distillation, and calibrated DNNs—to overcome missing user-preference data, sparse post-click conversions, sample-selection bias, and probability-calibration challenges.
AIFederated LearningOnline Advertising
0 likes · 13 min read

This article analyzes Meituan's delivery advertising system, detailing the shift from linear programming to an evolutionary‑strategy‑based multi‑action allocation (ES‑MACA), describing problem formalization, offline training, reward evaluation, online decision flow, extensive offline and online experiments, and future directions toward reinforcement learning.
AdvertisingMeituanOnline Advertising
0 likes · 28 min read

The article explains the drawbacks of the traditional generalized second‑price auction for online ads, proposes monitoring and probabilistically throttling fast‑spending campaigns using a time‑window based budget pacing model with a Pass‑Through Rate (PTR), and references the related research paper while also offering free resource links.
AIBudget PacingOnline Advertising
0 likes · 3 min read

The paper proposes a novel two‑stage auction for online ads that jointly optimizes a learned pre‑auction score with a second‑stage GSP auction, preserving incentive compatibility and achieving higher social welfare and revenue than traditional greedy methods, validated on public and industrial datasets.
Online Advertisingincentive compatibilitypre-auction scoring
0 likes · 25 min read

Alibaba’s research team had five papers accepted at the prestigious SIGIR 2022 conference in Madrid, covering innovations such as joint ad‑ranking and creative selection, personalized bundle generation, calibrated neural predictions, disentangled counterfactual regression, and cold‑start user recommendation, showcasing strong expertise in information retrieval and online advertising.
CalibrationOnline AdvertisingSIGIR 2022
0 likes · 8 min read

The paper introduces MAAB, a scalable multi‑agent reinforcement‑learning framework for online ad bidding that uses temperature‑regularized credit assignment, adaptive threshold agents, and mean‑field clustering to balance individual advertiser utility, platform revenue, and overall social welfare in competitive auction environments.
Multi-Agent Reinforcement LearningOnline Advertisingauto-bidding
0 likes · 28 min read

This article presents a comprehensive overview of using reinforcement learning to solve advertising bidding ROI optimization, covering historical foundations, methodological reasoning, system architecture, practical implementation details, challenges, evaluation metrics, and recommended algorithms for real‑world ad placement scenarios.
AdvertisingOnline AdvertisingROI optimization
0 likes · 17 min read


58 TechDec 16, 2021 · Artificial Intelligence
This talk presents the design and implementation of 58's commercial recruitment recommendation system, covering the business scenario, system architecture, regional and behavior‑based recall methods, various ranking models—including coarse‑ranking, dual‑tower, DIN‑bias, and multitask W3DA—and future optimization directions.
DBSCANEGESOnline Advertising
0 likes · 20 min read

At WSDM 2022, Alibaba’s advertising team presented four papers introducing a meta‑learning multi‑task multi‑scenario model for advertiser forecasting, a low‑cost Feature Co‑Action Network that boosts CTR prediction, an Adaptive Unified Allocation Framework that improves guaranteed display fulfillment and CTR, and a cooperative‑competitive multi‑agent auto‑bidding system that enhances both advertiser welfare and platform profit.
Meta-learningOnline Advertisingmulti-agent
0 likes · 11 min read

The paper presents MACG, a multi‑agent cooperative bidding game that integrates a global objective with individual advertiser goals, derives optimal bidding formulas, employs a strategy network and evolutionary search to tune parameters, and demonstrates over‑5% metric gains and stable 15‑day performance in Taobao’s online advertising platform.
Multi-Agent Reinforcement LearningMulti-Objective OptimizationOnline Advertising
0 likes · 18 min read

The paper presents Deep Uncertainty‑Aware Learning (DUAL), a scalable Bayesian deep‑learning framework that combines a neural feature extractor with a Gaussian‑process prior to model CTR prediction uncertainty, mitigates feedback‑loop bias, and enables confidence‑driven exploration (UCB and Thompson sampling) that improves long‑term utility while preserving accuracy.
Gaussian ProcessOnline AdvertisingUncertainty Modeling
0 likes · 15 min read

Alibaba Mama’s talk explains how calibrated probability models—evolving from simple Platt scaling to Bayesian isotonic regression and real‑time wave‑adjusted variants—improve click‑through and conversion predictions, enabling more accurate bidding, stable auctions, and fairer ad allocation despite data drift and sparsity.
AlgorithmCalibrationOnline Advertising
0 likes · 20 min read

This article introduces the concept of calibration in trustworthy machine learning, explains why accurate probability estimates are crucial for online advertising, reviews related research and evaluation metrics, and details the evolution of calibration algorithms such as Smoothed Isotonic Regression, Bayes‑SIR, real‑time optimizations, and post‑click conversion models, concluding with engineering deployment and future directions.
CalibrationOnline AdvertisingTrustworthy AI
0 likes · 18 min read

This article describes how Alibaba's advertising team tackled the challenges of modeling long‑term user interests for CTR prediction by co‑designing incremental computation services, introducing memory‑network‑based models (MIMN and HPMN), and achieving significant offline and online performance gains.
CTR predictionLong-Term InterestOnline Advertising
0 likes · 17 min read

Alibaba Mama combats the roughly 8.6 % abnormal traffic in China’s online ad market by distinguishing low‑quality from cheating clicks, employing a proactive perception layer, high‑dimensional visual analytics, and a dual‑stage real‑time and batch filtering system that also freezes fraudulent affiliate commissions and is continuously evaluated with precision‑recall and AUC metrics.
AlibabaOnline AdvertisingRisk Control
0 likes · 15 min read

An ecosystem‑wide intelligent promotion system applies a budget‑pacing algorithm with probabilistic throttling and fine‑ranking score adjustments in 5‑minute slots, guaranteeing uniform video exposure while minimizing impact on overall consumption, boosting daily exposure completion from under 5 % to up to 70 % and reducing watch‑time loss.
AlgorithmBudget PacingContent Distribution
0 likes · 10 min read

The article explains the exploration‑exploitation dilemma in recommendation systems, introduces the GuideBoot algorithm—an innovative guided bootstrap approach for contextual bandits—describes its Bayesian and non‑Bayesian foundations, presents experimental results on synthetic and real advertising data, and discusses an online learning extension.
Exploration-ExploitationGuideBootOnline Advertising
0 likes · 11 min read

The article presents Alibaba's display‑advertising team’s three‑stage computing‑power efficiency evolution, introduces the DCAF personalized power‑allocation algorithm with its Lagrangian formulation, and describes the AllSpark dynamic‑control framework that together enable a flexible, resource‑aware Transformers engine achieving significant business gains during high‑traffic events.
Online AdvertisingSystem Optimizationalgorithmic co-design
0 likes · 21 min read

This article describes how 58.com applied deep‑learning techniques—including feature engineering, sample construction, model evolution from Wide&Deep to DIN/DIEN and multi‑task learning—and system‑level optimizations to improve CTR/CPM performance in its large‑scale commercial ranking platform.
CTR predictionOnline AdvertisingSystem Optimization
0 likes · 38 min read

The article introduces COLD, a computing‑power‑aware online and lightweight deep pre‑ranking system for Alibaba's targeted ads, detailing its evolution from static CTR models to vector‑inner‑product models, its flexible network architecture with feature‑selection via SE blocks, engineering optimizations such as parallelism, column‑wise computation, Float16 and MPS, and demonstrates superior offline and online performance through extensive experiments.
COLDOnline Advertisingfeature selection
0 likes · 11 min read


58 TechJul 8, 2020 · Artificial Intelligence
This article introduces mainstream budget‑pacing methods for cost‑per‑click online ads, describes the 58.com business scenarios, details the pacing algorithm—including bid modification, probabilistic throttling, and reinforcement‑learning approaches—explains system design with PID control, and presents online experimental results and future directions.
Ad TechBudget PacingOnline Advertising
0 likes · 14 min read

This article explains the concepts of dynamic advertising, programmatic creative and dynamic product ads, target conversion bidding (OCPM/OCPX), the role and functions of Data Management Platforms (DMP), the fundamentals of ad auction mechanisms, and the design of optimal auction mechanisms for maximizing platform revenue.
DMPOnline Advertisingauction theory
0 likes · 24 min read

This article introduces the evolution, basic concepts, pricing models, and testing considerations of online advertising, covering ad formats, bidding mechanisms, key industry terms, and practical SDK testing points for mobile and video ads.
Ad TechOnline Advertisingad pricing models
0 likes · 8 min read

This article introduces the fundamentals of online advertising, covering its definition, evolution from contract to real‑time bidding models, key components such as DSP, ADX, ADN, RTB, various billing methods, and practical testing considerations for mobile ad SDKs.
Ad TechOnline Advertisingad testing
0 likes · 7 min read

This article presents a comprehensive overview of the evolution of Weibo's advertising strategy engineering architecture, detailing the system's growth from early banner ads to a sophisticated, multi‑layered online advertising platform that integrates algorithmic models, A/B experimentation, real‑time data pipelines, and precision targeting to support scalable, high‑performance ad delivery.
A/B testingAdvertisingOnline Advertising
0 likes · 19 min read

This article presents detailed study notes on the book “Computational Advertising”, covering an overview, ad system architecture, targeted advertising, online advertising, the E‑E algorithm, and additional insights, accompanied by illustrative diagrams to aid understanding of modern advertising technologies.
E&E algorithmOnline Advertisingad systems
0 likes · 3 min read

This talk by Sogou search advertising researcher Shupeng explains how deep learning techniques are applied to search ad tasks such as automated creative generation and click‑through‑rate prediction, covering system workflow, data pipelines, model evolution from linear models to Wide&Deep and NFM, evaluation metrics, and future directions.
CTR estimationOnline Advertisingautomated creative
0 likes · 33 min read

This article presents a comprehensive overview of 58.com’s search efficiency optimization, detailing the business background, ranking framework, data, algorithm, and engineering components, describing the three-stage ranking process, strategy and platform optimizations, feature engineering, model upgrades, and the resulting performance improvements.
AlgorithmOnline Advertisingefficiency optimization
0 likes · 12 min read


58 TechNov 9, 2018 · Artificial Intelligence
This article details how 58.com improved the efficiency of its search list ranking by moving from simple time‑based ordering to a comprehensive ranking framework that incorporates feedback strategies, basic machine‑learning models, feature upgrades, and advanced model upgrades, achieving significant gains in click‑through, conversion, and revenue across multiple business lines.
Online Advertisingclick-through ratefeature engineering
0 likes · 23 min read

This article summarizes Zeng Xianglin’s presentation on the stages of an algorithm engineer’s career—from academic Beta research and feature engineering through online deployment, model training, and deep‑learning applications—highlighting practical challenges and best practices in large‑scale advertising systems.
Algorithm EngineeringOnline Advertisingbig data
0 likes · 13 min read

This article explains how reinforcement learning, specifically the DDPG algorithm combined with Transformer-based networks, is applied to improve large‑scale news recall systems, detailing the business scenario, algorithm selection, model architecture, speed optimizations, training challenges, and observed online performance gains.
AIDDPGOnline Advertising
0 likes · 13 min read

Rocket Launching introduces a novel co‑training framework that jointly trains a lightweight network and a more powerful booster network, sharing parameters and using gradient‑blocking and hint loss to improve click‑through‑rate prediction accuracy while keeping online inference latency unchanged, validated on public datasets and Alibaba’s ad system.
CTR predictionOnline Advertisingco-training
0 likes · 13 min read

Online click‑through‑rate (CTR) prediction demands millisecond‑level response times, yet deep models are too slow; this paper introduces a “Rocket Launching” framework that jointly trains a lightweight net and a powerful booster net, sharing parameters and using gradient‑blocking and hint loss to improve accuracy without increasing inference latency.
CTR predictionOnline Advertisingco-training
0 likes · 13 min read

The paper “Optimized Cost per Click in Taobao Display Advertising” introduces a novel two‑level OCPC smart bidding algorithm that jointly optimizes advertiser ROI, user experience, and platform revenue, presents detailed mathematical formulations, offline and online experiments showing significant gains in GMV, CTR, CVR, and RPM across single‑item and banner ad placements.
OCPCOnline AdvertisingROI optimization
0 likes · 11 min read

This interview with Alibaba researcher Xu Yinghui reveals how the company built large‑scale deep reinforcement learning systems for search, recommendation, logistics and online advertising, detailing team structures, technical breakthroughs, training challenges, and future directions such as multi‑agent learning and GAN integration.
AIAlibabaOnline Advertising
0 likes · 20 min read
