Awesome-Deep-Learning-Papers-for-Search-Recommendation-Advertising

Awesome Deep Learning papers for industrial Search, Recommendation and Advertisement. They focus on Embedding, Matching, Pre-Ranking, Ranking (CTR/CVR prediction), Post-Ranking, Relevance-Ranking, LLM based ranking, Reinforcement Learning and so on.

Please cite our paper: Deep Learning to Rank in Industrial Search Engines, Recommender Systems, and Online Advertising - An Overview and New Perspectives, which was published in TOIS 2026.

00_Overview

• 2026 (TOIS) [DLTR] Deep Learning to Rank in Industrial Search Engines, Recommender Systems, and Online Advertising - An Overview and New Perspectives

01_Embedding

• 2013 (Google) (NIPS) [Word2vec] Distributed Representations of Words and Phrases and their Compositionality

• 2014 (KDD) [DeepWalk] DeepWalk - online learning of social representations

• 2015 (WWW) [LINE] LINE Large-scale Information Network Embedding

• 2016 (KDD) [Node2vec] node2vec - Scalable Feature Learning for Networks

• 2017 (ICLR) [GCN] Semi-supervised Classification with Graph Convolutional Networks

• 2017 (KDD) [Struc2vec] struc2vec - Learning Node Representations from Structural Identity

• 2017 (NIPS) [GraphSAGE] Inductive Representation Learning on Large Graphs

• 2018 (Alibaba) (KDD) *[Alibaba Embedding] Billion-scale Commodity Embedding for E-commerce Recommendation in Alibaba

• 2018 (ICLR) [GAT] Graph Attention Networks

• 2018 (Pinterest) (KDD) *[PinSage] Graph Convolutional Neural Networks for Web-Scale Recommender Systems

• 2018 (WSDM) [NetMF] Network embedding as matrix factorization - Unifying deepwalk, line, pte, and node2vec

• 2019 (Alibaba) (KDD) *[GATNE] Representation Learning for Attributed Multiplex Heterogeneous Network

02_Matching

• 1994 (CSCW) [User-CF] GroupLens - An Open Architecture for Collaborative Filtering of Netnews

• 1998 (Microsoft) Empirical Analysis of Predictive Algorithms for Collaborative Filtering

• 2003 (Amazon) [Item-CF] Amazon.com recommendations - item-to-item collaborative filtering

• 2009 (Computer) [MF] Matrix factorization techniques for recommender systems

• 2013 (Microsoft) (CIKM) [DSSM] Learning deep structured semantic models for web search using clickthrough data

• 2015 (KDD) [Sceptre] Inferring Networks of Substitutable and Complementary Products

• 2016 (Google) (RecSys) **[Youtube DNN] Deep Neural Networks for YouTube Recommendations

• 2018 (Airbnb) (KDD) *[Airbnb Embedding] Real-time Personalization using Embeddings for Search Ranking at Airbnb

• 2018 (Alibaba) (KDD) * [TDM] Learning Tree-based Deep Model for Recommender Systems

• 2018 (Pinterest) (KDD) *[PinSage] Graph Convolutional Neural Networks for Web-Scale Recommender Systems

• 2019 (Alibaba) (CIKM) **[MIND] Multi-Interest Network with Dynamic Routing for Recommendation at Tmall

• 2019 (Alibaba) (CIKM) *[SDM] SDM - Sequential deep matching model for online large-scale recommender system

• 2019 (Alibaba) (NIPS) *[JTM] Joint Optimization of Tree-based Index and Deep Model for Recommender Systems

• 2019 (Amazon) (KDD) Semantic Product Search

• 2019 (Baidu) (KDD) *[MOBIUS] MOBIUS - Towards the Next Generation of Query-Ad Matching in Baidu's Sponsored Search

• 2019 (Google) (RecSys) **[Two-Tower] Sampling-Bias-Corrected Neural Modeling for Large Corpus Item Recommendations

• 2019 (Google) (WSDM) *[Top-K Off-Policy] Top-K Off-Policy Correction for a REINFORCE Recommender System

• 2019 [Tencent] (KDD) A User-Centered Concept Mining System for Query and Document Understanding at Tencent

• 2020 (Alibaba) (Arxiv) [SWING] Large Scale Product Graph Construction for Recommendation in E-commerce

• 2020 (Alibaba) (ICML) [OTM] Learning Optimal Tree Models under Beam Search

• 2020 (Alibaba) (KDD) *[ComiRec] Controllable Multi-Interest Framework for Recommendation

• 2020 (Facebook) (KDD) **[Embedding for Facebook Search] Embedding-based Retrieval in Facebook Search

• 2020 (Google) (WWW) *[MNS] Mixed Negative Sampling for Learning Two-tower Neural Networks in Recommendations

• 2020 (JD) (CIKM) *[DecGCN] Decoupled Graph Convolution Network for Inferring Substitutable and Complementary Items

• 2020 (JD) (SIGIR) [DPSR] Towards Personalized and Semantic Retrieval - An End-to-EndSolution for E-commerce Search via Embedding Learning

• 2020 (Microsoft) (Arxiv) TwinBERT - Distilling Knowledge to Twin-Structured BERT Models for Efficient Retrieval

• 2021 (Alibaba) (KDD) * [MGDSPR] Embedding-based Product Retrieval in Taobao Search

• 2021 (Alibaba) (SIGIR) * [PDN] Path-based Deep Network for Candidate Item Matching in Recommenders

• 2021 (Amazon) (KDD) Extreme Multi-label Learning for Semantic Matching in Product Search

• 2021 (Baidu) (KDD) Pre-trained Language Model for Web-scale Retrieval in Baidu Search

• 2021 (Bytedance) (Arxiv) [DR] Deep Retrieval - Learning A Retrievable Structure for Large-Scale Recommendations

• 2021 (Meituan) (DLP-KDD) [DAT]A Dual Augmented Two-tower Model for Online Large-scale Recommendation

• 2022 (Alibaba) (CIKM) **[NANN] Approximate Nearest Neighbor Search under Neural Similarity Metric for Large-Scale Recommendation

• 2022 (Alibaba) (CIKM) [CLE-QR] Query Rewriting in TaoBao Search

• 2022 (Alibaba) **(CIKM) [MOPPR] Multi-Objective Personalized Product Retrieval in Taobao Search

• 2024 (Bytedance) (KDD) [Trinity] Trinity - Syncretizing Multi-:Long-Tail:Long-Term Interests All in One

• 2024 (Meta) (Arxiv) ** [GR] Actions Speak Louder than Words - Trillion-Parameter Sequential Transducers for Generative Recommendations

• 2025 (Bytedance) (Arxiv) [LongRetriever] LongRetriever - Towards Ultra-Long Sequence based Candidate Retrieval for Recommendation

• 2025 (Bytedance) (KDD) [VQ] Real-time Indexing for Large-scale Recommendation by Streaming Vector Quantization Retriever

• 2025 (JD) (KDD) [UniERF] UniERF - A Uniform Embedding-based Retrieval Framework for E-Commerce Search

• 2025 (Meta) (KDD) [MTMH] Optimizing Recall or Relevance? A Multi-Task Multi-Head Approach for Item-to-Item Retrieval in Recommendation

• 2025 (Meta) [RADAR ] RADAR - Recall Augmentation through Deferred Asynchronous Retrieval

• 2025 (Tencent) (Arxiv) An Efficient Embedding Based Ad Retrieval with GPU-Powered Feature Interaction

ANN

• 2017 (Arxiv) (Meta) [FAISS] Billion-scale similarity search with GPUs

• 2020 (PAMI) [HNSW] Efficient and Robust Approximate Nearest Neighbor Search Using Hierarchical Navigable Small World Graphs

• 2021 (TPAMI) [IVF-PQ] Product Quantization for Nearest Neighbor Search

Graph_Neural_Networks

• 2017 (ICLR) [GCN] Semi-Supervised Classification with Graph Convolutional Networks

• 2018 (ICLR) [GAT] Graph Attention Networks

• 2018 (Pinterest) (KDD) [PinSage] Graph Convolutional Neural Networks for Web-Scale Recommender Systems

• 2019 (Alibaba) (KDD) [IntentGC] IntentGC - a Scalable Graph Convolution Framework Fusing Heterogeneous Information for Recommendation

• 2019 (Alibaba) (KDD) [MEIRec] Metapath-guided Heterogeneous Graph Neural Network for Intent Recommendation

• 2019 (Alibaba) (SIGIR) [GIN] Graph Intention Network for Click-through Rate Prediction in Sponsored Search

• 2020 (Alibaba) (SIGIR) [ATBRG] ATBRG - Adaptive Target-Behavior Relational Graph Network for Effective Recommendation

LLM_Matching

• 2021 (Baidu) (KDD) Pre-trained Language Model for Web-scale Retrieval in Baidu Search

• 2023 (Google) (NIPS) [TIGER] Recommender Systems with Generative Retrieval

• 2024 (Alibaba) (WWW) [BEQUE] Large Language Model based Long-tail Query Rewriting in Taobao Search

• 2024 (Kuaishou) (Arxiv) [KuaiFormer] KuaiFormer - Transformer-Based Retrieval at Kuaishou

• 2024 (Meta) (Arxiv) ** [GR] Actions Speak Louder than Words - Trillion-Parameter Sequential Transducers for Generative Recommendations

• 2024 (Meta) (Arxiv) Unifying Generative and Dense Retrieval for Sequential Recommendation

• 2024 (Xiaohongshu) (WWW) [NoteLLM] NoteLLM - Multimodal Large Representation Models for Recommendation

• 2025 (Kuaishou) (Arxiv)[OneRec] OneRec - Unifying Retrieve and Rank with Generative Recommender and Preference Alignment

• 2025 (Alibaba) (Arxiv) [TBGRecall] TBGRecall - A Generative Retrieval Model for E-commerce Recommendation Scenarios

• 2025 (Arxiv) (Tencent) [RARE] Real-time Ad retrieval via LLM-generative Commercial Intention for Sponsored Search Advertising

• 2025 (Baidu) (Arxiv) [COBRA] Sparse Meets Dense -Unified Generative Recommendations with Cascaded Sparse-Dense Representations

• 2025 (Google) [PLUM] PLUM - Adapting Pre-trained Language Models for Industrial-scale Generative Recommendations

• 2025 (JD) (Arxiv) [GRAM] Generative Retrieval and Alignment Model - A New Paradigm for E-commerce Retrieval

• 2025 (Kuaishou) (AAAI) [Align3GR] Align3GR - Unified Multi-Level Alignment for LLM-based Generative Recommendation

• 2025 (Kuaishou) (Arxiv) [LARM] LLM-Alignment Live-Streaming Recommendationpdf

• 2025 (Kuaishou) (Arxiv) [LEARN] LEARN - Knowledge Adaptation from Large Language Model to Recommendation for Practical Industrial Application

• 2025 (Meta) (Arxiv) [DRAMA] DRAMA - Diverse Augmentation from Large Language Models to Smaller Dense Retrievers

• 2025 (Meta) (Arxiv) [ROO] Request-Only Optimization for Recommendation Systems

• 2025 (Pinterest) [PinRec] PinRec - Outcome-Conditioned, Multi-Token Generative Retrieval for Industry-Scale Recommendation Systems

• 2025 (Xiaohongshu) （KDD) [NoteLLM-2] NoteLLM-2 - Multimodal Large Representation Models for Recommendation

03_Pre-ranking

• 2020 (Alibaba) (DLP-KDD) [COLD] COLD - Towards the Next Generation of Pre-Ranking System

• 2022 (Huawei) (SIGIR) [RankFlow] RankFlow - JointOptimization ofMulti-Stage CascadeRanking SystemsasFlows

• 2022 （Huawei) (CIKM) [IntTower] IntTower - the Next Generation of Two-Tower Model for Pre-Ranking System

• 2023 (Alibaba) (CIKM) [COPR] COPR - Consistency-Oriented Pre-Ranking for Online Advertising

• 2023 (Alibaba) (KDD) [ASMOL] Rethinking the Role of Pre-ranking in Large-scale E-Commerce Searching System

• 2025 (Tencent) (Arxiv) [HIT] HIT Model - A Hierarchical Interaction-Enhanced Two-Tower Model for Pre-Ranking Systems

04_Ranking

• 2014 (ADKDD) (Facebook) Practical Lessons from Predicting Clicks on Ads at Facebook

• 2016 (Google) (DLRS) **[Wide & Deep] Wide & Deep Learning for Recommender Systems

• 2016 (Google) (RecSys) **[Youtube DNN] Deep Neural Networks for YouTube Recommendations

• 2018 (Alibaba) (KDD) **[DIN] Deep Interest Network for Click-Through Rate Prediction

• 2019 (Alibaba) (AAAI) **[DIEN] Deep Interest Evolution Network for Click-Through Rate Prediction

• 2019 (Alibaba) (KDD) [MIMN] Practice on Long Sequential User Behavior Modeling for Click-Through Rate Prediction

• 2019 (CIKM) ** [AutoInt] AutoInt -Automatic Feature Interaction Learning via Self-Attentive Neural Networks

• 2019 (Facebook) (Arxiv) [DLRM] (Facebook) Deep Learning Recommendation Model for Personalization and Recommendation Systems, Facebook

• 2019 (Google) (Recsys) ** [Youtube Multi-task] Recommending what video to watch next - a multitask ranking system

• 2019 (Google) (WWW) Towards Neural Mixture Recommender for Long Range Dependent User Sequences

• 2020 (Alibaba) (Arxiv) ** [SIM] Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Prediction

• 2020 (Alibaba) (NIPS) Neuron-level Structured Pruning using Polarization Regularizer

• 2020 (ICLR) Reformer - The Efficient Transformer

• 2020 (JD) (CIKM) **[DMT] Deep Multifaceted Transformers for Multi-objective Ranking in Large-Scale E-commerce Recommender Systems

• 2020 (SIGIR) [UBR4CTR] User Behavior Retrieval for Click-Through Rate Prediction

• 2020 (Tencent) (Recsys) ** [PLE] Progressive Layered Extraction (PLE) - A Novel Multi-Task Learning (MTL) Model for Personalized Recommendations

• 2021 (Alibaba) (Arxiv) [ETA] End-to-End User Behavior Retrieval in Click-Through Rate Prediction Model

• 2021 (Alibaba) (CIKM) * [ZEUS] Self-Supervised Learning on Users’ Spontaneous Behaviors for Multi-Scenario Ranking in E-commerce

• 2021 (Alibaba) (CIKM) [STAR] One Model to Serve All - Star Topology Adaptive Recommender for Multi-Domain CTR Prediction

• 2021 (Google) (WWW) * [DCN V2] DCN V2 - Improved Deep & Cross Network and Practical Lessons for Web-scale Learning to Rank Systems

• 2022 (Alibaba) (Arxiv) ** [ETA] Efficient Long Sequential User Data Modeling for Click-Through Rate Prediction

• 2022 (Alibaba) (WSDM) Modeling Users’ Contextualized Page-wise Feedback for Click-Through Rate Prediction in E-commerce Search

• 2022 (Meituan) (CIKM) [SDIM] Sampling Is All You Need on Modeling Long-Term User Behaviors for CTR Prediction

• 2022 (Meta) ** (Arxiv) DHEN - A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction

• 2023 (Alibaba) (Arxiv) [ESLM] Entire Space Learning Framework - Unbias Conversion Rate Prediction in Full Stages of Recommender System

• 2023 (Google) (Arxiv) On the Factory Floor - ML Engineering for Industrial-Scale Ads Recommendation Models

• 2023 (Google) ** (Arxiv) [Hiformer] Hiformer - Heterogeneous Feature Interactions Learning with Transformers for Recommender Systems

• 2023 (Kuaishou) (Arixiv) [TWIN] TWIN - TWo-stage Interest Network for Lifelong User Behavior Modeling in CTR Prediction at Kuaishou

• 2023 (Kuaishou) (CIKM) [QIN] Query-dominant User Interest Network for Large-Scale Search Ranking

• 2023 (Kuaishou) (KDD) [PEPNet] PEPNet - Parameter and Embedding Personalized Network for Infusing with Personalized Prior Information

• 2024 (Kuaishou) (CIKM) [TWINv2] TWIN V2 - Scaling Ultra-Long User Behavior Sequence Modeling for Enhanced CTR Prediction at Kuaishou

• 2024 (Meta) (Arxiv) ** [GR] Actions Speak Louder than Words - Trillion-Parameter Sequential Transducers for Generative Recommendations

• 2024 (Meta) ** (PMLR) [Wukong] Wukong - Towards a Scaling Law for Large-Scale Recommendation

• 2024 (Tencent) (KDD) [LCN] Cross-Domain LifeLong Sequential Modeling for Online Click-Through Rate Prediction

• 2025 (Alibaba) (Arxiv) [MUSE] MUSE - A Simple Yet Effective Multimodal Search-Based Framework for Lifelong User Interest Modeling

• 2025 (Bytedance) ** (Arxiv) [LONGER] LONGER - Scaling Up Long Sequence Modeling in Industrial Recommenders

• 2025 (Bytedance) ** (Arxiv) [RankMixer] RankMixer - Scaling Up Ranking Models in Industrial Recommenders

• 2025 (Kuaishou) (KDD) [HiT-LBM] Hierarchical Tree Search-based User Lifelong Behavior Modeling on Large Language Model

• 2025 (Meta) (Arxiv) [VISTA] Massive Memorization with Hundreds of Trillions of Parameters for Sequential Transducer Generative Recommenders

• 2025 (Meta) (KDD) DV365 - Extremely Long User History Modeling at Instagram

• 2025 (Pinterest) (Arxiv) [TransActV2]TransAct V2 - Lifelong User Action Sequence Modeling on Pinterest Recommendation

• 2025 （Bytedance) ** (Arxiv) [STCA] Make It Long, Keep It Fast - End-to-End 10k-Sequence Modeling at Billion Scale on Douyin

• 2026 (Meta) (KDD) [Lattice] Meta Lattice - Model Space Redesign for Cost-Effective Industry-Scale Ads Recommendations

Activation-Function

• 2020(Arxiv) [GLU] GLU Variants Improve Transformer

Calibration

• 2014 (ADKDD) (Facebook) Practical Lessons from Predicting Clicks on Ads at Facebook

• 2014 (TIST) Simple and scalable response prediction for display advertising

• 2023 Classifier Calibration with ROC-Regularized Isotonic Regression

Classic

• 2016 (ICLR) [GRU4Rec] Session-based Recommendations with Recurrent Neural Networks

• 2017 (Amazon) (IEEE) Two decades of recommender systems at Amazon.com

DNN

• 2019 (KDD) (Airbnb) Applying Deep Learning To Airbnb Search

• 2020 (Airbnb) (KDD) Improving Deep Learning For Airbnb Search

Delayed-Feedback-Problem

• 2008 (KDD) Learning Classifiers from Only Positive and Unlabeled Data

• 2014 (Criteo) (KDD) [DFM] Modeling Delayed Feedback in Display Advertising

• 2018 (Arxiv) [NoDeF] A Nonparametric Delayed Feedback Model for Conversion Rate Prediction

• 2019 (Twitter) (RecSys) Addressing Delayed Feedback for Continuous Training with Neural Networks in CTR prediction

• 2020 (AdKDD) Delayed Feedback Model with Negative Binomial Regression for Multiple Conversions

• 2020 (JD) (IJCAI) [TS-DL] An Attention-based Model for Conversion Rate Prediction with Delayed Feedback via Post-click Calibration

• 2020 (SIGIR) [DLA-DF] Dual Learning Algorithm for Delayed Conversions

• 2020 (WWW) [FSIW] A Feedback Shift Correction in Predicting Conversion Rates under Delayed Feedback

• 2021 (Alibaba) (AAAI) [ES-DFM] Capturing Delayed Feedback in Conversion Rate Prediction via Elapsed-Time Sampling

• 2021 (Alibaba) (AAAI) [ESDF] Delayed Feedback Modeling for the Entire Space Conversion Rate Prediction

• 2021 (Alibaba) (Arxiv) [Defer] Real Negatives Matter - Continuous Training with Real Negatives for Delayed Feedback Modeling

• 2021 (Google) (Arxiv) Handling many conversions per click in modeling delayed feedback

• 2021 (Tencent) (SIGIR) Counterfactual Reward Modification for Streaming Recommendation with Delayed Feedback

• 2022 (Alibaba) (WWW) [DEFUSE] Asymptotically Unbiased Estimation for Delayed Feedback Modeling via Label Correction

Distill

• 2020 (Alibaba) (KDD) *[Privileged Features Distillation] Privileged Features Distillation at Taobao Recommendations

• 2024 (Google) Self-Auxiliary Distillation for Sample Efficient Learning in Google-Scale Recommenders

• 2025 (Bytedance) (KDD) [HA-PFD] Hardness-aware Privileged Features Distillation with Latent Alignment for CVR Prediction

• 2025 (Kuaishou) [MIKD] Mutual Information-aware Knowledge Distillation for Short Video Recommendation

• 2026 （Bytedance) (Arxiv) [Rec-Distill] Rec-Distill - An Industrial Distillation Pipeline for Large-Scale Recommendation Models

Experiment

• 2010 (Google) Overlapping Experiment Infrastructure - More, Better, Faster Experimentation

• 2019 (Alibaba) (KDD) OptimizedCost perClickin TaobaoDisplayAdvertising

• 2022 (Google) (KDD) Scale Calibration of Deep Ranking Models

Feature-Crossing

• 2010 (ICDM) [FM] Factorization machines

• 2013 (Google) (KDD) [LR] Ad Click Prediction - a View from the Trenches

• 2016 (Arxiv) [PNN] Product-based Neural Networks for User Response Prediction

• 2016 (Criteo) (Recsys) [FFM] Field-aware Factorization Machines for CTR Prediction

• 2016 (ECIR) [FNN] Deep Learning over Multi-field Categorical Data – A Case Study on User Response Prediction

• 2016 (KDD) [Deepintent] Deepintent - Learning attentions for online advertising with recurrent neural networks

• 2016 (Microsoft) (KDD) [Deep Crossing] Deep Crossing - Web-scale modeling without manually crafted combinatorial features

• 2017 (Google) (ADKDD) [DCN] Deep & CrossNetwork for Ad Click Predictions

• 2017 (Huawei) (IJCAI) [DeepFM] DeepFM - A Factorization-Machine based Neural Network for CTR Prediction

• 2017 (IJCAI) [AFM] Attentional Factorization Machines Learning the Weight of Feature Interactions via Attention Networks

• 2017 (SIGIR) [NFM] Neural Factorization Machines for Sparse Predictive Analytics

• 2017 (WWW) [NCF] Neural Collaborative Filtering

• 2018 (Google) (WSDM) [Latent Cross] Latent Cross Making Use of Context in Recurrent Recommender Systems

• 2018 (KDD) [xDeepFM] xDeepFM - Combining Explicit and Implicit Feature Interactions for Recommender Systems

• 2018 (TOIS) [PNN] Product-Based Neural Networks for User Response Prediction over Multi-Field Categorical Data

• 2019 (CIKM) ** [AutoInt] AutoInt - Automatic Feature Interaction Learning via Self-Attentive Neural Networks

• 2019 (Huawei) (WWW) [FGCNN] Feature Generation by Convolutional Neural Network for Click-Through Rate Prediction

• 2019 (Sina) (Arxiv) [FAT-DeepFFM] FAT-DeepFFM - Field Attentive Deep Field-aware Factorization Machine

• 2019 (Tencent) (AAAI) [IFM] Interaction-aware Factorization Machines for Recommender Systems

• 2020 (Baidu) (KDD) [CAN] Combo-Attention Network for Baidu Video Advertising

• 2021 (Google) (NIPS) [MLP-Mixer] MLP-Mixer - An all-MLP Architecture for Vision

• 2021 (Google) (WWW) * [DCN V2] DCN V2 - Improved Deep & Cross Network and Practical Lessons for Web-scale Learning to Rank Systems

• 2022 (Alibaba) (WSDM) * [CAN] CAN - Feature Co-Action Network for Click-Through Rate Prediction

• 2022 (Meta) ** (Arxiv) DHEN - A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction

• 2023 (CIKM) * [GDCN] Towards Deeper, Lighter and Interpretable Cross Network for CTR Prediction

• 2023 (Google) ** (Arxiv) [Hiformer] Hiformer - Heterogeneous Feature Interactions Learning with Transformers for Recommender Systems

• 2023 (Sina) (CIKM) [MemoNet] MemoNet - Memorizing All Cross Features’ Representations Efficiently via Multi-Hash Codebook Network for CTR Prediction

• 2024 (Google) (ICLR) From Sparse to Soft Mixtures of Experts

• 2024 (Meta) ** (PMLR) [Wukong] Wukong - Towards a Scaling Law for Large-Scale Recommendation

• 2024 （LinkedIn) (KDD) [RDCN] LiRank - Industrial Large Scale Ranking Models at LinkedIn

• 2025 (Alibaba) [HHFT] HHFT - Hierarchical Heterogeneous Feature Transformer for Recommendation Systems

• 2025 (Bytedance) (Arxiv) [RankMixer] RankMixer - Scaling Up Ranking Models in Industrial Recommenders

• 2025 (Bytedance) ** (Arxiv) [Pyramid Mixer] Pyramid Mixer - Multi-dimensional Multi-period Interest Modeling for Sequential Recommendation

• 2025 (Bytedance) ** (Arxiv) [RankMixer] RankMixer - Scaling Up Ranking Models in Industrial Recommenders

• 2025 (Bytedance) ** (CIKM) [RankMixer] RankMixer - Scaling Up Ranking Models in Industrial Recommenders

• 2025 (Google) ** (Arxiv) [Hiformer] Hiformer - Heterogeneous Feature Interactions Learning with Transformers for Recommender Systems

• 2025 (Meta) (CIKM) [InterFormer] InterFormer - Effective Heterogeneous Interaction Learning for Click-Through Rate Prediction

• 2025 (Tencent) (Arxiv) [D-MoE] Enhancing CTR Prediction with De-correlated Expert Networks

• 2025 （Alibaba) [FAT] From Scaling to Structured Expressivity - Rethinking Transformers for CTR Prediction

• 2025 （Bytedance) (Arxiv) [OneTrans] OneTrans - Unified Feature Interaction and Sequence Modeling with One Transformer in Industrial Recommender

• 2026 (Alibaba) (Arxiv) [EST] EST - Towards Efficient Scaling Laws in Click-Through Rate Prediction via Unified Modeling

• 2026 (Alibaba) (Arxiv) [MGDIN] Deferred is Better - A Framework for Multi-Granularity Deferred Interaction of Heterogeneous Features

• 2026 (Bytedance) (Arxiv) [MixFormer] MixFormer - Co-Scaling Up Dense and Sequence in Industrial Recommenders

• 2026 (Bytedance) (Arxiv) [TokenMixer-Large] TokenMixer-Large - Scaling Up Large Ranking Models in Industrial Recommenders

• 2026 (Bytedance) (Arxiv) [UG-Sep] Compute Only Once - UG-Separation for Efficient Large Recommendation Models

• 2026 (Bytedance) (Arxiv) [Zenith] Zenith - Scaling up Ranking Models for Billion-scale Livestreaming Recommendation

• 2026 (Coupon) [Arxiv] On the Practice of Scaling Search Conversion Rate Prediction

• 2026 (Kuaishou) (Arxiv) [UniMixer] UniMixer - A Unified Architecture for Scaling Laws in Recommendation Systems

• 2026 (Tencent) (Arxiv) [RankUp] RankUp - Towards High-rank Representations for Large Scale Advertising Recommender Systems

• 2026 （Bytedance) (Arxiv) [Rec-Distill] Rec-Distill - An Industrial Distillation Pipeline for Large-Scale Recommendation Models

• 2026 （Tencent) (Arxiv) [TokenFormer] TokenFormer - Unify the Multi-Field and Sequential Recommendation Worlds

Feature_Importance

• 2022 (Kuaishou) (Arxiv) [LPFS] LPFS - Learnable Polarizing Feature Selection for Click-Through Rate Prediction

• 2024 (Huawei)(KDD) ERASE - Benchmarking Feature Selection Methods for Deep Recommender Systems

Gating

• 2014 (TASLP) * [LHUC] Learning Hidden Unit Contributions for Unsupervised Acoustic Model Adaptation

• 2018 (CVPR) * [SENet] Squeeze-and-Excitation Networks

• 2019 (Sina) (Recsys) [FiBiNET] FiBiNET - combining feature importance and bilinear feature interaction for click-through rate prediction

• 2020 (Sina) (Arxiv) [GateNet] GateNet - Gating-Enhanced Deep Network for Click-Through Rate Prediction

• 2021 (Sina) (Arxiv) [ContextNet] ContextNet - A Click-Through Rate Prediction Framework Using Contextual information to Refine Feature Embedding

• 2021 (Sina) （DLP-KDD) [MaskNet] MaskNet - Introducing Feature-Wise Multiplication to CTR Ranking Models by Instance-Guided Mask

• 2023 (Kuaishou) (KDD) [PEPNet] PEPNet - Parameter and Embedding Personalized Network for Infusing with Personalized Prior Information

• 2023 (Sina) (CIKM) [FiBiNet++] FiBiNet++ - Reducing Model Size by Low Rank Feature Interaction Layer for CTR Prediction

• 2025 (Bytedance) [ADS] Adaptive Domain Scaling for Personalized Sequential Modeling in Recommenders

LLM_Ranking

• 2019 (CIKM) [AutoInt] AutoInt -Automatic Feature Interaction Learning via Self-Attentive Neural Networks

• 2020 (Arxiv) Scaling Laws for Neural Language Models

• 2021 (Baidu) (KDD) Pre-trained Language Model based Ranking in Baidu Search

• 2023 (Arxiv) [E4SRec] E4SRec - An Elegant Effective Efficient Extensible Solution of Large Language Models for Sequential Recommendation

• 2024 (Alibaba) (Arxiv) [BAHE] Breaking the Length Barrier - LLM-Enhanced CTR Prediction in Long Textual User Behaviors

• 2024 (Bytedance) (Arxiv) [HLLM] HLLM - Enhancing Sequential Recommendations via Hierarchical Large Language Models for Item and User Modeling

• 2024 (Google) (Arxiv) LLMs for User Interest Exploration in Large-scale Recommendation Systems

• 2024 (Google) (Arxiv) [CALRec] CALRec - Contrastive Alignment of Generative LLMs for Sequential Recommendation

• 2024 (Kuaishou) (Arxiv) [LEARN] LEARN - Knowledge Adaptation from Large Language Model to Recommendation for Practical Industrial Application

• 2024 (Kuaishou) (KDD) [NAR4Rec] Non-autoregressive Generative Models for Reranking Recommendation

• 2024 (Meta) (Arxiv) Unifying Generative and Dense Retrieval for Sequential Recommendation

• 2024 (Meta) (Arxiv) [SUM] Scaling User Modeling - Large-scale Online User Representations for Ads Personalization in Meta

• 2025 (Alibaba) (Arxiv) Unlocking Scaling Law in Industrial Recommendation Systems with a Three-step Paradigm based Large User Model

• 2025 (Alibaba) (Arxiv) [HeterRec] Hierarchical Causal Transformer with Heterogeneous Information for Expandable Sequential Recommendation

• 2025 (Alibaba) (Arxiv) [URM] Large Language Models Are Universal Recommendation Learners

• 2025 (Alibaba) (KDD) [GPSD] Scaling Transformers for Discriminative Recommendation via Generative Pretraining

• 2025 (Alibaba) (WWW) Explainable LLM-driven Multi-dimensional Distillation for E-Commerce Relevance Learning

• 2025 (Amazon) (Arxiv) SynerGen - Contextualized Generative Recommender for Unified Search and Recommendation

• 2025 (Arxiv) (Pinterest) [PinRec] PinRec - Outcome-Conditioned, Multi-Token Generative Retrieval for Industry-Scale Recommendation Systems

• 2025 (Arxiv) (Xiaohongshu) [GenRank] Towards Large-scale Generative Ranking

• 2025 (Baidu) (Arxiv) [COBRA] Sparse Meets Dense -Unified Generative Recommendations with Cascaded Sparse-Dense Representations

• 2025 (Google) (Arxiv) User Feedback Alignment for LLM-powered Exploration in Large-scale Recommendation Systems

• 2025 (Google) (Arxiv) [STAR] STAR - A Simple Training-free Approach for Recommendations using Large Language Models

• 2025 (Kuaishou) (Arxiv) [GenSAR] Unified Generative Search and Recommendation

• 2025 (Kuaishou) (Arxiv) [LARM] LLM-Alignment Live-Streaming Recommendationpdf

• 2025 (Kuaishou) (Arxiv) [LEARN] LEARN - Knowledge Adaptation from Large Language Model to Recommendation for Practical Industrial Application

• 2025 (Kuaishou) (Arxiv) [OneRec-V2] OneRec Technical Report v2

• 2025 (Kuaishou) (Arxiv) [OneRec] OneRec - Unifying Retrieve and Rank with Generative Recommender and Preference Alignment

• 2025 (Kuaishou) (Arxiv) [OneRec] OneRec Technical Report

• 2025 (Meituan) (Arxiv) [DFGR] Action is All You Need - Dual-Flow Generative Ranking Network for Recommendation

• 2025 (Meituan) (Arxiv) [MTGR] MTGR - Industrial-Scale Generative Recommendation Framework in Meituan

• 2025 (Meituan) (Arxiv) [UniROM] One Model to Rank Them All - Unifying Online Advertising with End-to-End Learning

• 2025 (Meta) (Arxiv) [HyperCast] Realizing Scaling Laws in Recommender Systems - A Foundation–Expert Paradigm for Hyperscale Model Deployment

• 2025 (Microsoft) (KDD)Towards Web-scale Recommendations with LLMs - From Quality-aware Ranking to Candidate Generation

• 2025 (Pinterest) (Arxiv) [PinFM] PinFM - Foundation Model for User Activity Sequences at a Billion-scale Visual Discovery Platform

• 2025 (Tencent) (Arxiv) [GPR] GPR - Towards a Generative Pre-trained One-Model Paradigm for Large-Scale Advertising Recommendation

• 2025 (eBay) (Arxiv) LLMDistill4Ads - Using Cross-Encoders to Distill from LLM Signals for Advertiser Keyphrase Recommendations at eBay

• 2025 （Bytedance) (Arxiv) [OneTrans] OneTrans - Unified Feature Interaction and Sequence Modeling with One Transformer in Industrial Recommender

• 2025 （Meta) (Arxiv) [ExFM] External Large Foundation Model - How to Efficiently Serve Trillions of Parameters for Online Ads Recommendation

• 2026 (Alibaba) (Arxiv) [SORT] SORT - A Systematically Optimized Ranking Transformer for Industrial-scale Recommenders

• 2026 (Bytedance) (Arxiv) [HyFormer] HyFormer - Revisiting the Roles of Sequence Modeling and Feature Interaction in CTR Prediction

• 2026 (Bytedance) (Arxiv) [MixFormer] MixFormer - Co-Scaling Up Dense and Sequence in Industrial Recommenders

• 2026 (Kuaishou) (Arxiv) [GR4AD] Generative Recommendation for Large-Scale Advertising

• 2026 (Meta) (Arxiv) [Kunlun] Kunlun - Establishing Scaling Laws for Massive-Scale Recommendation Systems through Unified Architecture Design

• 2026 (Meta) (Arxiv) [WHALE] WHALE - A Scalable Unified Model for Recommendation with Wukong-HSTU Architecture

• 2026 （Bytedance) (Arxiv) [MDL] MDL - A Unified Multi-Distribution Learner in Large-scale Industrial Recommendation through Tokenization

Loss

• 2015 (Twitter) (KDD) Click-through Prediction for Advertising in Twitter Timeline

• 2022 (Google) (KDD) Scale Calibration of Deep Ranking Models

• 2023 (Alibaba) (KDD) Joint Optimization of Ranking and Calibration with Contextualized Hybrid Model

• 2023 (Google) (CIKM) Regression Compatible Listwise Objectives for Calibrated Ranking with Binary Relevance

• 2024 (Tencent) (KDD) Understanding the Ranking Loss for Recommendation with Sparse User Feedback

Multi-Modal

• 2018 (Alibaba) (CIKM) [Image CTR] Image Matters - Visually Modeling User Behaviors Using Advanced Model Server

• 2024 (Alibaba) (CIKM) Enhancing Taobao Display Advertising with Multimodal Representations - Challenges, Approaches and Insights

• 2026 (Alibaba) (WSDM) [MOON] MOON - Generative MLLM-based Multimodal Representation Learning for E-commerce Product Understanding

Multi-domain-Multi-Scenario

• 2014 (TASLP) * [LHUC] Learning Hidden Unit Contributions for Unsupervised Acoustic Model Adaptation

• 2015 (Microsoft) (WWW) A Multi-View Deep Learning Approach for Cross Domain User Modeling in Recommendation Systems

• 2018 (Google) (KDD) ** [MMoE] Modeling task relationships in multi-task learning with multi-gate mixture-of-experts

• 2019 (Alibaba) (CIKM) [WE-CAN] Cross-domain Attention Network with Wasserstein Regularizers for E-commerce Search

• 2020 (Alibaba) (Arxiv) [SAML] Scenario-aware and Mutual-based approach for Multi-scenario Recommendation in E-Commerce

• 2020 (Alibaba) (CIKM) [HMoE] Improving Multi-Scenario Learning to Rank in E-commerce by Exploiting Task Relationships in the Label Space

• 2020 (Alibaba)(CIKM) [MiNet] MiNet - Mixed Interest Network for Cross-Domain Click-Through Rate Prediction

• 2020 (Tencent) (Recsys) ** [PLE] Progressive Layered Extraction (PLE) - A Novel Multi-Task Learning (MTL) Model for Personalized Recommendations

• 2021 (Alibaba) (CIKM) * [ZEUS] Self-Supervised Learning on Users’ Spontaneous Behaviors for Multi-Scenario Ranking in E-commerce

• 2021 (Alibaba) (CIKM) ** [STAR] One Model to Serve All - Star Topology Adaptive Recommender for Multi-Domain CTR Prediction

• 2021 (Google) (ICLR) HyperGrid Transformers - Towards A Single Model for Multiple Tasks

• 2021 (Kwai) (Arxiv) [POSO] POSO - Personalized Cold Start Modules for Large-scale Recommender Systems

• 2022 (Alibaba) (CIKM) AdaSparse - Learning Adaptively Sparse Structures for Multi-Domain Click-Through Rate Prediction

• 2022 (Alibaba) (NIPS) ** [APG] APG - Adaptive Parameter Generation Network for Click-Through Rate Prediction

• 2023 (Alibaba) (CIKM) [HC2] Hybrid Contrastive Constraints for Multi-Scenario Ad Ranking

• 2023 (Alibaba) (CIKM) [MMN] Masked Multi-Domain Network - Multi-Type and Multi-Scenario Conversion Rate Prediction with a Single Model

• 2023 (Alibaba) (CIKM) [Rec4Ad] Rec4Ad - A Free Lunch to Mitigate Sample Selection Bias for Ads CTR Prediction in Taobao

• 2023 (Alibaba) (SIGIR) [MARIA] Multi-Scenario Ranking with Adaptive Feature Learning

• 2023 (CIKM) [HAMUR] HAMUR - Hyper Adapter for Multi-Domain Recommendation

• 2023 (Huawei) (CIKM) [DFFM] DFFM - Domain Facilitated Feature Modeling for CTR Prediction

• 2023 (Kuaishou) (KDD) * [PEPNet] PEPNet - Parameter and Embedding Personalized Network for Infusing with Personalized Prior Information

• 2023 (Tencent) (KDD) Scenario-Adaptive Feature Interaction for Click-Through Rate Prediction

• 2024 (Alibaba) (CIKM) * [MultiLoRA] MultiLoRA - Multi-Directional Low-Rank Adaptation for Multi-Domain Recommendation

• 2024 (Alibaba) (RecSys) * [MLoRA] MLoRA - Multi-Domain Low-Rank Adaptive Network for Click-Through Rate Prediction

• 2024 (Kuaishou) (SIGIR) [M3oE] M3oE - Multi-Domain Multi-Task Mixture-of-Experts Recommendation Framework

• 2024 (Tencent) (KDD) [LCN] Cross-Domain LifeLong Sequential Modeling for Online Click-Through Rate Prediction

• 2024 (WSDM) Exploring Adapter-based Transfer Learning for Recommender Systems - Empirical Studies and Practical Insights

• 2025 (Kuaishou) (KDD) [HoME] HoME - Hierarchy of Multi-Gate Experts for Multi-Task Learning at Kuaishou

Multi-task

• (2018) (ICML) GradNorm - Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks

• 2014 (TASLP) [LHUC] Learning Hidden Unit Contributions for Unsupervised Acoustic Model Adaptation

• [2017 (Google) (ICLR) [Sparsely-Gated MOE] Outr