2026-01-01 至 2026-08-24 历史扫描清单与实现批次¶
本页不是把原始关键词命中冒充入选论文。机器可读全集见
2026-historical-candidates.json：每个去重后的新候选都保留检索来源、初筛桶和计划状态。
漏斗与状态¶
| 项目 | 数量 |
|---|---|
| 去重后的新候选 | 3906 |
| 需要全文审查 | 404 |
| 当前 PR 已实现 | 74 |
| 固定后续实现队列 | 0 |
| 仍待全文决定、未承诺实现 | 331 |
| P2 或查询碰撞（保留审计记录） | 3501 |
fulltext-review-backlog 不是拒绝。只有核验机构、正文实验、代码状态和与现有实现的增量后，才能晋级后续批次或写入带原因的终态。工业搜广推继续执行量化线上 A/B/明确全流量硬门槛。
固定实现批次¶
B00：当前 PR 已完成¶
- 2607.26369 ClockRoPE: Random Fourier Rotations for Temporal Routine Modeling — PR #120 / ClockRoPE
- 2607.27475 OneShot: Index-in-Ranking with Neural Scoring for Large-Scale Retrieval — PR #120 / OneShot
- 2607.24789 NEXT: Reasoning-Driven Video Recommendation via a Vision-Language Model — PR #120 / NEXT
- 2606.26859 AgentX: Towards Agent-Driven Self-Iteration of Industrial Recommender Systems — PR #120 / AgentX
B01：8 月工业生成推荐与多模态（已完成）¶
- 2608.21012 From a Static Multi-Level Small Semantic Codebook to a Dynamic Single-Level Large Semantic Codebook for Generative Recommendation — 搜广推与 LLM 应用
- 2608.18322 Multimedia Asset Personalization via Multimodal Embeddings at Netflix — 搜广推与 LLM 应用
- 2608.17613 Once Generated, Ranked: End-to-End Generative Slate Recommendation with Unified Semantic-Collaborative IDs — 搜广推与 LLM 应用
- 2608.09634 IntHQ: Task-Interactive Hierarchical Query on Dual-Stream Representations for Generative Recommendation — 搜广推与 LLM 应用
- 2608.07989 PushDualGen: Enabling LLMs to Generate Semantic IDs with Interpretable Copy for Industrial Push Recommendation — 搜广推与 LLM 应用
B02：7 月工业生成推荐、Agent harness 与搜索（已完成）¶
- 2607.29241 RecHarness: A Bandit-Routed Agentic Harness for Self-Evolving Recommender Systems — 搜广推与 LLM 应用
- 2607.29213 GALA: Generative Aligned Learning for Adaptive Multimodal Representation in the Taobao Shangou Recommender System — 搜广推与 LLM 应用
- 2607.27789 From Understanding to Action: Feedback-Grounded Policy Discovery for Generative Recommendation — 搜广推与 LLM 应用
- 2607.14835 LLM-Based Re-Ranking for Real Estate Search — 搜广推与 LLM 应用
- 2607.14418 Adaptive Ad Load Design for Sponsored Search Markets: Evidence, Theory, and Deployment — 搜广推与 LLM 应用
- 2607.26073 Guess Where You Go: Generative Next Point-of-Interest Recommendation in Amap — 搜广推与 LLM 应用
- 2606.31031 GenPage: Towards End-to-End Generative Homepage Construction at Netflix — 搜广推与 LLM 应用
B03：6–5 月序列建模、生成搜索与多场景排序（已完成）¶
- 2606.19108 JourneyFormer: Encoding Airbnb Guest Journey with Sequence Modeling — 搜广推与 LLM 应用
- 2605.26717 L2Rec: Towards Dual-View Understanding of LLMs for Personalized Recommendation — 搜广推与 LLM 应用
- 2605.25514 From Item-Only to Query-Item: Query-Conditioned Generative Search with QGS in Quark — 搜广推与 LLM 应用
- 2605.23702 TubiFM: Unified Item, Carousel, and Search Ranking for Streaming Discovery — 搜广推与 LLM 应用
- 2605.21752 PEARL: Unbiased Percentile Estimation via Contrastive Learning for Industrial-Scale Livestream Recommendation — 搜广推与 LLM 应用
- 2605.17863 DADF: A Distribution-Aware Debiasing Framework for Watch-Time Regression in Recommender Systems — 搜广推与 LLM 应用
B04：推荐 RL、知识迁移与 Semantic ID（已完成）¶
- 2605.16344 A Production-Ready RL Framework for Personalized Utility Tuning with Pareto Sweeping in Pinterest Recommender Systems — 搜广推与 LLM 应用
- 2605.05730 Effective Knowledge Transfer for Multi-Task Recommendation Models — 搜广推与 LLM 应用
- 2604.23522 Beyond Static Collision Handling: Adaptive Semantic ID Learning for Multimodal Recommendation at Industrial Scale — 搜广推与 LLM 应用
- 2604.12234 UniRec: Bridging the Expressive Gap between Generative and Discriminative Recommendation via Chain-of-Attribute — 搜广推与 LLM 应用
- 2603.24226 UniScale: Synergistic Entire Space Data and Model Scaling for Search Ranking — 搜广推与 LLM 应用
- 2603.22916 GateSID: Adaptive Gating for Balancing Semantic and Collaborative Signals in Recommendation — 搜广推与 LLM 应用
B05：电商生成、搜索融合与工业排序（已完成）¶
- 2603.19710 AIGQ: An End-to-End Hybrid Generative Architecture for E-commerce Query Recommendation — 搜广推与 LLM 应用
- 2603.19585 SaFRO: Satisfaction-Aware Fusion via Dual-Relative Policy Optimization for Short-Video Search — 搜广推与 LLM 应用
- 2603.03988 SORT: A Systematically Optimized Ranking Transformer for Industrial-scale Recommenders — 搜广推与 LLM 应用
- 2603.00632 Stop Treating Collisions Equally: Qualification-Aware Semantic ID Learning for Recommendation at Industrial Scale — 搜广推与 LLM 应用
- 2602.20995 Generative Pseudo-Labeling for Pre-Ranking with LLMs — 搜广推与 LLM 应用
- 2602.17058 A Long-term Value Prediction Framework In Video Ranking — 搜广推与 LLM 应用
B06：2 月召回、广告、长序列与 LLM 排序（已完成）¶
- 2602.12968 RGAlign-Rec: Ranking-Guided Alignment for Latent Query Reasoning in Recommendation Systems — 搜广推与 LLM 应用
- 2602.12354 An Industrial-Scale Sequential Recommender for LinkedIn Feed Ranking — 搜广推与 LLM 应用
- 2602.11410 CADET: Context-Conditioned Ads CTR Prediction With a Decoder-Only Transformer — 搜广推与 LLM 应用
- 2602.09744 DiffuReason: Bridging Latent Reasoning and Generative Refinement for Sequential Recommendation — 搜广推与 LLM 应用
- 2602.09401 SARM: LLM-Augmented Semantic Anchor for End-to-End Live-Streaming Ranking — 搜广推与 LLM 应用
- 2602.09194 ML-DCN: Masked Low-Rank Deep Crossing Network Towards Scalable Ads Click-through Rate Prediction at Pinterest — 搜广推与 LLM 应用
- 2602.01023 Unifying Ranking and Generation in Query Auto-Completion via Retrieval-Augmented Generation and Multi-Objective Alignment — 搜广推与 LLM 应用
B07：LLM 架构、长上下文、KV cache 与评测基础设施（已完成）¶
- 2608.12831 Fast A/B/n Testing: Exact Multi-Policy Comparison via Tree-Coupled Feedback Sharing — LLM 后训练、搜广推与 LLM 应用
- 2608.10296 Cracks in the Foundation: Seemingly Minor Architectural Choices Impact Long Context Extension — 基础模型
- 2608.08878 DistillCache: KL-Guided Adaptive KV-Cache Eviction for Memory-Efficient LLM Inference — 基础模型、LLM 后训练
- 2608.06849 Autonomy-of-Heads: Data-Free Sparse Attention from Frozen Query-Key Geometry — 基础模型
- 2608.05000 Towards Physics of Multimodal Pretraining: Knowledge Flow, Modality Synergy, Early Unification, and Recipes — 基础模型
- 2608.01672 Learning What to Remember: Test-Time Training via Context Distillation — 基础模型
- 2608.02032 DART: Decoded Attention over Recurrent States for Efficient Long-Context Sequence Modeling — 基础模型
- 2607.29032 TransMem: Transforming Hidden States into Memory for Large Language Models — Agent、基础模型
- 2607.17715 C\(^2\)KV: Compressed and Composable KV Cache Reuse for Efficient LLM Inference — 基础模型
B08：OPD 与多教师/过程蒸馏（已完成）¶
- 2608.19408 Beyond Imitation: Filtering On-Policy Distillation by Reasoning Progress — LLM 后训练
- 2608.09745 SR-OPSD: Self-Referenced On-Policy Self-Distillation — LLM 后训练
- 2608.05802 On-Policy Delta Distillation for Multilingual Math Reasoning — LLM 后训练
- 2608.03673 CausalOPD: First-Wrong-Step Supervision for Distilling Causal Chain Reasoning — LLM 后训练
- 2608.03092 SMOPD: Multi-Reward Reinforcement Learning via Specialize-and-Merge Online Policy Distillation — LLM 后训练
- 2608.00782 Distill Where You Fail: Recovering Learning Signals of Negative RL-Groups from Adaptive Teacher Guidance — LLM 后训练
B09：Rubric、外部 rollout 与多奖励 RL（已完成）¶
- 2608.16072 Learn What's Left, Not What's Mastered: Saturation Aware Advantage Reweighting for Multi-Reward Policy Optimization — LLM 后训练
- 2608.11669 Rubric Dropout: A Simple Way to Mitigate Reward Hacking in Rubric-as-Reward RL — LLM 后训练
- 2608.01717 Beyond On-Policy Exploration: Integrating External Policy Rollouts for Reinforcement Learning in Diffusion Language Models — LLM 后训练
- 2607.28026 Contrastive Reinforced Policy Optimization via Privileged Self-Distillation — Agent、LLM 后训练
- 2607.26873 SERPO: Self-Evolving Rubric Policy Optimization for Open-Ended Test-Time Reinforcement Learning — LLM 后训练
- 2607.19331 ISO: An RLVR-Native Optimization Stack — LLM 后训练
B10：Agentic RL 与长时序 credit assignment（已完成）¶
- 2608.19842 SAPO: Single-Rollout Autoregressive Policy Optimization for Agentic Reinforcement Learning — Agent
- 2608.19197 SPADE: Self-Play in Adaptive Synthetic Executable Environments — Agent、LLM 后训练
- 2608.18682 RTPO: Reverse-Turn Policy Optimization for Stabilizing Agentic RL Training — Agent
- 2608.17289 PlanPO: Group Planning-Aware Policy Optimization for Multi-Turn Agentic LLMs — Agent
- 2608.16156 TRCA: Transition-wise Rubric Credit Assignment for Long-horizon LLM Agents — Agent
- 2608.11967 LoongReflect: Boosting Long-Horizon Reflection in Search Agents via Global Perspective Distillation — Agent、基础模型、LLM 后训练
B11：Agent 记忆、工具规划与自进化系统（已完成）¶
- 2608.15703 HyMem: Hierarchical Context Management for Long-Horizon Agents via Information Isolation — Agent
- 2608.09380 OpenLoopEvolve: A Verifiable Self-Evolution Framework for Loop Policies in Long-Horizon Complex Tasks — Agent
- 2608.06811 Coupling Planning with Episodic Memory in LLM Agents for Software Issue Resolution — Agent
- 2608.03468 ToolLIFT: Lifting Tool-Specific Trajectories into Function-Level Graphs for Generalizable Tool Planning — Agent
- 2608.02650 HyperAgent: Planning and Acting over Tool-Schema Hypergraphs for Tool-Use LLM Agents — Agent
- 2607.28527 MANTA: Multi-Agent Network Topology Adaptation for Self-Evolving Multi-Agent Systems — Agent
全部 404 个全文审查候选¶
状态含义：Bxx 为固定实现批次；B00 已完成；待全文 表示尚未承诺实现，但不会从账本消失。
| 日期 | 领域 | 论文 | 初筛 | 计划 |
|---|---|---|---|---|
| 2026-08-21 | Agent | 2608.21101 ClawSentry: A Progressive Multi-Tier Security Monitor for Safeguarding Autonomous LLM Agents | manual-review | 待全文 |
| 2026-08-21 | 基础模型 | 2608.21030 COMET: Contrastive Motion-Enhanced Temporal Reasoning for Video Multimodal Large Language Models | manual-review | 待全文 |
| 2026-08-21 | 搜广推与 LLM 应用 | 2608.21012 From a Static Multi-Level Small Semantic Codebook to a Dynamic Single-Level Large Semantic Codebook for Generative Recommendation | industrial-fulltext-review | B01 |
| 2026-08-21 | Agent | 2608.20844 TRACE: Agentic Catalog Enrichment with Multi-source Evidence Grounding | manual-review | 待全文 |
| 2026-08-21 | Agent | 2608.20631 Weighted Memory Tree: Remembering What Matters for Long-Horizon LLM Agents | manual-review | 待全文 |
| 2026-08-20 | 基础模型 / LLM 后训练 | 2608.20331 G-CARL: Grounded Checklist-Aligned Reward Learning for Patient-Oriented Medical Report Interpretation | manual-review, p2-deferred-review | 待全文 |
| 2026-08-20 | 基础模型 | 2608.20315 Explainable Transformer Models for Clinical Prediction Tasks on Structured Electronic Health Records | manual-review | 待全文 |
| 2026-08-20 | Agent | 2608.20314 MidTool: Mid-training Data Synthesis for Agentic Tool Use | manual-review | 待全文 |
| 2026-08-20 | Agent | 2608.19842 SAPO: Single-Rollout Autoregressive Policy Optimization for Agentic Reinforcement Learning | manual-review | B10 |
| 2026-08-20 | Agent / LLM 后训练 | 2608.19803 MileGPO: Milestone Inference with Local Evidence for Graph-Based Policy Optimization of Long-Horizon LLM Agents | manual-review, p2-deferred-review | 待全文 |
| 2026-08-20 | Agent / 基础模型 | 2608.19662 ReCache: Efficient KV Cache Reuse and Compression for Tool-Augmented LLM Agents | manual-review, p2-deferred-review | 待全文 |
| 2026-08-20 | Agent | 2608.19564 Remember, Verify, or Ask? Cross-Family Evaluation of Memory Commitment in LLM Agents | manual-review | 待全文 |
| 2026-08-19 | LLM 后训练 | 2608.19408 Beyond Imitation: Filtering On-Policy Distillation by Reasoning Progress | manual-review | B08 |
| 2026-08-19 | 基础模型 | 2608.19395 HYDRA: A Heterogeneous Chiplet DSE Framework for Serving Dynamic Hybrid LLM Workloads | manual-review | 待全文 |
| 2026-08-19 | Agent / LLM 后训练 | 2608.19197 SPADE: Self-Play in Adaptive Synthetic Executable Environments | manual-review, p2-deferred-review | B10 |
| 2026-08-19 | Agent | 2608.18884 Training-Free Inference-Time Self-Reflection and Cost-Bounded Early Stopping for Large Language Models | manual-review | 待全文 |
| 2026-08-19 | 基础模型 | 2608.18733 Flama: a Python framework for development and deployment of production-ready APIs, machine learning, and LLM services | manual-review | 待全文 |
| 2026-08-19 | Agent | 2608.18682 RTPO: Reverse-Turn Policy Optimization for Stabilizing Agentic RL Training | manual-review | B10 |
| 2026-08-19 | 基础模型 | 2608.18656 FlashAttention for Scalable Vector Architectures | manual-review | 待全文 |
| 2026-08-19 | 搜广推与 LLM 应用 | 2608.18637 PILOT Technical Report | manual-review | 待全文 |
| 2026-08-19 | 基础模型 | 2608.18628 When Safety Overrides Vision: Exploring Dynamics between Vision Influence and Safety Alignment in Vision-Language Models | manual-review | 待全文 |
| 2026-08-19 | 基础模型 / LLM 后训练 | 2608.18578 Compress and Forget: bitsandbytes Quantization Amplifies Proactive Interference in LLMs | manual-review, p2-deferred-review | 待全文 |
| 2026-08-19 | Agent | 2608.18423 FM-Bench: A Benchmark for Long-Horizon Management with Competing Agents | manual-review | 待全文 |
| 2026-08-18 | 搜广推与 LLM 应用 | 2608.18322 Multimedia Asset Personalization via Multimodal Embeddings at Netflix | industrial-fulltext-review | B01 |
| 2026-08-18 | Agent / 基础模型 / LLM 后训练 | 2608.18008 Policy-Invariant Reward Shaping from LLM Feedback: A Framework for Hybrid RL Agents | manual-review, p2-deferred-review | 待全文 |
| 2026-08-18 | 基础模型 | 2608.17950 Do Large Language Models Play Six Degrees of Separation? Measuring Topological Compression in Long-Context Manifolds | manual-review | 待全文 |
| 2026-08-18 | 搜广推与 LLM 应用 | 2608.17613 Once Generated, Ranked: End-to-End Generative Slate Recommendation with Unified Semantic-Collaborative IDs | industrial-fulltext-review | B01 |
| 2026-08-18 | Agent / LLM 后训练 | 2608.17524 Evaluating RL Explainability Methods by How Much They Help Fix Bugs in Agents | manual-review, p2-deferred-review | 待全文 |
| 2026-08-18 | 搜广推与 LLM 应用 | 2608.17316 Empowering Compact LLMs with Fusion of Layer-wise Exits for Recommendation | manual-review | 待全文 |
| 2026-08-18 | Agent | 2608.17289 PlanPO: Group Planning-Aware Policy Optimization for Multi-Turn Agentic LLMs | manual-review | B10 |
| 2026-08-17 | Agent / 基础模型 | 2608.17209 Teach and Grow: An Agent-Centered Architecture for General Robot Learning | manual-review, p2-deferred-review | 待全文 |
| 2026-08-17 | Agent / 基础模型 | 2608.18171 Looped Language Models Improve Compositional Tool Calling | manual-review, p2-deferred-review | 待全文 |
| 2026-08-17 | Agent / 基础模型 | 2608.16889 Don't Drop the BATON: Long-Horizon Robot Manipulation via Agentic Subtask Exploration and Transition-aware Memory | manual-review, p2-deferred-review | 待全文 |
| 2026-08-17 | 基础模型 | 2608.16844 Proteus: Incremental Memory Activation for Long-Context Sequence Modeling | manual-review | 待全文 |
| 2026-08-17 | 基础模型 | 2608.16514 Matched Outcomes, Divergent Gaze: How Foveated MLLMs Search Compared to Humans | manual-review | 待全文 |
| 2026-08-17 | Agent | 2608.16447 HaReCAP: Habitual-action Grounding for Recursive Large Language Model Agents | manual-review | 待全文 |
| 2026-08-17 | 基础模型 / LLM 后训练 | 2608.16419 PertMind: Eliciting Emergent Biological Reasoning in LLM via Reinforcement Learning on Cellular Perturbation Data | manual-review, p2-deferred-review | 待全文 |
| 2026-08-17 | 基础模型 / LLM 后训练 | 2608.16316 Deep Thought Alignment: Trajectory-Level Latent Distillation for Video Reasoning | manual-review, p2-deferred-review | 待全文 |
| 2026-08-17 | Agent / 基础模型 | 2608.16168 QUMem: Personalized Memory for Query-Conditioned User-State Inference in LLM Agents | manual-review, p2-deferred-review | 待全文 |
| 2026-08-17 | Agent | 2608.16156 TRCA: Transition-wise Rubric Credit Assignment for Long-horizon LLM Agents | manual-review | B10 |
| 2026-08-17 | LLM 后训练 | 2608.16072 Learn What's Left, Not What's Mastered: Saturation Aware Advantage Reweighting for Multi-Reward Policy Optimization | manual-review | B09 |
| 2026-08-16 | 基础模型 | 2608.15962 SEER: Long-Context Reasoning via Selective Visual-Text Compression | manual-review | 待全文 |
| 2026-08-16 | Agent / LLM 后训练 / 搜广推与 LLM 应用 | 2608.15949 Ask to Be Sure: Informative Interactions for Confident Multi-Turn LLM Recommendation | manual-review | 待全文 |
| 2026-08-16 | Agent | 2608.15703 HyMem: Hierarchical Context Management for Long-Horizon Agents via Information Isolation | manual-review | B11 |
| 2026-08-16 | 基础模型 / LLM 后训练 | 2608.15602 FluxBin: Flexible LUT-based Ultra-low-bit LLM Inference by Algorithm-Kernel Synergy | manual-review, p2-deferred-review | 待全文 |
| 2026-08-16 | Agent | 2608.15579 Kozuchi Agent: A Language-Agnostic Open-Weight Agent for Software Repair | manual-review | 待全文 |
| 2026-08-15 | Agent | 2608.15175 LAPF: LLM-Agent-Based Path Finder Using the UAVScenes Dataset | manual-review | 待全文 |
| 2026-08-15 | Agent / 基础模型 | 2608.15032 Handoff-H1: An Orchestrated Vision-Agent System for Material Quantity Takeoff from Construction Blueprints | manual-review, p2-deferred-review | 待全文 |
| 2026-08-15 | Agent | 2608.15016 Hierarchical Agentic Incident Response with Digital-Twin-Validated Attack Inference | manual-review | 待全文 |
| 2026-08-14 | 基础模型 | 2608.14198 MINT: A Universal Zero-Shot Predictor for Transaction Data | manual-review | 待全文 |
| 2026-08-14 | Agent / 搜广推与 LLM 应用 | 2608.14068 MACS: A Hybrid Multi-Agent Framework for Reliable Conversational E-Commerce Recommendation | manual-review | 待全文 |
| 2026-08-14 | Agent / 基础模型 | 2608.13900 Agentic Transaction: Towards ACID-Compliant Agent Systems | manual-review, p2-deferred-review | 待全文 |
| 2026-08-14 | LLM 后训练 | 2608.13854 Bootstrapping Niche Multilingual Code Translation via Reinforcement Learning with Execution-Based Verifiable Supervision | manual-review | 待全文 |
| 2026-08-13 | Agent | 2608.13681 Fine-Tuning Qwen3-27B for C-to-Rust Code Translation: A Three-Stage Curriculum of Pretraining, Debugging-Aware SFT, and Task-Specific SFT | manual-review | 待全文 |
| 2026-08-13 | Agent / 基础模型 / LLM 后训练 | 2608.13505 Intern-S2-Preview: Scientific Agentic Foundation Model | manual-review, p2-deferred-review | 待全文 |
| 2026-08-13 | Agent / 基础模型 | 2608.13463 MLLM-Routed Heterogeneous Ensembles for Robust Cross-Dataset Image Classification | manual-review, p2-deferred-review | 待全文 |
| 2026-08-13 | 基础模型 | 2608.13426 Reduced Matrix Multiplication: Input-Adaptive Matrix-Product Reduction for LLM Inference | manual-review | 待全文 |
| 2026-08-13 | Agent | 2608.13317 StateBridge: Training-free Hidden-state Alignment for Latent Communication in LLM Multi-Agent Systems | manual-review | 待全文 |
| 2026-08-13 | 搜广推与 LLM 应用 | 2608.12986 STAR: Structured Tokenization and Target-Aware Interest Representation for PCVR Prediction | manual-review | 待全文 |
| 2026-08-13 | LLM 后训练 / 搜广推与 LLM 应用 | 2608.12831 Fast A/B/n Testing: Exact Multi-Policy Comparison via Tree-Coupled Feedback Sharing | industrial-fulltext-review, query-collision | B07 |
| 2026-08-13 | 搜广推与 LLM 应用 | 2608.12778 DrEM: Dual-Side Robust Ensemble Ranking from Noisy User Preference Predictions in Video Recommendation | industrial-fulltext-review | 待全文 |
| 2026-08-12 | 基础模型 | 2608.12099 RT-SEMamba: Real-Time Speech Enhancement Mamba via Progressive Knowledge Distillation | manual-review | 待全文 |
| 2026-08-12 | 搜广推与 LLM 应用 | 2608.11973 Sci-Surf: Navigating Scientific Literature Discovery through Human Feedback and Intelligent Summarization | manual-review | 待全文 |
| 2026-08-12 | Agent / 基础模型 / LLM 后训练 | 2608.11967 LoongReflect: Boosting Long-Horizon Reflection in Search Agents via Global Perspective Distillation | manual-review, p2-deferred-review | B10 |
| 2026-08-12 | LLM 后训练 | 2608.11715 When the API Speaks the Wrong Language: Revisiting Post-Training for Multilingual Tool Use | manual-review | 待全文 |
| 2026-08-12 | LLM 后训练 | 2608.11669 Rubric Dropout: A Simple Way to Mitigate Reward Hacking in Rubric-as-Reward RL | manual-review | B09 |
| 2026-08-11 | LLM 后训练 | 2608.11350 Self-Evolving Embodied Agents via Skill-Harness Evolution | manual-review | 待全文 |
| 2026-08-11 | Agent / LLM 后训练 | 2608.11191 Test-Time Self-Evolving GUI Visual Grounding via Reflection-Guided On-Policy Self-Distillation | manual-review, p2-deferred-review | 待全文 |
| 2026-08-11 | Agent / 基础模型 / LLM 后训练 | 2608.11152 Scheduling Mixed RL Rollouts Beyond Prefix Locality | manual-review, p2-deferred-review | 待全文 |
| 2026-08-11 | 基础模型 | 2608.10908 Order Matters: LVLMs as Judges for Temporal Reasoning in Image Sequences | manual-review | 待全文 |
| 2026-08-11 | LLM 后训练 | 2608.10812 Reference-Free Post-Training of Open Large Language Models for Multilingual Machine Translation | manual-review | 待全文 |
| 2026-08-11 | Agent | 2608.10530 On Understanding, Identifying, and Mitigating Vulnerabilities in Agentic Large Language Models | manual-review | 待全文 |
| 2026-08-11 | 搜广推与 LLM 应用 | 2608.10447 Towards Efficient Reasoning in LLM-Based Recommender Systems via Model Merging | manual-review | 待全文 |
| 2026-08-11 | Agent / 基础模型 | 2608.10430 Actionable Hallucination Detection: Translating Latent Uncertainty into Agentic Critique | manual-review, p2-deferred-review | 待全文 |
| 2026-08-10 | 基础模型 | 2608.10296 Cracks in the Foundation: Seemingly Minor Architectural Choices Impact Long Context Extension | manual-review | B07 |
| 2026-08-10 | 基础模型 / LLM 后训练 / 搜广推与 LLM 应用 | 2608.10182 From Prediction to Incrementality: Causal Optimization for Large-Scale Targeting and Recommendation | industrial-fulltext-review, p2-deferred-review, query-collision | 待全文 |
| 2026-08-10 | Agent / LLM 后训练 | 2608.10126 Procedural Fairness Failures in RLHF from Preference Averaging | manual-review, p2-deferred-review | 待全文 |
| 2026-08-10 | LLM 后训练 | 2608.10090 CHORUS: Complementary Experts for High-Coverage Testbench Stimulus Generation | manual-review | 待全文 |
| 2026-08-10 | LLM 后训练 | 2608.09745 SR-OPSD: Self-Referenced On-Policy Self-Distillation | manual-review | B08 |
| 2026-08-10 | 搜广推与 LLM 应用 | 2608.09634 IntHQ: Task-Interactive Hierarchical Query on Dual-Stream Representations for Generative Recommendation | industrial-fulltext-review | B01 |
| 2026-08-10 | 搜广推与 LLM 应用 | 2608.09605 TSPORec: Token Selection via Preference Optimization for LLM-Based Sequential Recommendation | manual-review | 待全文 |
| 2026-08-10 | Agent | 2608.09380 OpenLoopEvolve: A Verifiable Self-Evolution Framework for Loop Policies in Long-Horizon Complex Tasks | manual-review | B11 |
| 2026-08-10 | LLM 后训练 | 2608.09217 Beyond Solvability: Task Learnability as a Static Prior for LLM RL Post-Training | manual-review | 待全文 |
| 2026-08-10 | Agent | 2608.09119 Motif 3: Technical Report | manual-review | 待全文 |
| 2026-08-09 | 基础模型 / LLM 后训练 | 2608.08878 DistillCache: KL-Guided Adaptive KV-Cache Eviction for Memory-Efficient LLM Inference | manual-review, p2-deferred-review | B07 |
| 2026-08-09 | 基础模型 / LLM 后训练 | 2608.08744 Can We Optimize the Performance-Carbon Emission Break-Even Point?: The Quest for Greener LLMs | manual-review, query-collision | 待全文 |
| 2026-08-08 | 基础模型 | 2608.08081 RotaryQuant: Fitting 120B MoE Models on Consumer Hardware via Fused Compressed-Space Attention | manual-review | 待全文 |
| 2026-08-08 | 搜广推与 LLM 应用 | 2608.07989 PushDualGen: Enabling LLMs to Generate Semantic IDs with Interpretable Copy for Industrial Push Recommendation | industrial-fulltext-review | B01 |
| 2026-08-07 | 基础模型 / 搜广推与 LLM 应用 | 2608.07816 Preserving Item Semantics for Free: Rethinking Token Initialization in LLM-Based Generative Recommendation | manual-review, p2-deferred-review | 待全文 |
| 2026-08-07 | 搜广推与 LLM 应用 | 2608.10008 Do LLM Recommenders Know When They're Hallucinating? Auditing Confidence Calibration in Catalog Faithfulness | manual-review | 待全文 |
| 2026-08-07 | Agent / LLM 后训练 | 2608.07418 ResidencyRL: Reinforcement Learning in Simulated Clinical Environments | manual-review, p2-deferred-review | 待全文 |
| 2026-08-07 | 基础模型 | 2608.07193 An AI4AI Framework for Visual Token Pruning | manual-review | 待全文 |
| 2026-08-07 | 搜广推与 LLM 应用 | 2608.07055 Teacher Retains Full Tokens, Student Merges Efficiently: TM20K for E-Commerce Sequence Modeling in Ad Recommendation | industrial-fulltext-review | 待全文 |
| 2026-08-07 | 基础模型 | 2608.06901 Prune Once: Retraining-Free Task-Agnostic Pruning for Vision-Language Models | manual-review | 待全文 |
| 2026-08-07 | 基础模型 | 2608.06849 Autonomy-of-Heads: Data-Free Sparse Attention from Frozen Query-Key Geometry | manual-review | B07 |
| 2026-08-07 | Agent | 2608.06811 Coupling Planning with Episodic Memory in LLM Agents for Software Issue Resolution | manual-review | B11 |
| 2026-08-07 | 搜广推与 LLM 应用 | 2608.06792 Progressive Alignment of Recommender Foundation Model through Multi-Phase Post-Training | industrial-fulltext-review | 待全文 |
| 2026-08-06 | LLM 后训练 | 2608.06526 GRASP: Reinforcing Language Model Anonymizers with Group Relative Policy Optimization | manual-review | 待全文 |
| 2026-08-06 | Agent | 2608.06474 WebGrader: Training LLMs for Web Development with Self-Evolving Programmatic Grader | manual-review | 待全文 |
| 2026-08-06 | 搜广推与 LLM 应用 | 2608.06068 Cleo: A Transparent and Controllable Chatbot for Conversational Commerce | manual-review | 待全文 |
| 2026-08-06 | LLM 后训练 | 2608.05802 On-Policy Delta Distillation for Multilingual Math Reasoning | manual-review | B08 |
| 2026-08-06 | Agent | 2608.05792 When Agentic AI Meets Integrated Sensing and Communication | manual-review | 待全文 |
| 2026-08-06 | 搜广推与 LLM 应用 | 2608.05655 Is Personalized Modality Weighting Actually Personalized? A Controlled Audit of Per-User Weighting Claims in Multimodal Recommenders | priority-fulltext-review | 待全文 |
| 2026-08-05 | 搜广推与 LLM 应用 | 2608.07593 Weather- and Location-Aware Agentic Dining Recommendation: Leveraging LLM World Knowledge for Region-Sensitive Contextual Reasoning | priority-fulltext-review | 待全文 |
| 2026-08-05 | 基础模型 | 2608.05000 Towards Physics of Multimodal Pretraining: Knowledge Flow, Modality Synergy, Early Unification, and Recipes | manual-review | B07 |
| 2026-08-05 | 搜广推与 LLM 应用 | 2608.04807 WatchLens: A Configurable Platform for Online Video Recommendation Experiments | industrial-fulltext-review | 待全文 |
| 2026-08-05 | 基础模型 | 2608.04678 Kathleen Writes: Autoregressive Generation and Data Scaling Without Attention | manual-review | 待全文 |
| 2026-08-05 | LLM 后训练 | 2608.04646 Evaluating Theory of Mind in Reasoning Models: Robustness over Reasoning | manual-review | 待全文 |
| 2026-08-04 | 基础模型 | 2608.03930 Logic Before Language: Pre-pretraining on Formal Derivations Fosters Skill Acquisition and Compressibility | manual-review | 待全文 |
| 2026-08-04 | 搜广推与 LLM 应用 | 2608.03899 ATLAS: Learning to Recommend Across Unseen Domains | manual-review | 待全文 |
| 2026-08-04 | 基础模型 | 2608.03855 Bi-semantic Chemical Embedder for Joint Representation Learning of SMILES and Natural Language | manual-review | 待全文 |
| 2026-08-04 | Agent | 2608.03800 Autoreflection: How Agentic Strange Loops Turn Human Culture into AI Infrastructure | manual-review | 待全文 |
| 2026-08-04 | 搜广推与 LLM 应用 | 2608.03722 When Outputs Disperse, Does Epistemic Revision Follow? A Black-Box Diagnostic for Machine Collectives | priority-fulltext-review | 待全文 |
| 2026-08-04 | LLM 后训练 | 2608.03673 CausalOPD: First-Wrong-Step Supervision for Distilling Causal Chain Reasoning | manual-review | B08 |
| 2026-08-04 | 搜广推与 LLM 应用 | 2608.03659 How Closely Do LLM Reviews Align with Human Peer Review? | priority-fulltext-review | 待全文 |
| 2026-08-04 | LLM 后训练 | 2608.03545 Hi-TTRL: Regulating Consensus with Hints for Test-Time Reinforcement Learning | manual-review | 待全文 |
| 2026-08-04 | Agent / 基础模型 / LLM 后训练 | 2608.03502 Hybrid LLM-Augmented Reinforcement Learning Agents for Complex Sequential Decision Tasks | manual-review, p2-deferred-review | 待全文 |
| 2026-08-04 | Agent | 2608.03468 ToolLIFT: Lifting Tool-Specific Trajectories into Function-Level Graphs for Generalizable Tool Planning | manual-review | B11 |
| 2026-08-04 | 基础模型 | 2608.11249 Diffuse to Compress: Leveraging Diffusion LMs for Lossless Compression | manual-review | 待全文 |
| 2026-08-04 | 搜广推与 LLM 应用 | 2608.03272 Attacking and Defending Multi-Agent Collaborative Filtering Systems Through Connectivity | manual-review | 待全文 |
| 2026-08-04 | LLM 后训练 | 2608.03204 Aligning Large Vision-Language Models at Test Time: A Trajectory-Guided Structured Sampling Approach | manual-review | 待全文 |
| 2026-08-04 | LLM 后训练 | 2608.03092 SMOPD: Multi-Reward Reinforcement Learning via Specialize-and-Merge Online Policy Distillation | manual-review | B08 |
| 2026-08-04 | LLM 后训练 | 2608.03077 PAMT: Process-Aligned Reinforcement Learning for Multi-Domain Machine Translation | manual-review | 待全文 |
| 2026-08-04 | 基础模型 / LLM 后训练 | 2608.03048 PI-Mem: Pushing Long-Context Reasoning to 3.6M Tokens with Parallel-Iterative Memory | manual-review, p2-deferred-review | 待全文 |
| 2026-08-03 | 基础模型 / LLM 后训练 | 2608.02867 BODHI: Do LLMs Branch Out and Discover Heterogeneous Inferences? | manual-review, p2-deferred-review | 待全文 |
| 2026-08-03 | LLM 后训练 | 2608.02831 Reinforcement Learning with Evolving Rubrics as Rewards for Audio Reasoning | manual-review | 待全文 |
| 2026-08-03 | 基础模型 | 2608.02560 Structured Memory for Edge Language Models: Persistent Context and Corpus Retrieval via O(1) SSM State Injection | manual-review | 待全文 |
| 2026-08-03 | Agent / LLM 后训练 | 2608.02391 Cooperative Coevolution for Resource-Constrained Agentic LLM Post-Training | manual-review, p2-deferred-review | 待全文 |
| 2026-08-03 | LLM 后训练 | 2608.02139 Self-Improving Large Language Models via Progressive Experience Evolution | manual-review | 待全文 |
| 2026-08-03 | 基础模型 / LLM 后训练 | 2608.02110 IACM-RL: Intent-Aware Context Management and Reinforcement Learning for Complex Tool Invocation under Dynamic Intent Fluctuations | manual-review, p2-deferred-review | 待全文 |
| 2026-08-03 | LLM 后训练 | 2608.02087 Instruction-Conditioned Exploration for Reinforcement Learning with Self-Distillation to an Unconditioned Policy | manual-review | 待全文 |
| 2026-08-03 | 基础模型 | 2608.02032 DART: Decoded Attention over Recurrent States for Efficient Long-Context Sequence Modeling | manual-review | B07 |
| 2026-08-03 | 基础模型 | 2608.02031 Learning-Based Collaborative MEC for LLM Inference with Soft-Deadline Awareness via Transformer-Enhanced PPO | manual-review | 待全文 |
| 2026-08-03 | LLM 后训练 | 2608.01804 LEAP: Lean Environment-Feedback via Adaptive Pruning for Code RL in GPU Kernel Generation | manual-review | 待全文 |
| 2026-08-03 | 搜广推与 LLM 应用 | 2608.01732 X-KGRank: A Knowledge Graph RAG Framework for Explainable Recommendations via Pattern Mining and LLM Re-Ranking | manual-review | 待全文 |
| 2026-08-03 | Agent | 2608.01719 MNC: Scope-Bound Semantic Declassification for Private LLM-Agent Communication | manual-review | 待全文 |
| 2026-08-03 | LLM 后训练 | 2608.01717 Beyond On-Policy Exploration: Integrating External Policy Rollouts for Reinforcement Learning in Diffusion Language Models | manual-review | B09 |
| 2026-08-03 | Agent | 2608.01684 GABench: A Comprehensive Benchmark for Evaluating LLM Agents on Graph Analysis Tasks | manual-review | 待全文 |
| 2026-08-03 | Agent / LLM 后训练 | 2608.01678 Progressive Agent Skill Generation via Reinforcement Learning | manual-review, p2-deferred-review | 待全文 |
| 2026-08-03 | 基础模型 | 2608.01676 Understanding Sparse Attention Selectivity in Long-Context Foundation Models via Counterfactual Evaluation | manual-review | 待全文 |
| 2026-08-03 | 基础模型 | 2608.01672 Learning What to Remember: Test-Time Training via Context Distillation | manual-review | B07 |
| 2026-08-03 | Agent | 2608.02683 \(S^3\): Improving Agent Safety through Multi-Stage Defense | manual-review | 待全文 |
| 2026-08-03 | Agent | 2608.01558 Securing Agentic AI: From Per-Action Checks to Trajectory Assurance | manual-review | 待全文 |
| 2026-08-02 | 基础模型 | 2608.01536 Celty: SpMspV GPU Kernel and SIMT Co-Design for Efficient Dual-Sparse LLM Inference | manual-review | 待全文 |
| 2026-08-02 | LLM 后训练 | 2608.01359 EviSD: Evidence-Conditioned Self-Distillation for Search-Augmented Agents | manual-review | 待全文 |
| 2026-08-02 | 搜广推与 LLM 应用 | 2608.00938 GRACE: Generative Recommender Acceleration Engine for Real-Time Ads Retrieval | manual-review | 待全文 |
| 2026-08-01 | 搜广推与 LLM 应用 | 2608.00816 Exponential Reward Weighting for Fine-Tuning Generative Recommenders under Sparse and Noisy Feedback | manual-review | 待全文 |
| 2026-08-01 | LLM 后训练 | 2608.00782 Distill Where You Fail: Recovering Learning Signals of Negative RL-Groups from Adaptive Teacher Guidance | manual-review | B08 |
| 2026-08-01 | Agent | 2608.00718 Adversarial Attacks in Multi-Agent LLM Pipelines: Unveiling Structural Vulnerabilities in Agentic AI Architectures | manual-review | 待全文 |
| 2026-08-01 | Agent | 2608.00558 AiFlow: Token-Native Reactive Orchestration with Bounded Backpressure for Streaming LLM Applications | manual-review | 待全文 |
| 2026-07-31 | Agent / LLM 后训练 | 2608.00335 RMSWeb: Reflection, Failure-Mode Mining, and Salvage-DS for Web Agent Reinforcement Learning | manual-review, p2-deferred-review | 待全文 |
| 2026-07-31 | Agent | 2608.02650 HyperAgent: Planning and Acting over Tool-Schema Hypergraphs for Tool-Use LLM Agents | manual-review | B11 |
| 2026-07-31 | LLM 后训练 | 2608.00220 Verifier-Induced Support Reshaping in On-Policy Optimization | manual-review | 待全文 |
| 2026-07-31 | Agent / LLM 后训练 | 2608.00175 Inference-Time Policy Alignment for Fair Reinforcement Learning | manual-review, p2-deferred-review | 待全文 |
| 2026-07-31 | 搜广推与 LLM 应用 | 2608.11241 RecSys Factory: Bounding LLM Agent Autonomy to Decision Points in the Industrial Recommender Lifecycle | manual-review | 待全文 |
| 2026-07-31 | 搜广推与 LLM 应用 | 2607.29241 RecHarness: A Bandit-Routed Agentic Harness for Self-Evolving Recommender Systems | industrial-fulltext-review | B02 |
| 2026-07-31 | 搜广推与 LLM 应用 | 2607.29213 GALA: Generative Aligned Learning for Adaptive Multimodal Representation in the Taobao Shangou Recommender System | industrial-fulltext-review | B02 |
| 2026-07-31 | LLM 后训练 | 2607.29185 Learning Latent Reasoning Traces for Scalar Reward Models End-to-End | manual-review | 待全文 |
| 2026-07-31 | Agent / 基础模型 | 2607.29032 TransMem: Transforming Hidden States into Memory for Large Language Models | manual-review, p2-deferred-review | B07 |
| 2026-07-31 | 搜广推与 LLM 应用 | 2607.28971 Don't Contrast the Impossible: Region-Constrained Batching for Contrastive User Modeling on a Local Community Platform | industrial-fulltext-review | 待全文 |
| 2026-07-31 | 搜广推与 LLM 应用 | 2607.28940 TransX: Scaling Transformer-based Recommendation via Behavioral and Serving Stream Crossings | industrial-fulltext-review | 待全文 |
| 2026-07-30 | 搜广推与 LLM 应用 | 2607.28895 LLM-Based Generative Retrieval for Snapchat Content Recommendation | industrial-fulltext-review | 待全文 |
| 2026-07-30 | Agent | 2607.28840 Benchmarks Are Not Validation: A System-Level View of Financial LLM Applications | manual-review | 待全文 |
| 2026-07-30 | Agent / 基础模型 / LLM 后训练 | 2607.28826 Distilling Knowledge from Large Language Models into Lightweight Reinforcement Learning Agents for Autonomous Cyber Operations | manual-review, p2-deferred-review | 待全文 |
| 2026-07-30 | Agent | 2607.28527 MANTA: Multi-Agent Network Topology Adaptation for Self-Evolving Multi-Agent Systems | manual-review | B11 |
| 2026-07-30 | LLM 后训练 | 2607.28457 SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute | manual-review | 待全文 |
| 2026-07-30 | LLM 后训练 | 2607.28301 HARGO: Heterogeneity-Aware Reward-Guided Optimization for RL Post-Training of LLMs on HPC Tasks | manual-review | 待全文 |
| 2026-07-30 | LLM 后训练 | 2607.28127 FinSMART: Financial Sentiment Analysis for Algorithmic Trading through Market-Aligned Reinforcement Learning | manual-review | 待全文 |
| 2026-07-30 | LLM 后训练 | 2607.28077 LEEPS: Latent-Guided Explore-Exploit Prompt Sampling for Efficient RLVR in Large Language Models | manual-review | 待全文 |
| 2026-07-30 | Agent / LLM 后训练 | 2607.28026 Contrastive Reinforced Policy Optimization via Privileged Self-Distillation | manual-review, p2-deferred-review | B09 |
| 2026-07-30 | 搜广推与 LLM 应用 | 2607.27944 Interpretable Representation via LLM-Driven Generative Disentanglement for Local-Life Service Recommendation | manual-review | 待全文 |
| 2026-07-30 | 搜广推与 LLM 应用 | 2607.27789 From Understanding to Action: Feedback-Grounded Policy Discovery for Generative Recommendation | industrial-fulltext-review | B02 |
| 2026-07-30 | 搜广推与 LLM 应用 | 2607.27760 Hierarchical Latent Reasoning for LLM-based Recommendation | manual-review | 待全文 |
| 2026-07-30 | 搜广推与 LLM 应用 | 2607.27647 LoopMemGR: From Behavior Logs to Evolving Memory for Generative Recommendation | manual-review | 待全文 |
| 2026-07-30 | LLM 后训练 | 2607.27610 Kalman Meets Curriculum: Efficient Dynamic Prompt Selection for Adaptive RL Finetuning | manual-review | 待全文 |
| 2026-07-30 | Agent | 2607.27562 DeepResearch Agent System | manual-review | 待全文 |
| 2026-07-29 | 搜广推与 LLM 应用 | 2607.27475 OneShot: Index-in-Ranking with Neural Scoring for Large-Scale Retrieval | industrial-fulltext-review | B00 |
| 2026-07-29 | 搜广推与 LLM 应用 | 2607.26893 Beyond Action Imitation: Learning a Decision-Aware User Simulator for Online Advertising | manual-review | 待全文 |
| 2026-07-29 | LLM 后训练 | 2607.26873 SERPO: Self-Evolving Rubric Policy Optimization for Open-Ended Test-Time Reinforcement Learning | manual-review | B09 |
| 2026-07-29 | Agent | 2607.26724 UrbanDS: A Graph-Guided LLM Multi-Agent System for Data-Intensive Urban Tasks | manual-review | 待全文 |
| 2026-07-29 | 搜广推与 LLM 应用 | 2607.26621 WhisperRec: Latent Reasoning for Efficient Foundation Recommendation Models | manual-review | 待全文 |
| 2026-07-29 | LLM 后训练 | 2608.14644 DUET: Dual-Teacher On-Policy Distillation via Same-Weight Disagreement for Prohibition Compliance | manual-review | 待全文 |
| 2026-07-29 | 搜广推与 LLM 应用 | 2607.26500 Multi-Decoder OneRec: Controllable Generative Retrieval for Multi-Objective Industrial Recommendation | industrial-fulltext-review | 待全文 |
| 2026-07-29 | LLM 后训练 | 2607.26457 DHRCL:Training Code LLMs with Dense Hierarchical Rewards and Curriculum Learning | manual-review | 待全文 |
| 2026-07-29 | 搜广推与 LLM 应用 | 2607.26427 PSG: Pair-Space Generation for Efficient Generative Reranking | manual-review | 待全文 |
| 2026-07-29 | 搜广推与 LLM 应用 | 2607.26418 DIRECTOR: Dynamic Index-based Recommendation with Transport-Optimized Retrieval | industrial-fulltext-review | 待全文 |
| 2026-07-29 | 搜广推与 LLM 应用 | 2607.26369 ClockRoPE: Random Fourier Rotations for Temporal Routine Modeling | industrial-fulltext-review | B00 |
| 2026-07-29 | Agent / LLM 后训练 | 2607.26358 Post-Training at the Edge of Detectability: A Game-Theoretic Approach to Fine-Tuning | manual-review, p2-deferred-review | 待全文 |
| 2026-07-28 | 搜广推与 LLM 应用 | 2607.25823 Hypothesis-Driven Shelf Generation for Personalised Recommendation | manual-review | 待全文 |
| 2026-07-28 | 搜广推与 LLM 应用 | 2607.25640 LLM-as-a-Judge for Evaluating System Responses in Conversational Music Recommendation | manual-review | 待全文 |
| 2026-07-28 | 搜广推与 LLM 应用 | 2607.25346 The Case Against Generation for Retrieval: Discriminative Language Models as Effective Retrievers | manual-review | 待全文 |
| 2026-07-28 | 搜广推与 LLM 应用 | 2607.25344 Reward Guided Decoding for Generative Recommendation | industrial-fulltext-review | 待全文 |
| 2026-07-28 | 搜广推与 LLM 应用 | 2607.25339 SPARC: Sequence-aware Progressive Attribute Routing and Compression Framework for Generative Recommendation | manual-review | 待全文 |
| 2026-07-28 | 搜广推与 LLM 应用 | 2607.25276 FunnelAL: Retrieve-then-Rank Active Learning for Single-Class Discovery | manual-review | 待全文 |
| 2026-07-28 | LLM 后训练 | 2607.26094 Meta-Learned Reward Shaping for Reinforcement Learning from Human Feedback | manual-review | 待全文 |
| 2026-07-27 | 搜广推与 LLM 应用 | 2607.25110 Memory Layer: Train the In-Model Cache for Recommendation Models | industrial-fulltext-review | 待全文 |
| 2026-07-27 | Agent / LLM 后训练 | 2607.25091 Towards Robust Reinforcement Learning for Small-Scale Language Model Agents | manual-review, p2-deferred-review | 待全文 |
| 2026-07-27 | LLM 后训练 | 2607.24900 Inverse RL Helps Align AI by Imitating Humans | manual-review | 待全文 |
| 2026-07-27 | LLM 后训练 | 2607.24522 FlowCTS: On-policy Continuous Trajectory Supervision of Flow Models | manual-review | 待全文 |
| 2026-07-27 | 搜广推与 LLM 应用 | 2607.24213 Integrating Factual and Normative Industrial Knowledge via Constraint-Aware Graph Attention for Process Plan Recommendation | manual-review | 待全文 |
| 2026-07-27 | 搜广推与 LLM 应用 | 2607.24092 ConAlign: Conditional Alignment Framework for Balancing Biased and Unbiased Recommendation | industrial-fulltext-review | 待全文 |
| 2026-07-27 | 基础模型 / 搜广推与 LLM 应用 | 2607.24025 SpecFormer: Mitigating Embedding and Attention Collapse via Spectral-Aware Transformer for Recommendation | industrial-fulltext-review, p2-deferred-review | 待全文 |
| 2026-07-27 | 搜广推与 LLM 应用 | 2607.23986 MEMOIR: Temporal Behavioral Memory for Recommendation Across the Preference-Drift Spectrum | manual-review | 待全文 |
| 2026-07-26 | 搜广推与 LLM 应用 | 2607.24869 Ranked by Position: Order Sensitivity as an Exploitable Attack Surface in LLM Listwise Recommenders | manual-review | 待全文 |
| 2026-07-26 | Agent | 2607.23678 Focus Is All You Need: Adaptive Goal-aware Attention Orchestration for Multi-Agent Graph Systems | manual-review | 待全文 |
| 2026-07-26 | 搜广推与 LLM 应用 | 2607.23561 Towards a Relevance Posterior in Neural Information Access | manual-review | 待全文 |
| 2026-07-26 | LLM 后训练 | 2607.23488 Learning Sampling Parameters for Diffusion Models | manual-review | 待全文 |
| 2026-07-26 | LLM 后训练 | 2607.23420 LA-RL: Label-Aware Self-Reflection for Reinforcement Learning in Information Extraction | manual-review | 待全文 |
| 2026-07-25 | LLM 后训练 | 2607.23125 Self-Boosting Vision-Language Models with Noisy Student On-Policy Self-Distillation | manual-review | 待全文 |
| 2026-07-25 | 搜广推与 LLM 应用 | 2607.23038 EGR: Embedding-Native Generative Retrieval with a Shared LLM | manual-review | 待全文 |
| 2026-07-24 | 搜广推与 LLM 应用 | 2607.24846 Two Views, One Voice: Evidence-Grounded Conversational Music Recommendation | manual-review | 待全文 |
| 2026-07-24 | LLM 后训练 | 2608.07531 Search-G1: Grounded Search Agents via Representation-Based Intrinsic Rewards | manual-review | 待全文 |
| 2026-07-24 | 基础模型 / LLM 后训练 | 2607.21971 Teaching LLMs to Self-Evolve: Cultivating Core Meta-Skills with Reinforcement Learning | manual-review, p2-deferred-review | 待全文 |
| 2026-07-23 | 搜广推与 LLM 应用 | 2607.21519 Diffusion Language Model for Recommendation | manual-review | 待全文 |
| 2026-07-23 | LLM 后训练 | 2607.21090 Training Large Language Models for Self-Explanation Faithfulness | manual-review | 待全文 |
| 2026-07-23 | LLM 后训练 | 2607.20952 The Weight of Silence: A Causal Case for Weights Over the Scratchpad in Latent Chess Reasoning | manual-review | 待全文 |
| 2026-07-23 | 搜广推与 LLM 应用 | 2607.24829 Improving Rare Medication Recommendation with Counterfactual Data Augmentation and Large Language Models | manual-review | 待全文 |
| 2026-07-22 | LLM 后训练 | 2607.20090 Reinforcement Learning for Large Language Model Selective Evidence Adoption from Contaminated Retrieval Results | manual-review | 待全文 |
| 2026-07-21 | LLM 后训练 | 2607.19331 ISO: An RLVR-Native Optimization Stack | manual-review | B09 |
| 2026-07-21 | 搜广推与 LLM 应用 | 2608.11219 From Monolithic to Modular: Segment-level Automatic Prompt Optimization | priority-fulltext-review | 待全文 |
| 2026-07-21 | LLM 后训练 | 2607.19219 Beyond Score Prediction: LLM-Based Essay Scoring and Feedback Generation via Reinforcement Learning with Rubric Rewards | manual-review | 待全文 |
| 2026-07-21 | Agent / 基础模型 / LLM 后训练 | 2607.19450 REGEN: Replay-recycling for Expert-to-Generalist distillation with Offline Reinforcement Learning | manual-review, p2-deferred-review | 待全文 |
| 2026-07-21 | Agent / 基础模型 | 2608.07527 DocAtlas: Long-Document Understanding as Mutable-State Interaction | manual-review, p2-deferred-review | 待全文 |
| 2026-07-20 | Agent | 2607.18485 Trusted Credentials, Untrusted Behavior: Benchmarking LLM-Agent Security in High-Performance Computing | manual-review | 待全文 |
| 2026-07-20 | Agent | 2607.18147 LLMs and Agentic AI Systems for Smart Grids: A Tutorial on Architectures and Applications | manual-review | 待全文 |
| 2026-07-20 | 基础模型 | 2607.17715 C\(^2\)KV: Compressed and Composable KV Cache Reuse for Efficient LLM Inference | manual-review | B07 |
| 2026-07-19 | Agent | 2607.17038 Reward-Driven LLM Agent Workflows: Synthesizing POMDP Routing and Self-Correction for Autonomous Decision-Making | manual-review | 待全文 |
| 2026-07-18 | 搜广推与 LLM 应用 | 2607.16633 Beyond Fixed Depths and Widths: Optimizing Textual Decoding Tries in LLM-based Generative Recommendation | manual-review | 待全文 |
| 2026-07-16 | 搜广推与 LLM 应用 | 2607.15440 Stochastic Reset Pathfinding: Path-Level Regret for Cascading Bandits over Graph Paths | priority-fulltext-review | 待全文 |
| 2026-07-16 | Agent | 2607.15257 SearchOS-V1: Towards Robust Open-Domain Information-Seeking Agent Collaboration | manual-review | 待全文 |
| 2026-07-16 | 搜广推与 LLM 应用 | 2607.14835 LLM-Based Re-Ranking for Real Estate Search | industrial-fulltext-review | B02 |
| 2026-07-16 | 搜广推与 LLM 应用 | 2607.14604 Accelerating A/B-Tests with Counterfactual Estimation: Reducing Variance through Policy Overlap | priority-fulltext-review | 待全文 |
| 2026-07-15 | LLM 后训练 | 2607.13753 Post-Training Shifts Confidence: A Three-Stage Analysis of How SFT, RL, and OPD Shape CoT Calibration | manual-review | 待全文 |
| 2026-07-13 | 搜广推与 LLM 应用 | 2607.26073 Guess Where You Go: Generative Next Point-of-Interest Recommendation in Amap | industrial-fulltext-review | B02 |
| 2026-07-12 | 搜广推与 LLM 应用 | 2607.10910 ZoRRO: A Zero-Weight Personalized Recommender System for Scalable News Recommendation | industrial-fulltext-review | 待全文 |
| 2026-07-11 | 搜广推与 LLM 应用 | 2607.10239 Multilingual Semantic Retrieval for Apple Music Search | industrial-fulltext-review | 待全文 |
| 2026-07-11 | 搜广推与 LLM 应用 | 2607.10235 Consensus vs. Dissent: Dynamic LLM Modeling of Subjective Preferences in Group Recommenders | manual-review | 待全文 |
| 2026-07-10 | 搜广推与 LLM 应用 | 2607.10016 Tokenizing Numerical and Embedding Features for LLM RecSys | manual-review | 待全文 |
| 2026-07-10 | 搜广推与 LLM 应用 | 2607.09988 An LLM-powered Agentic Recommendation System for Connected TV Content Discovery | manual-review | 待全文 |
| 2026-07-09 | 搜广推与 LLM 应用 | 2607.08703 MPFlow: Learning Budgeted Max-Flow Optimization on the Lightning Network with Deep Graph Reinforcement Learning | industrial-fulltext-review | 待全文 |
| 2026-07-09 | Agent / 基础模型 | 2607.08497 Cognitive-structured Multimodal Agent for Multimodal Understanding, Generation, and Editing | manual-review, p2-deferred-review | 待全文 |
| 2026-07-09 | 搜广推与 LLM 应用 | 2607.08365 DaV-Gen: End-to-End Generative Retrieval via Draft-and-Verify | manual-review | 待全文 |
| 2026-07-08 | 搜广推与 LLM 应用 | 2607.06979 Robust Federated Learning Under Real-World Client Churn | priority-fulltext-review | 待全文 |
| 2026-07-08 | 搜广推与 LLM 应用 | 2607.06963 Large Language Models (LLMs) and Generative AI in Cybersecurity and Privacy: A Survey of Dual-Use Risks, AI-Generated Malware, Explainability, and Defensive Strategies | priority-fulltext-review | 待全文 |
| 2026-07-05 | 搜广推与 LLM 应用 | 2607.04270 LBR: Towards Mitigating Length Bias in Large Language Models for Recommendation | manual-review | 待全文 |
| 2026-07-03 | 基础模型 | 2607.03089 STELLA: Efficient Sensor-to-LLM Translation for On-Device Human Activity Recognition | manual-review | 待全文 |
| 2026-07-01 | 搜广推与 LLM 应用 | 2607.01170 Diffusion-GR2: Diffusion Generative Reasoning Re-ranker | manual-review | 待全文 |
| 2026-06-30 | 搜广推与 LLM 应用 | 2606.31984 GR2 Technical Report | manual-review | 待全文 |
| 2026-06-30 | Agent / 搜广推与 LLM 应用 | 2606.31693 ShopX: A Foundation Model for Intent-to-Item Fulfillment in Agentic Shopping | manual-review, p2-deferred-review | 待全文 |
| 2026-06-30 | 搜广推与 LLM 应用 | 2606.31031 GenPage: Towards End-to-End Generative Homepage Construction at Netflix | industrial-fulltext-review | B02 |
| 2026-06-29 | 搜广推与 LLM 应用 | 2606.29946 POEM: Partial-Order Enhanced Real-Time Sequential Modeling for Recommendation | industrial-fulltext-review | 待全文 |
| 2026-06-28 | 搜广推与 LLM 应用 | 2606.29341 Monosemanticity in Recommender Systems | manual-review | 待全文 |
| 2026-06-27 | 基础模型 / 搜广推与 LLM 应用 | 2606.28933 FinInvest-GTCN: Explainable Graph-Temporal-Causal Modeling for Risk-Aware Investment Decision Optimization | p2-deferred-review, priority-fulltext-review | 待全文 |
| 2026-06-27 | 搜广推与 LLM 应用 | 2607.24789 NEXT: Reasoning-Driven Video Recommendation via a Vision-Language Model | manual-review | B00 |
| 2026-06-26 | 搜广推与 LLM 应用 | 2606.28059 Fast and Feasible: Permutation-based Constrained Reranking for Revenue Maximization | industrial-fulltext-review | 待全文 |
| 2026-06-25 | 搜广推与 LLM 应用 | 2606.27058 UniFormer: Efficient and Unified Model-Centric Scaling for Industrial Recommendation | industrial-fulltext-review | 待全文 |
| 2026-06-25 | 搜广推与 LLM 应用 | 2606.26859 AgentX: Towards Agent-Driven Self-Iteration of Industrial Recommender Systems | manual-review | B00 |
| 2026-06-25 | LLM 后训练 | 2606.26787 AIGP: An LLM-Based Framework for Long-Term Value Alignment in E-Commerce Pricing | manual-review | 待全文 |
| 2026-06-24 | 搜广推与 LLM 应用 | 2606.25496 Recommendation as Generation: Unifying Personalized Video Generation and Recommendation at Industrial Scale | industrial-fulltext-review | 待全文 |
| 2026-06-22 | 搜广推与 LLM 应用 | 2606.23911 Scaling Dense Retrieval with LLM-Annotated Training Data: Structured Mining and Progressive Curriculum for E-Commerce Sponsored Search | industrial-fulltext-review | 待全文 |
| 2026-06-22 | 搜广推与 LLM 应用 | 2606.23057 Who Owns the AI Recommendation? A Multi-Industry Empirical Map of Brand Category Ownership Across Large Language Models | priority-fulltext-review | 待全文 |
| 2026-06-19 | 搜广推与 LLM 应用 | 2606.21590 Radial Basis Function Networks as Projection Heads in Self-Supervised Learning | priority-fulltext-review | 待全文 |
| 2026-06-17 | 搜广推与 LLM 应用 | 2606.19108 JourneyFormer: Encoding Airbnb Guest Journey with Sequence Modeling | industrial-fulltext-review | B03 |
| 2026-06-17 | 搜广推与 LLM 应用 | 2606.18814 LensKit-Auto | priority-fulltext-review | 待全文 |
| 2026-06-17 | 搜广推与 LLM 应用 | 2606.18750 Ensuring Trustworthy Online A/B Testing: Addressing Five Key Questions on CUPED | industrial-fulltext-review | 待全文 |
| 2026-06-16 | 搜广推与 LLM 应用 | 2606.18451 A Cross-Model VLM-Judge Protocol for Single-Image 3D Mesh Quality (and Why Cheap Proxies Fall Short) | priority-fulltext-review | 待全文 |
| 2026-06-15 | 搜广推与 LLM 应用 | 2606.16838 OneRank: Unified Transformer-Native Ranking Architecture for Multi-Task Recommendation | industrial-fulltext-review | 待全文 |
| 2026-06-15 | 搜广推与 LLM 应用 | 2606.16641 PIANO: Personalized Reranking via Information Aggregation Node for Music Search Optimization | industrial-fulltext-review | 待全文 |
| 2026-06-12 | 搜广推与 LLM 应用 | 2606.14932 Retrieval-as-a-Service:A System-Oriented Analysis of Industrial Retrieval Pipelines in Web Systems | manual-review | 待全文 |
| 2026-06-12 | 搜广推与 LLM 应用 | 2606.14817 Combining Retrieval-Augmented Text Generation with LLMs for Reading Content Recommendations | priority-fulltext-review | 待全文 |
| 2026-06-11 | LLM 后训练 | 2606.14801 QPILOTS: Efficient Test-Time Q-Steering for Flow Policies | manual-review | 待全文 |
| 2026-06-11 | 搜广推与 LLM 应用 | 2606.12984 SkillChain: Closing the Loop on Skill Evolution for Image-Based E-Commerce AI Assistants | industrial-fulltext-review | 待全文 |
| 2026-06-10 | 搜广推与 LLM 应用 | 2606.12281 CCKS: Consensus-based Communication and Knowledge Sharing | priority-fulltext-review | 待全文 |
| 2026-06-10 | 搜广推与 LLM 应用 | 2606.12198 LLM-Based User Personas for Recommendations at Scale | industrial-fulltext-review | 待全文 |
| 2026-06-09 | 搜广推与 LLM 应用 | 2606.10907 From Prompt to Purchase: How AI Brand Recommendations Move Consumers on the Open Web | priority-fulltext-review | 待全文 |
| 2026-06-09 | 搜广推与 LLM 应用 | 2606.10357 Atomic Intent Reasoning: Bringing LLM Semantics to Industrial Cross-Domain Recommendations | industrial-fulltext-review | 待全文 |
| 2026-06-08 | 搜广推与 LLM 应用 | 2606.10243 DUET -- Dual User Embedding Transformers for Offsite Conversion Prediction | industrial-fulltext-review | 待全文 |
| 2026-06-07 | 搜广推与 LLM 应用 | 2606.08604 Gryphon: A Unified Architecture for Semantic-ID Generation and Item-Level Scoring in Industrial Recommendations | industrial-fulltext-review | 待全文 |
| 2026-06-07 | LLM 后训练 / 搜广推与 LLM 应用 | 2606.08480 Adaptive Loss Balancing for Noise-Robust GRPO in Generative Recommendation | industrial-fulltext-review, p2-deferred-review | 待全文 |
| 2026-06-07 | 搜广推与 LLM 应用 | 2606.08466 ToolRec: Calibrated Preference Alignment for Query Recommendation in On-Device Assistants | industrial-fulltext-review | 待全文 |
| 2026-06-05 | 搜广推与 LLM 应用 | 2606.07068 Bias in Filter Feature Selection Evaluation: A Meta-Analysis of Datasets, Baselines, and Experimental Design Choices | priority-fulltext-review | 待全文 |
| 2026-06-05 | 搜广推与 LLM 应用 | 2606.06970 SSRLive: Live Streaming Recommendation with Dynamic Semantic ID | industrial-fulltext-review | 待全文 |
| 2026-06-03 | 搜广推与 LLM 应用 | 2606.04448 Bridging Short Videos and Live Streams: Reasoning-Guided Multimodal LLMs for Cross-Domain Representation Learning | industrial-fulltext-review | 待全文 |
| 2026-06-03 | 搜广推与 LLM 应用 | 2606.04387 Rethinking Sales Lead Scoring with LLM-based Hierarchical Preference Ranking | industrial-fulltext-review | 待全文 |
| 2026-06-02 | 搜广推与 LLM 应用 | 2606.04110 Variance Reduction for Heavy-Tailed Monetization Metrics in Ranking Experiments via Post-Stratification | industrial-fulltext-review | 待全文 |
| 2026-06-02 | LLM 后训练 / 搜广推与 LLM 应用 | 2606.03866 Taiji: Pareto Optimal Policy Optimization with Semantics-IDs Trade-off for Industrial LLM-Enhanced Recommendation | industrial-fulltext-review, p2-deferred-review | 待全文 |
| 2026-05-29 | 搜广推与 LLM 应用 | 2606.00282 Synthetic Data from Cross-Domain Events for Large-Scale Recommendation Systems | industrial-fulltext-review | 待全文 |
| 2026-05-27 | 搜广推与 LLM 应用 | 2605.29141 Toward User Preference Alignment in LLM Recommendation via Explicit Context Feedback | priority-fulltext-review | 待全文 |
| 2026-05-27 | 搜广推与 LLM 应用 | 2605.28888 Generative Spatiotemporal Intent Sequence Recommendation via Implicit Reasoning in Amap | industrial-fulltext-review | 待全文 |
| 2026-05-26 | 搜广推与 LLM 应用 | 2605.27704 Joint Optimization of Relevance and Engagement in Multi-Task Ranking for E-Commerce with Efficient LLM Supervision | industrial-fulltext-review | 待全文 |
| 2026-05-26 | 搜广推与 LLM 应用 | 2607.19357 Stochastic Primal-Dual Decoding for Multiobjective Generative Recommender Systems | industrial-fulltext-review | 待全文 |
| 2026-05-26 | 搜广推与 LLM 应用 | 2605.27103 MuChator: Enabling Active Music Discovery via Conversational Music LLMs in Douyin Music | industrial-fulltext-review | 待全文 |
| 2026-05-26 | 搜广推与 LLM 应用 | 2605.27043 Causal Representation Learning for Generalisable Recommendation | industrial-fulltext-review | 待全文 |
| 2026-05-26 | 搜广推与 LLM 应用 | 2605.26717 L2Rec: Towards Dual-View Understanding of LLMs for Personalized Recommendation | industrial-fulltext-review | B03 |
| 2026-05-26 | 搜广推与 LLM 应用 | 2605.26424 Uniboost: Global Coordination with Value Alignment for Fair and Efficient Traffic Allocation | industrial-fulltext-review | 待全文 |
| 2026-05-25 | 搜广推与 LLM 应用 | 2605.25726 SIREN: Unified Multi-Granularity Semantic Interaction for Multi-Modal Lifelong User Interest Modeling | industrial-fulltext-review | 待全文 |
| 2026-05-25 | 搜广推与 LLM 应用 | 2605.25514 From Item-Only to Query-Item: Query-Conditioned Generative Search with QGS in Quark | industrial-fulltext-review | B03 |
| 2026-05-24 | 搜广推与 LLM 应用 | 2605.25007 Meta-Modal Agent: Sequential Evidence Routing for Missing-Modality Candidate Reranking | priority-fulltext-review | 待全文 |
| 2026-05-22 | 搜广推与 LLM 应用 | 2605.23702 TubiFM: Unified Item, Carousel, and Search Ranking for Streaming Discovery | industrial-fulltext-review | B03 |
| 2026-05-20 | 搜广推与 LLM 应用 | 2605.21752 PEARL: Unbiased Percentile Estimation via Contrastive Learning for Industrial-Scale Livestream Recommendation | industrial-fulltext-review | B03 |
| 2026-05-19 | 搜广推与 LLM 应用 | 2605.20559 Group-Aware Matrix Estimation and Latent Subspace Recovery | priority-fulltext-review | 待全文 |
| 2026-05-18 | 搜广推与 LLM 应用 | 2605.18696 Ensembling Tabular Foundation Models - A Diversity Ceiling And A Calibration Trap | priority-fulltext-review | 待全文 |
| 2026-05-18 | 搜广推与 LLM 应用 | 2605.17863 DADF: A Distribution-Aware Debiasing Framework for Watch-Time Regression in Recommender Systems | industrial-fulltext-review | B03 |
| 2026-05-15 | 搜广推与 LLM 应用 | 2605.16479 Policy-Grounded Dynamic Facet Suggestions for Job Search | industrial-fulltext-review | 待全文 |
| 2026-05-13 | 搜广推与 LLM 应用 | 2605.13052 RAG-Enhanced Large Language Models for Dynamic Content Expiration Prediction in Web Search | industrial-fulltext-review | 待全文 |
| 2026-05-12 | 搜广推与 LLM 应用 | 2605.11732 AgentDisCo: Towards Disentanglement and Collaboration in Open-ended Deep Research Agents | priority-fulltext-review | 待全文 |
| 2026-05-11 | 搜广推与 LLM 应用 | 2605.11118 A Cascaded Generative Approach for e-Commerce Recommendations | industrial-fulltext-review | 待全文 |
| 2026-05-11 | 搜广推与 LLM 应用 | 2605.10367 AgentGR: Semantic-aware Agentic Group Decision-Making Simulator for Group Recommendation | priority-fulltext-review | 待全文 |
| 2026-05-09 | 搜广推与 LLM 应用 | 2605.08731 Choosing a JPEG Decoder for PyTorch DataLoaders: Workload-Specific Throughput on Four CPUs | priority-fulltext-review | 待全文 |
| 2026-05-08 | 搜广推与 LLM 应用 | 2605.07129 RRCM: Ranking-Driven Retrieval over Collaborative and Meta Memories for LLM Recommendation | priority-fulltext-review | 待全文 |
| 2026-05-08 | 搜广推与 LLM 应用 | 2605.16344 A Production-Ready RL Framework for Personalized Utility Tuning with Pareto Sweeping in Pinterest Recommender Systems | industrial-fulltext-review | B04 |
| 2026-05-07 | 搜广推与 LLM 应用 | 2605.06981 Bridging Textual Profiles and Latent User Embeddings for Personalization | priority-fulltext-review | 待全文 |
| 2026-05-07 | 搜广推与 LLM 应用 | 2605.05855 Bridging Passive and Active: Enhancing Conversation Starter Recommendation via Active Expression Modeling | industrial-fulltext-review | 待全文 |
| 2026-05-07 | 搜广推与 LLM 应用 | 2605.05730 Effective Knowledge Transfer for Multi-Task Recommendation Models | industrial-fulltext-review | B04 |
| 2026-05-04 | 搜广推与 LLM 应用 | 2606.07546 Beyond Item IDs: Scaling Short-Form-Video Recommendation via Semantic-Native Long Sequence Modeling | industrial-fulltext-review | 待全文 |
| 2026-05-01 | 搜广推与 LLM 应用 | 2605.00353 Negative Data Mining for Contrastive Learning in Dense Retrieval at IKEA.com | industrial-fulltext-review | 待全文 |
| 2026-04-30 | 搜广推与 LLM 应用 | 2605.00068 Human-in-the-Loop Meta Bayesian Optimization for Fusion Energy and Scientific Applications | priority-fulltext-review | 待全文 |
| 2026-04-30 | 搜广推与 LLM 应用 | 2604.27321 Toward Autonomous SOC Operations: End-to-End LLM Framework for Threat Detection, Query Generation, and Resolution in Security Operations | priority-fulltext-review | 待全文 |
| 2026-04-29 | 搜广推与 LLM 应用 | 2604.26390 Meta-Learning and Targeted Differential Privacy to Improve the Accuracy-Privacy Trade-off in Recommendations | priority-fulltext-review | 待全文 |
| 2026-04-28 | 搜广推与 LLM 应用 | 2604.25834 Action-Aware Generative Sequence Modeling for Short Video Recommendation | industrial-fulltext-review | 待全文 |
| 2026-04-28 | 搜广推与 LLM 应用 | 2604.25732 Personalized Multi-Interest Modeling for Cross-Domain Recommendation to Cold-Start Users | priority-fulltext-review | 待全文 |
| 2026-04-26 | 搜广推与 LLM 应用 | 2604.23522 Beyond Static Collision Handling: Adaptive Semantic ID Learning for Multimodal Recommendation at Industrial Scale | industrial-fulltext-review | B04 |
| 2026-04-25 | 搜广推与 LLM 应用 | 2604.23088 Code Broker: A Multi-Agent System for Automated Code Quality Assessment | priority-fulltext-review | 待全文 |
| 2026-04-20 | 搜广推与 LLM 应用 | 2604.18146 Modular Representation Compression: Adapting LLMs for Efficient and Effective Recommendations | industrial-fulltext-review | 待全文 |
| 2026-04-20 | 搜广推与 LLM 应用 | 2604.17878 RankUp: Towards High-rank Representations for Large Scale Advertising Recommender Systems | industrial-fulltext-review | 待全文 |
| 2026-04-17 | 搜广推与 LLM 应用 | 2604.15937 Polarization by Default: Auditing Recommendation Bias in LLM-Based Content Curation | priority-fulltext-review | 待全文 |
| 2026-04-15 | 搜广推与 LLM 应用 | 2604.14352 PROXIMA: A Reliability Scoring Framework for Proxy Metrics in Online Controlled Experiments | industrial-fulltext-review | 待全文 |
| 2026-04-15 | 搜广推与 LLM 应用 | 2604.13796 Driving Engagement in Daily Fantasy Sports with a Scalable and Urgency-Aware Ranking Engine | industrial-fulltext-review | 待全文 |
| 2026-04-14 | 搜广推与 LLM 应用 | 2604.12965 Efficient Retrieval Scaling with Hierarchical Indexing for Large Scale Recommendation | priority-fulltext-review | 待全文 |
| 2026-04-14 | 搜广推与 LLM 应用 | 2604.14223 TRACE: A Conversational Framework for Sustainable Tourism Recommendation with Agentic Counterfactual Explanations | priority-fulltext-review | 待全文 |
| 2026-04-14 | 搜广推与 LLM 应用 | 2605.04076 A Regulatory Governance Framework for AI-Driven Financial Fraud Detection in U.S. Banking: Integrating OCC, SR 11-7, CFPB, and FinCEN Compliance Requirements for Model Development, Validation, and Monitoring Lifecycles | priority-fulltext-review | 待全文 |
| 2026-04-14 | 搜广推与 LLM 应用 | 2604.12234 UniRec: Bridging the Expressive Gap between Generative and Discriminative Recommendation via Chain-of-Attribute | industrial-fulltext-review | B04 |
| 2026-04-13 | 搜广推与 LLM 应用 | 2604.11440 R3-VAE: Reference Vector-Guided Rating Residual Quantization VAE for Generative Recommendation | industrial-fulltext-review | 待全文 |
| 2026-04-12 | 搜广推与 LLM 应用 | 2604.10471 SID-Coord: Coordinating Semantic IDs for ID-based Ranking in Short-Video Search | industrial-fulltext-review | 待全文 |
| 2026-04-09 | 搜广推与 LLM 应用 | 2604.08181 Long-Term Embeddings for Balanced Personalization | industrial-fulltext-review | 待全文 |
| 2026-04-08 | 搜广推与 LLM 应用 | 2605.27374 ICG: Improving Cover Image Generation via MLLM-based Prompting and Personalized Preference Alignment | priority-fulltext-review | 待全文 |
| 2026-04-06 | 搜广推与 LLM 应用 | 2604.09698 Evaluating Scene-based In-Situ Item Labeling for Immersive Conversational Recommendation | priority-fulltext-review | 待全文 |
| 2026-04-05 | 搜广推与 LLM 应用 | 2604.03949 Semantic IDs for Recommender Systems at Snapchat: Use Cases, Technical Challenges, and Design Choices | industrial-fulltext-review | 待全文 |
| 2026-03-30 | 搜广推与 LLM 应用 | 2603.28124 RCLRec: Reverse Curriculum Learning for Modeling Sparse Conversions in Generative Recommendation | industrial-fulltext-review | 待全文 |
| 2026-03-27 | 搜广推与 LLM 应用 | 2603.26085 AgenticRS-Architecture: System Design for Agentic Recommender Systems | industrial-fulltext-review | 待全文 |
| 2026-03-26 | 搜广推与 LLM 应用 | 2603.25070 An Explainable Ensemble Learning Framework for Crop Classification with Optimized Feature Pyramids and Deep Networks | priority-fulltext-review | 待全文 |
| 2026-03-26 | 搜广推与 LLM 应用 | 2603.24963 Design Once, Deploy at Scale: Template-Driven ML Development for Large Model Ecosystems | priority-fulltext-review | 待全文 |
| 2026-03-25 | 搜广推与 LLM 应用 | 2603.24226 UniScale: Synergistic Entire Space Data and Model Scaling for Search Ranking | industrial-fulltext-review | B04 |
| 2026-03-24 | 搜广推与 LLM 应用 | 2603.22916 GateSID: Adaptive Gating for Balancing Semantic and Collaborative Signals in Recommendation | industrial-fulltext-review | B04 |
| 2026-03-23 | 搜广推与 LLM 应用 | 2603.21481 TagLLM: A Fine-Grained Tag Generation Approach for Note Recommendation | industrial-fulltext-review | 待全文 |
| 2026-03-20 | 搜广推与 LLM 应用 | 2603.20062 The End of Rented Discovery: How AI Search Redistributes Power Between Hotels and Intermediaries | priority-fulltext-review | 待全文 |
| 2026-03-20 | 搜广推与 LLM 应用 | 2603.19710 AIGQ: An End-to-End Hybrid Generative Architecture for E-commerce Query Recommendation | industrial-fulltext-review | B05 |
| 2026-03-20 | 搜广推与 LLM 应用 | 2603.19665 GenFacet: End-to-End Generative Faceted Search via Multi-Task Preference Alignment in E-Commerce | industrial-fulltext-review | 待全文 |
| 2026-03-20 | 搜广推与 LLM 应用 | 2603.19585 SaFRO: Satisfaction-Aware Fusion via Dual-Relative Policy Optimization for Short-Video Search | industrial-fulltext-review | B05 |
| 2026-03-19 | 搜广推与 LLM 应用 | 2603.18765 Implicit Grading Bias in Large Language Models: How Writing Style Affects Automated Assessment Across Math, Programming, and Essay Tasks | priority-fulltext-review | 待全文 |
| 2026-03-17 | 搜广推与 LLM 应用 | 2604.13057 A Multi-Model Approach to English-Bangla Sentiment Classification of Government Mobile Banking App Reviews | priority-fulltext-review | 待全文 |
| 2026-03-17 | 搜广推与 LLM 应用 | 2603.16088 RecBundle: A Next-Generation Geometric Paradigm for Explainable Recommender Systems | priority-fulltext-review | 待全文 |
| 2026-03-12 | 搜广推与 LLM 应用 | 2603.11486 Quantized Inference for OneRec-V2 | industrial-fulltext-review | 待全文 |
| 2026-03-04 | 搜广推与 LLM 应用 | 2603.04227 Constraint-Aware Generative Re-ranking for Multi-Objective Optimization in Advertising Feeds | industrial-fulltext-review | 待全文 |
| 2026-03-04 | 搜广推与 LLM 应用 | 2603.03988 SORT: A Systematically Optimized Ranking Transformer for Industrial-scale Recommenders | industrial-fulltext-review | B05 |
| 2026-03-01 | 搜广推与 LLM 应用 | 2603.00980 Beyond the Flat Sequence: Hierarchical and Preference-Aware Generative Recommendations | industrial-fulltext-review | 待全文 |
| 2026-02-28 | 搜广推与 LLM 应用 | 2603.00632 Stop Treating Collisions Equally: Qualification-Aware Semantic ID Learning for Recommendation at Industrial Scale | industrial-fulltext-review | B05 |
| 2026-02-28 | 搜广推与 LLM 应用 | 2603.00502 Trinity: A Scenario-Aware Recommendation Framework for Large-Scale Cold-Start Users | industrial-fulltext-review | 待全文 |
| 2026-02-27 | 搜广推与 LLM 应用 | 2602.23717 Recommending Search Filters To Improve Conversions At Airbnb | industrial-fulltext-review | 待全文 |
| 2026-02-26 | 搜广推与 LLM 应用 | 2602.23530 Unified Learning-to-Rank for Multi-Channel Retrieval in Large-Scale E-Commerce Search | industrial-fulltext-review | 待全文 |
| 2026-02-25 | 搜广推与 LLM 应用 | 2602.21600 AQR-HNSW: Accelerating Approximate Nearest Neighbor Search via Density-aware Quantization and Multi-stage Re-ranking | priority-fulltext-review | 待全文 |
| 2026-02-25 | 搜广推与 LLM 应用 | 2603.19249 Spelling Correction in Healthcare Query-Answer Systems: Methods, Retrieval Impact, and Empirical Evaluation | priority-fulltext-review | 待全文 |
| 2026-02-24 | 搜广推与 LLM 应用 | 2602.20995 Generative Pseudo-Labeling for Pre-Ranking with LLMs | manual-review | B05 |
| 2026-02-23 | 搜广推与 LLM 应用 | 2603.06631 T-REX: Transformer-Based Category Sequence Generation for Grocery Basket Recommendation | industrial-fulltext-review | 待全文 |
| 2026-02-20 | 搜广推与 LLM 应用 | 2602.18348 Explaining AutoClustering: Uncovering Meta-Feature Contribution in AutoML for Clustering | priority-fulltext-review | 待全文 |
| 2026-02-20 | 搜广推与 LLM 应用 | 2602.17976 In-Context Pure Exploration in Continuous Decision Spaces | priority-fulltext-review | 待全文 |
| 2026-02-19 | 搜广推与 LLM 应用 | 2602.17058 A Long-term Value Prediction Framework In Video Ranking | industrial-fulltext-review | B05 |
| 2026-02-16 | LLM 后训练 / 搜广推与 LLM 应用 | 2602.15005 Learning User Interests via Reasoning and Distillation for Cross-Domain News Recommendation | industrial-fulltext-review, p2-deferred-review | 待全文 |
| 2026-02-14 | 搜广推与 LLM 应用 | 2602.13581 Climber-Pilot: A Non-Myopic Generative Recommendation Model Towards Better Instruction-Following | industrial-fulltext-review | 待全文 |
| 2026-02-13 | 搜广推与 LLM 应用 | 2602.13134 Awakening Dormant Users: Generative Recommendation with Counterfactual Functional Role Reasoning | industrial-fulltext-review | 待全文 |
| 2026-02-13 | 搜广推与 LLM 应用 | 2602.12972 Jointly Optimizing Debiased CTR and Uplift for Coupons Marketing: A Unified Causal Framework | industrial-fulltext-review | 待全文 |
| 2026-02-13 | 搜广推与 LLM 应用 | 2602.12968 RGAlign-Rec: Ranking-Guided Alignment for Latent Query Reasoning in Recommendation Systems | industrial-fulltext-review | B06 |
| 2026-02-13 | 搜广推与 LLM 应用 | 2602.12593 RQ-GMM: Residual Quantized Gaussian Mixture Model for Multimodal Semantic Discretization in CTR Prediction | industrial-fulltext-review | 待全文 |
| 2026-02-13 | 搜广推与 LLM 应用 | 2602.12564 CAPTS: Channel-Aware, Preference-Aligned Trigger Selection for Multi-Channel Item-to-Item Retrieval | industrial-fulltext-review | 待全文 |
| 2026-02-12 | 搜广推与 LLM 应用 | 2602.12354 An Industrial-Scale Sequential Recommender for LinkedIn Feed Ranking | industrial-fulltext-review | B06 |
| 2026-02-12 | 搜广推与 LLM 应用 | 2602.12041 Compress, Cross and Scale: Multi-Level Compression Cross Networks for Efficient Scaling in Recommender Systems | industrial-fulltext-review | 待全文 |
| 2026-02-12 | 搜广推与 LLM 应用 | 2602.11562 LASER: An Efficient Target-Aware Segmented Attention Framework for End-to-End Long Sequence Modeling | industrial-fulltext-review | 待全文 |
| 2026-02-11 | 搜广推与 LLM 应用 | 2602.11410 CADET: Context-Conditioned Ads CTR Prediction With a Decoder-Only Transformer | industrial-fulltext-review | B06 |
| 2026-02-11 | 搜广推与 LLM 应用 | 2602.10455 Compute Only Once: UG-Separation for Efficient Large Recommendation Models | industrial-fulltext-review | 待全文 |
| 2026-02-10 | 搜广推与 LLM 应用 | 2602.09901 QP-OneModel: A Unified Generative LLM for Multi-Task Query Understanding in Xiaohongshu Search | industrial-fulltext-review | 待全文 |
| 2026-02-10 | 搜广推与 LLM 应用 | 2602.09744 DiffuReason: Bridging Latent Reasoning and Generative Refinement for Sequential Recommendation | industrial-fulltext-review | B06 |
| 2026-02-10 | 搜广推与 LLM 应用 | 2602.09401 SARM: LLM-Augmented Semantic Anchor for End-to-End Live-Streaming Ranking | industrial-fulltext-review | B06 |
| 2026-02-10 | 搜广推与 LLM 应用 | 2602.09386 SMES: Towards Scalable Multi-Task Recommendation via Expert Sparsity | industrial-fulltext-review | 待全文 |
| 2026-02-09 | 搜广推与 LLM 应用 | 2602.09194 ML-DCN: Masked Low-Rank Deep Crossing Network Towards Scalable Ads Click-through Rate Prediction at Pinterest | industrial-fulltext-review | B06 |
| 2026-02-09 | 搜广推与 LLM 应用 | 2602.08530 PIT: A Dynamic Personalized Item Tokenizer for End-to-End Generative Recommendation | industrial-fulltext-review | 待全文 |
| 2026-02-08 | 搜广推与 LLM 应用 | 2602.07987 Learning to Alleviate Familiarity Bias in Video Recommendation | industrial-fulltext-review | 待全文 |
| 2026-02-03 | 搜广推与 LLM 应用 | 2602.03324 SCASRec: A Self-Correcting and Auto-Stopping Model for Generative Route List Recommendation | industrial-fulltext-review | 待全文 |
| 2026-02-01 | 搜广推与 LLM 应用 | 2602.01023 Unifying Ranking and Generation in Query Auto-Completion via Retrieval-Augmented Generation and Multi-Objective Alignment | industrial-fulltext-review | B06 |
| 2026-01-31 | 搜广推与 LLM 应用 | 2602.02582 Uncertainty and Fairness Awareness in LLM-Based Recommendation Systems | priority-fulltext-review | 待全文 |
| 2026-01-31 | 搜广推与 LLM 应用 | 2602.00758 Temporal Leakage in Search-Engine Date-Filtered Web Retrieval: A Retrospective Forecasting Case Study | priority-fulltext-review | 待全文 |
| 2026-01-30 | 搜广推与 LLM 应用 | 2601.22820 User-Adaptive Meta-Learning for Cold-Start Medication Recommendation with Uncertainty Filtering | priority-fulltext-review | 待全文 |
| 2026-01-29 | 搜广推与 LLM 应用 | 2601.21285 Zenith: Scaling up Ranking Models for Billion-scale Livestreaming Recommendation | industrial-fulltext-review | 待全文 |
| 2026-01-28 | 搜广推与 LLM 应用 | 2601.20215 Towards End-to-End Alignment of User Satisfaction via Questionnaire in Video Recommendation | industrial-fulltext-review | 待全文 |
| 2026-01-28 | 搜广推与 LLM 应用 | 2601.20199 MERGE: Next-Generation Item Indexing Paradigm for Large-Scale Streaming Recommendation | industrial-fulltext-review | 待全文 |
| 2026-01-26 | 搜广推与 LLM 应用 | 2601.18664 S\(^2\)GR: Stepwise Semantic-Guided Reasoning in Latent Space for Generative Recommendation | industrial-fulltext-review | 待全文 |
| 2026-01-26 | 搜广推与 LLM 应用 | 2604.09549 Beyond Offline A/B Testing: Context-Aware Agent Simulation for Recommender System Evaluation | industrial-fulltext-review | 待全文 |
| 2026-01-25 | 搜广推与 LLM 应用 | 2601.17836 Unleashing the Potential of Sparse Attention on Long-term Behaviors for CTR Prediction | industrial-fulltext-review | 待全文 |
| 2026-01-24 | 搜广推与 LLM 应用 | 2601.17472 Adversarial Alignment and Disentanglement for Cross-Domain CTR Prediction with Domain-Encompassing Features | industrial-fulltext-review | 待全文 |
| 2026-01-20 | 搜广推与 LLM 应用 | 2601.14333 Hierarchical Contextual Uplift Bandits for Catalog Personalization | industrial-fulltext-review | 待全文 |
| 2026-01-11 | 搜广推与 LLM 应用 | 2601.06873 Applying Embedding-Based Retrieval to Airbnb Search | industrial-fulltext-review | 待全文 |
| 2026-01-08 | 搜广推与 LLM 应用 | 2601.04674 PROMISE: Process Reward Models Unlock Test-Time Scaling Laws in Generative Recommendations | industrial-fulltext-review | 待全文 |
| 2026-01-08 | 搜广推与 LLM 应用 | 2601.04554 Exploring Recommender System Evaluation: A Multi-Modal User Agent Framework for A/B Testing | industrial-fulltext-review | 待全文 |
| 2026-01-06 | 搜广推与 LLM 应用 | 2601.02955 Rethinking Multi-objective Ranking Ensemble in Recommender System: From Score Fusion to Rank Consistency | industrial-fulltext-review | 待全文 |
| 2026-01-05 | 搜广推与 LLM 应用 | 2601.02002 Exploring Approaches for Detecting Memorization of Recommender System Data in Large Language Models | priority-fulltext-review | 待全文 |
执行约束¶
- 每批开始前重新读论文正文，若不存在声称的证据则从批次移出，并在机器账本写明理由。
- 每篇必须有独立机制代码、公开数据/mini-suite 指标、完整论文信息、原文关键图和复现边界。
- 可接入 evolve 的机制必须同时注册 mutation；负结果照常保存。
- B01–B11 全部关闭前，后续扫描仍以本历史账本为基线；关闭后才恢复近期增量扫描。