Most large-scale recommender systems follow a multi-stage cascade of retrieval, pre-ranking, ranking, and re-ranking. A key challenge at the pre-ranking stage arises from the heterogeneity of training instances sampled from coarse-grained retrieval results, fine-grained ranking signals, and exposure feedback. Our analysis reveals that prevailing pre-ranking methods, which indiscriminately mix heterogeneous samples, suffer from gradient conflicts: hard samples dominate training while easy ones remain underutilized, leading to suboptimal performance. We further show that the common practice of uniformly scaling model complexity across all samples is inefficient, as it overspends computation on easy cases and slows training without proportional gains. To address these limitations, this paper presents Heterogeneity-Aware Adaptive Pre-ranking (HAP), a unified framework that mitigates gradient conflicts through conflict-sensitive sampling coupled with tailored loss design, while adaptively allocating computational budgets across candidates. Specifically, HAP disentangles easy and hard samples, directing each subset along dedicated optimization paths. Building on this separation, it first applies lightweight models to all candidates for efficient coverage, and further engages stronger models on the hard ones, maintaining accuracy while reducing cost. This approach not only improves pre-ranking effectiveness but also provides a practical perspective on scaling strategies in industrial recommender systems. HAP has been deployed in the Toutiao production system for 9 months, yielding up to 0.4% improvement in user app usage duration and 0.05% in active days, without additional computational cost. We also release a large-scale industrial hybrid-sample dataset to enable the systematic study of source-driven candidate heterogeneity in pre-ranking.
论文检索
输入标题、作者或关键词，从 954 篇学术成果中精准定位
In recent years, the success of large language models (LLMs) has driven the exploration of scaling laws in recommender systems. However, models that demonstrate scaling laws are actually challenging to deploy in industrial settings for modeling long sequences of user behaviors, due to the high computational complexity of the standard self-attention mechanism. Despite various sparse self-attention mechanisms proposed in other fields, they are not fully suited for recommendation scenarios. This is because user behaviors exhibit personalization and temporal characteristics: different users have distinct behavior patterns, and these patterns change over time, with data from these users differing significantly from data in other fields in terms of distribution. To address these challenges, we propose SparseCTR, an efficient and effective model specifically designed for long-term behaviors of users. To be precise, we first segment behavior sequences into chunks in a personalized manner to avoid separating continuous behaviors and enable parallel processing of sequences. Based on these chunks, we propose a three-branch sparse self-attention mechanism to jointly identify users' global interests, interest transitions, and short-term interests. Furthermore, we design a composite relative temporal encoding via learnable, head-specific bias coefficients, better capturing sequential and periodic relationships among user behaviors. Extensive experimental results show that SparseCTR not only improves efficiency but also outperforms state-of-the-art methods. More importantly, it exhibits an obvious scaling law phenomenon, maintaining performance improvements across three orders of magnitude in FLOPs. In online A/B testing, SparseCTR increased CTR by 1.72% and CPM by 1.41%. Our source code is available at https://github.com/laiweijiang/SparseCTR.
Short-video platforms have rapidly become a new generation of information retrieval systems, where users formulate queries to access desired videos. However, user queries, especially long-tail ones, often suffer from spelling errors, incomplete phrasing, and ambiguous intent, resulting in mismatches between user expectations and retrieved results. While large language models (LLMs) have shown success in long-tail query rewriting within e-commerce, they struggle on short-video platforms, where proprietary content such as short videos, live streams, micro dramas, and user social networks falls outside their training distribution. To address this challenge, we introduce CardRewriter, an LLM-based framework that incorporates domain-specific knowledge to enhance long-tail query rewriting. For each query, our method aggregates multi-source knowledge relevant to the query and summarizes it into an informative and query-relevant knowledge card. This card then guides the LLM to better capture user intent and produce more effective query rewrites. We optimize CardRewriter using a two-stage training pipeline: supervised fine-tuning followed by group relative policy optimization, with a tailored reward system balancing query relevance and retrieval effectiveness. Offline experiments show that CardRewriter substantially improves rewriting quality for queries targeting proprietary content. Online A/B testing further confirms significant gains in long-view rate (LVR) and click-through rate (CTR), along with a notable reduction in initiative query reformulation rate (IQRR). Since September 2025, CardRewriter has been deployed on Kuaishou, one of China's largest short-video platforms, serving hundreds of millions of users daily.
Trends on short-video platforms evolve at a rapid pace, with new content issues emerging every day that fall outside the coverage of existing annotation policies. However, traditional human-driven discovery of emerging issues is too slow, which leads to delayed updates of annotation policies and poses a major challenge for effective content governance. In this work, we propose an automatic issue discovery method based on multimodal LLM agents. Our approach automatically recalls short videos containing potential new issues and applies a two-stage clustering strategy to group them, with each cluster corresponding to a newly discovered issue. The agent then generates updated annotation policies from these clusters, thereby extending coverage to these emerging issues. Our agent has been deployed in the real system. Both offline and online experiments demonstrate that this agent-based method significantly improves the effectiveness of emerging-issue discovery (with an F1 score improvement of over 20%) and enhances the performance of subsequent issue governance (reducing the view count of problematic videos by approximately 15%). More importantly, compared to manual issue discovery, it greatly reduces time costs and substantially accelerates the iteration of annotation policies.
Generative Recommenders (GRs), exemplified by the Hierarchical Sequential Transduction Unit (HSTU), have emerged as a powerful paradigm for modeling long user interaction sequences. However, we observe that their ''flat-sequence'' assumption overlooks the rich, intrinsic structure of user behavior. This leads to two key limitations: a failure to capture the temporal hierarchy of session-based engagement, and computational inefficiency, as dense attention introduces significant noise that obscures true preference signals within semantically sparse histories, which deteriorates the quality of the learned representations. To this end, we propose a novel framework named HPGR (Hierarchical and Preference-aware Generative Recommender), built upon a two-stage paradigm that injects these crucial structural priors into the model to handle the drawback. Specifically, HPGR comprises two synergistic stages. First, a structure-aware pre-training stage employs a session-based Masked Item Modeling (MIM) objective to learn a hierarchically-informed and semantically rich item representation space. Second, a preference-aware fine-tuning stage leverages these powerful representations to implement a Preference-Guided Sparse Attention mechanism, which dynamically constrains computation to only the most relevant historical items, enhancing both efficiency and signal-to-noise ratio. Empirical experiments on a large-scale proprietary industrial dataset from APPGallery and an online A/B test verify that HPGR achieves state-of-the-art performance over multiple strong baselines, including HSTU and MTGR.
Short-video recommenders such as Douyin must exploit extremely long user histories without breaking latency or cost budgets. We present an end-to-end system that scales long-sequence modeling to 10k -length histories in production. First, we introduce Stacked Target-to-History Cross Attention (STCA), which replaces history self-attention with stacked cross-attention from the target to the history, reducing complexity from quadratic to linear in sequence length and enabling efficient end-to-end training. Second, we propose Request Level Batching (RLB), a user-centric batching scheme that aggregates multiple targets for the same user/request to share the user-side encoding, substantially lowering sequence-related storage, communication, and compute without changing the learning objective. Third, we design a length-extrapolative training strategy---train on shorter windows, infer on much longer ones---so the model generalizes to 10k histories without additional training cost. Across offline and online experiments, we observe predictable, monotonic gains as we scale history length and model capacity, mirroring the scaling law behavior observed in large language models. Deployed at full traffic on Douyin, our system delivers significant improvements on key engagement metrics while meeting production latency, demonstrating a practical path to scaling end-to-end long-sequence recommendation to the 10k regime.
Auto-bidding plays an essential role in online advertising, allowing agents to automatically adjust bids for advertisers. Recently, the rise of Marketing Management service in e-commerce platforms has driven the evolution from auto-bidding to auto-marketing, enabling merchants to delegate their advertising bidding and product's coupon discounting decisions to agents. Auto-marketing requires agents to jointly decide on bidding and coupon discounting. Furthermore, compared to classic static constraints, auto-marketing agent faces a self-funding constraint (where the budget for both bidding and coupon discounting is entirely derived from the agent's commission revenue). Existing rule-based or RL-based methods often struggle with dynamic environments and complex sequential dependencies. To overcome these limitations, we propose a Generative Auto-Marketing framework (GAM), designed for performing joint sequential decisions on bidding and coupon discounting, and optimizing business objectives through post-training alignment. Furthermore, GAM employs a flexible, constraint-aware reward alignment module, and utilizes Group Relative Policy Optimization (GRPO) to align the pre-trained model, thus empirically balancing objective maximization and constraint satisfaction. We construct an offline simulation environment based on large-scale real-world dataset, and demonstrate the effectiveness of GAM through extensive experimental results.
Query-product relevance prediction is fundamental to e-commerce search and has become even more critical in the era of AI-powered shopping, where semantic understanding and complex reasoning directly shape the user experience and exert an indirect, yet substantial impact on business conversion. Large Language Models (LLMs) enable generative, reasoning-based approaches, typically aligned via supervised fine-tuning (SFT) or preference optimization methods like Direct Preference Optimization (DPO). However, the increasing complexity of business rules and user queries exposes the inability of existing methods to endow models with robust reasoning capacity for long-tail and challenging cases. Efforts to address this via reinforcement learning strategies like Group Relative Policy Optimization (GRPO) often suffer from sparse terminal rewards, offering insufficient guidance for multi-step reasoning, which in turn slows convergence. To address these challenges, we propose TaoSR-AGRL, an Adaptive Guided Reinforcement Learning framework for LLM-based relevance prediction in Taobao Search Relevance. TaoSR-AGRL introduces two key innovations: (1) Rule-aware Reward Shaping, which decomposes the final relevance judgment into dense, structured rewards aligned with domain-specific relevance criteria; and (2) Adaptive Guided Replay, which identifies low-accuracy rollouts during training and injects targeted ground-truth guidance to steer the policy away from stagnant, rule-violating reasoning patterns toward compliant trajectories. TaoSR-AGRL was evaluated on large-scale datasets and through online evaluations on Taobao Search. It consistently outperforms DPO and GRPO baselines, improving relevance accuracy and rule adherence, with measurable gains in user engagement and stable training dynamics. The model has been deployed on Taobao, serving hundreds of millions of users.
In e-commerce search, the diverse ways in which users express intentions lead to lexical and semantic gaps between queries and product descriptions, making query rewriting (QR) indispensable for improving matching efficiency. With the development of LLMs, QR has evolved from discriminative approaches to various LLM-based alignment methods. However, these methods typically treat all queries uniformly, without fundamentally distinguishing their rewriting difficulty or underlying linguistic issues, making the rewritten query deviate from human expectations. To address this limitation, we propose AWHCP (Aligning with Human Cognition and Preference), a novel framework that adopts a human-centric perspective and introduces the Problem–Intention–Fix–Rewrite (PIFR) paradigm. Built upon PIFR, AWHCP establishes a multi-granularity alignment training framework that simultaneously aligns with both system retrieval preferences and human rewriting behaviors. First, we construct high-quality PIFR-structured data and perform supervised fine-tuning to enable the model to learn human-like rewriting patterns. Second, we apply beam search to generate multiple candidates and leverage system-side feedback signals to conduct coarse-grained direct preference alignment, endowing the model with initial difficulty-aware reasoning capabilities. Third, we introduce a multi-dimensional rewrite quality judgment model trained via Group Relative Policy Optimization (GRPO), enabling fine-grained alignment with nuanced human rewriting preferences. Deployed on 1688's main search engine since August 2025, AWHCP has demonstrated strong effectiveness through extensive offline evaluations and large-scale online A/B tests, leading to a +3.9% gain in UV-L2O.
We present BanditLP, a scalable multi-stakeholder contextual bandit framework that unifies neural Thompson Sampling (TS) for learning objective-specific outcomes with a large-scale linear program (LP) for constrained action selection at serving time. The methodology is application-agnostic, compatible with arbitrary neural architectures, and deployable at web scale, with an LP solver capable of handling billions of variables. Experiments on public benchmarks and synthetic data show consistent gains over strong baselines. We apply this approach in LinkedIn's email marketing system and demonstrate business win, illustrating the value of integrated exploration and constrained optimization in production.
Web-scale editorial crawling must balance coverage and freshness within tight politeness and request budgets. Nonetheless, in production systems, manual seed management remains common despite being inefficient—oversampling redundant seeds while missing high-yield ones. URLBank replaces manual curation with a label-free controller that infers optimal seed selection directly from temporal crawl telemetry. It identifies candidate entry points, estimates stability—the persistence of links across crawler revolutions—and productivity—the rate of first-seen publications—and ranks them through greedy marginal gain on a shared-credit coverage objective. In a shadow A/B evaluation spanning 5,238 sites, URLBank consistently achieves higher coverage, greater efficiency, and earlier discovery under identical conditions. Gains remain stable across Top-K budgets, approaching near-complete coverage with far fewer seeds. Deployed alongside Meltwater's production crawler Pulitzer, URLBank operates with versioned policies, ranked prefixes for crawl budgets, and integrated health diagnostics, making allocation transparent, auditable, and reversible. Together, these results demonstrate that temporal signals, through an interpretable greedy objective, yield large, measurable improvements in industrial-scale coverage, resource efficiency, and freshness.
Traditional search engines return uniform results for identical queries, overlooking users' personalized intents. While personalized search has been extensively studied, research on query rewriting for personalized intents has been constrained by traditional approaches like statistical co-occurrence and synonym expansion. Current work primarily addresses multi-turn dialogue scenarios in Conversational AI rather than exploring applications in large-scale search engines. This typically stems from the absence of real search scenario data and the difficulty of inferring users' intents through personalized reasoning. While Large Language Models (LLMs) combined with Chain-of-Thought (CoT) capabilities provide possibilities for personalized reasoning, CoT introduces additional reasoning overhead that is difficult to accept in online scenarios requiring low latency. To address this, this paper proposes PicQue (Personalized Efficient Query Rewrite), a personalized query rewriting model training pipeline aimed at achieving high accuracy with low latency. PicQue contains a two-stage training that first employs a novel Hybrid Supervised Fine-Tuning strategy to retain the model's reasoning capabilities while allowing the decoding process to skip CoT, thereby obtaining CoT's accuracy gains without increasing latency. Building on this foundation, PicQue conducts second-stage reinforcement learning using Group Relative Policy Optimization (GRPO) to further improve rewriting accuracy and reduce the risk of over-rewriting. We also propose a Guided Search strategy to optimize GRPO training, alleviating the reduction in training sample utilization when all sampling rollouts are wrong. Extensive offline and online experiments demonstrate PicQue's effectiveness. In offline metrics, PicQue achieves over 7% improvement in rewriting accuracy compared to baseline methods and compresses up to 95% of decoding tokens. In online A/B tests, user satisfaction increases by 1.78% and query change rate decreases by 0.71%, achieving significant online gains.
Promoting cold items to achieve rapid growth remains a fundamental challenge in billion-scale recommendation systems, as traditional natural/organic recommendation approaches primarily focus on Click-Through Rate (CTR) optimization, which naturally limits the exposure and spread of cold items. Recently, the AliBoost (V1) framework introduced boosting strategies to promote cold items to users most likely to click them. However, it still follows the same CTR-oriented optimization approach, thereby limiting long-term ecosystem health. In this work, we present the CTR-growth balanced boosting framework AliBoostV2, which explicitly considers the growth value of boosting candidate users and selects optimal users to balance immediate CTR goals with long-term growth potential. AliBoostV2 includes two key innovations: (1) a tailored Growth Potential Prediction module using counterfactual reasoning to estimate the additional natural traffic generated by each potential boosting exposure, and (2) a Dynamic CTR-Growth Boosting strategy that dynamically captures users' different interaction patterns across various time periods and delivers to users who can both click and contribute to growth simultaneously. AliBoostV2 has been deployed in production across Alibaba and Taobao's main platforms over the past six months, successfully cold-starting over one billion new items. Compared to the AliBoost (V1) framework, our approach achieves significant improvements of over 17.54% in both clicks and gross merchandise value (GMV) for cold items within a 180-day period. Extensive online analyses and rigorous A/B testing demonstrate the effectiveness of AliBoostV2 in addressing critical ecosystem challenges in billion-scale recommendation.
A practical log anomaly detection system for large-scale web services, suitable for real industrial deployment, should exhibit three key characteristics: fidelity to decisive lexical cues, human-auditable explanations, and low inference latency. Existing approaches often compromise one or more of these requirements by discarding crucial tokens during parsing, obscuring decisions through dense embeddings, or incurring heavy computational costs. We introduce Smart Eye, a deployable two-stage key pattern retrieval framework that employs large language models (LLMs) strictly as proposal engines, coupled with a deterministic verifier. In Stage I, the system enhances recall of anomalous logs under a configurable false positive budget using budgeted maximum coverage selection. In Stage II, precision is monotonically improved by refining parent patterns into specialized child patterns and introducing targeted exclusions. These essential patterns, extracted by LLMs and then confirmed, will be formulated into regular expressions to enable efficient online matching of anomalous logs. We provide both theoretical justification and a practical implementation that runs efficiently on commodity CPUs. When evaluated on three real world Huawei Cloud Core log datasets, Smart Eye achieves state-of-the-art anomaly detection with a average F1 score of 0.9877, outperforming the strongest neural baseline. Moreover, it produces human interpretable evidence chains and supports low latency inference. These results demonstrate that combining LLM-guided proposal generation with deterministic verification forms a robust and deployable alternative to traditional end to end embedding pipelines for industrial log analysis.
Recommender Systems (RS) are fundamental to modern online services. While most existing approaches optimize for short-term engagement, recent work has begun to explore reinforcement learning (RL) to model long-term user value. However, these efforts face significant challenges due to the vast, dynamic action spaces inherent in RS, which hinder stable policy learning. To resolve this bottleneck, we introduce Hierarchical Semantic RL (HSRL), which reframes RL-based recommendation over a fixed Semantic Action Space (SAS). HSRL encodes items as Semantic IDs (SIDs) for policy learning, and maps SIDs back to their original items via a fixed lookup during execution. To align decision-making with SID generation, the Hierarchical Policy Network (HPN) operates in a coarse-to-fine manner, employing hierarchical residual state modeling to refine each level's context from the previous level's residual, thereby reducing representation–decision mismatch. In parallel, a Multi-level Critic (MLC) provides token-level value estimates, enabling fine-grained credit assignment. Across public benchmarks and a large-scale production dataset from a leading short-video advertising platform, HSRL consistently surpasses state-of-the-art baselines. In online deployment over a 7-day A/B testing, it delivers an 18.421% ADVV lift and a 1.251% increase in Revenue, supporting HSRL as a scalable paradigm for RL-based recommendation.
The rapid development of large language models (LLMs) has significantly enhanced the capabilities of AI-native applications, offering substantial improvements in user experience across various sectors. In particular, the integration of LLMs with external APIs has become critical for services such as Baidu Maps, which leverages ERNIE Bot to provide real-time, intelligent responses through its AI Guide service. However, as user queries diversify, the ability to dynamically discover, design, and integrate new APIs has become increasingly essential. This paper addresses the challenges of automating the real-world API discovery, design, and integration process, focusing on mitigating human labor costs and biases while ensuring the creation of high-quality training data. To this end, we propose an Automatic Real-world API Discovery and Deployment (ARADD) framework to efficiently discover new real-world APIs suitable for query solving and automatically master them with minimal labor cost. Specifically, we firstly propose a Multi-Stage LLM-empowered Iterative Intent Extraction method, which integrates a closed-source LLM with our lightweight agent to capture each new intent accurately and efficiently. Secondly, we propose a Contextual-Aware API Design and Self-Instruct Data Generation module to discover APIs suitable for the captured new intent and generate training data pairs of this intent. Finally, a Two-Stage Data Filtering module is introduced to distill the most influential data point for fine-tuning the agent model. Extensive experiments on a real-world log dataset and the online service side validate the effectiveness of our proposed framework.