VQL: An End-to-End Context-Aware Vector Quantization Attention for Ultra-Long User Behavior Modeling

In large-scale recommender systems, ultra-long user behavior sequences encode rich signals of evolving interests. Extending sequence length generally improves accuracy, but directly modeling such sequences in production is infeasible due to latency and memory constraints. Existing solutions fall into two categories: (1) top-k retrieval, which truncates the sequence and may discard most attention mass when L >> k; and (2) encoder-based compression, which preserves coverage but often over-compresses and fails to incorporate key context such as temporal gaps or target-aware signals. Neither class achieves a good balance of low-loss compression, context awareness, and efficiency. We propose VQL, a context-aware Vector Quantization Attention framework for ultra-long behavior modeling, with three innovations. (1) Key-only quantization: only attention keys are quantized, while values remain intact; we prove that softmax normalization yields an error bound independent of sequence length, and a codebook loss directly supervises quantization quality. This also enables L-free inference via offline caches. (2) Multi-scale quantization: attention heads are partitioned into groups, each with its own small codebook, which reduces quantization error while keeping cache size fixed. (3) Efficient context injection: static features (e.g., item category, modality) are directly integrated, and relative position is modeled via a separable temporal kernel. All context is injected without enlarging the codebook, so cached representations remain query-independent. Experiments on three large-scale datasets (KuaiRand-1K, KuaiRec, TMALL) show that VQL consistently outperforms strong baselines, achieving higher accuracy while reducing inference latency, establishing a new state of the art in balancing accuracy and efficiency for ultra-long sequence recommendation.

Expert-Guided Diffusion Planner for Auto-bidding

Auto-bidding is extensively applied in advertising systems, serving a multitude of advertisers. Generative bidding is gradually gaining traction due to its robust planning capabilities and generalizability. In contrast to traditional reinforcement learning-based bidding, generative bidding does not rely on the Markov Decision Process (MDP) exhibiting superior planning capabilities in long-horizon scenarios. Conditional diffusion modeling approaches have demonstrated significant potential in the realm of auto-bidding. However, relying solely on return as the optimality condition is weak to guarantee the generation of genuinely optimal decision sequences, lacking personalized structural information. Moreover, diffusion models' t-step autoregressive generation mechanism inherently carries timeliness risks. To address these issues, we propose a novel conditional diffusion modeling method based on expert trajectory guidance combined with a skip-step sampling strategy to enhance generation efficiency. We have validated the effectiveness of this approach through extensive offline experiments and achieved statistically significant results in online A/B testing, achieving an increase of 11.29% in conversion and a 12.35% in revenue compared with the baseline.

R4ec: A Reasoning, Reflection, and Refinement Framework for Recommendation Systems

Harnessing Large Language Models (LLMs) for recommendation systems has emerged as a prominent avenue, drawing substantial research interest. However, existing approaches primarily involve basic prompt techniques for knowledge acquisition, which resemble System-1 thinking. This makes these methods highly sensitive to errors in the reasoning path, where even a small mistake can lead to an incorrect inference. To this end, in this paper, we propose $R^{4}$ec, a reasoning, reflection and refinement framework that evolves the recommendation system into a weak System-2 model. Specifically, we introduce two models: an actor model that engages in reasoning, and a reflection model that judges these responses and provides valuable feedback. Then the actor model will refine its response based on the feedback, ultimately leading to improved responses. We employ an iterative reflection and refinement process, enabling LLMs to facilitate slow and deliberate System-2-like thinking. Ultimately, the final refined knowledge will be incorporated into a recommendation backbone for prediction. We conduct extensive experiments on Amazon-Book and MovieLens-1M datasets to demonstrate the superiority of $R^{4}$ec. We also deploy $R^{4}$ec on a large scale online advertising platform, showing 2.2\% increase of revenue. Furthermore, we investigate the scaling properties of the actor model and reflection model.

Graph Learning for Cooperative Cell-Free ISAC Systems: From Optimization to Estimation

Cell-free integrated sensing and communication (ISAC) systems have emerged as a promising paradigm for sixth-generation (6G) networks, enabling simultaneous high-rate data transmission and high-precision radar sensing through cooperative distributed access points (APs). Fully exploiting these capabilities requires a unified design that bridges system-level optimization with multi-target parameter estimation. This paper proposes an end-to-end graph learning approach to close this gap, modeling the entire cell-free ISAC network as a heterogeneous graph to jointly design the AP mode selection, user association, precoding, and echo signal processing for multi-target position and velocity estimation. In particular, we propose two novel heterogeneous graph learning frameworks: a dynamic graph learning framework and a lightweight mirror-based graph attention network (mirror-GAT) framework. The dynamic graph learning framework employs structural and temporal attention mechanisms integrated with a three-dimensional convolutional neural network (3D-CNN), enabling superior performance and robustness in cell-free ISAC environments. Conversely, the mirror-GAT framework significantly reduces computational complexity and signaling overhead through a bi-level iterative structure with share adjacency. Simulation results validate that both proposed graph-learning-based frameworks achieve significant improvements in multi-target position and velocity estimation accuracy compared to conventional heuristic and optimization-based designs. Particularly, the mirror-GAT framework demonstrates substantial reductions in computational time and signaling overhead, underscoring its suitability for practical deployments.

Personalized Tree based progressive regression model for watch-time prediction in short video recommendation

In online video platforms, accurate watch time prediction has become a fundamental and challenging problem in video recommendation. Previous research has revealed that the accuracy of watch time prediction highly depends on both the transformation of watch-time labels and the decomposition of the estimation process. TPM (Tree based Progressive Regression Model) achieves State-of-the-Art performance with a carefully designed and effective decomposition paradigm. TPM discretizes the watch time into several ordinal intervals and organizes them into a binary decision tree, where each node corresponds to a specific interval. At each non-leaf node, a binary classifier is used to determine the specific interval in which the watch time variable most likely falls, based on the prediction outcome at its parent node. The tree structure serves as the core of TPM, as it defines the decomposition of watch time estimation and determines how the ordinal intervals are discretized. However, in TPM, the tree is predefined as a full binary tree, which may be sub-optimal for the following reasons. First, a full binary tree implies an equal partitioning of the watch time space, which may struggle to capture the complexity of real-world watch time distributions. Second, instead of relying on a globally fixed tree structure, we advocate for a personalized, data-driven tree that can be learned in an end-to-end manner. Therefore, we propose PTPM to enable a highly personalized decomposition of watch estimation with better efficacy and efficiency. Moreover, we reveal that TPM is affected by selection bias due to conditional modeling and devise a simple approach to address it. We conduct extensive experiments on both offline datasets and online environments. PTPM has been fully deployed in core traffic scenarios and serves more than 400 million users per day.

Hierarchical Tree Search-based User Lifelong Behavior Modeling on Large Language Model

Large Language Models (LLMs) have garnered significant attention in Recommendation Systems (RS) due to their extensive world knowledge and robust reasoning capabilities. However, a critical challenge lies in enabling LLMs to effectively comprehend and extract insights from massive user behaviors. Current approaches that directly leverage LLMs for user interest learning face limitations in handling long sequential behaviors, effectively extracting interest, and applying interest in practical scenarios. To address these issues, we propose a Hierarchical Tree Search-based User Lifelong Behavior Modeling framework (HiT-LBM). HiT-LBM integrates Chunked User Behavior Extraction (CUBE) and Hierarchical Tree Search for Interest (HTS) to capture diverse interests and interest evolution of user. CUBE divides user lifelong behaviors into multiple chunks and learns the interest and interest evolution within each chunk in a cascading manner. HTS generates candidate interests through hierarchical expansion and searches for the optimal interest with process rating model to ensure information gain for each behavior chunk. Additionally, we design Temporal-Ware Interest Fusion (TIF) to integrate interests from multiple behavior chunks, constructing a comprehensive representation of user lifelong interests. The representation can be embedded into any recommendation model to enhance performance. Extensive experiments demonstrate the effectiveness of our approach, showing that it surpasses state-of-the-art methods.

Navigate the Unknown: Enhancing LLM Reasoning with Intrinsic Motivation Guided Exploration

Reinforcement learning (RL) has emerged as a pivotal method for improving the reasoning capabilities of Large Language Models (LLMs). However, prevalent RL approaches such as Proximal Policy Optimization (PPO) and Group-Regularized Policy Optimization (GRPO) face critical limitations due to their reliance on sparse outcome-based rewards and inadequate mechanisms for incentivizing exploration. These limitations result in inefficient guidance for multi-step reasoning processes. Specifically, sparse reward signals fail to deliver effective or sufficient feedback, particularly for challenging problems. Furthermore, such reward structures induce systematic biases that prioritize exploitation of familiar trajectories over novel solution discovery. These shortcomings critically hinder performance in complex reasoning tasks, which inherently demand iterative refinement across ipntermediate steps. To address these challenges, we propose an Intrinsic Motivation guidEd exploratioN meThOd foR LLM Reasoning (i-MENTOR), a novel method designed to both deliver dense rewards and amplify explorations in the RL-based training paradigm. i-MENTOR introduces three key innovations: trajectory-aware exploration rewards that mitigate bias in token-level strategies while maintaining computational efficiency; dynamic reward scaling to stabilize exploration and exploitation in large action spaces; and advantage-preserving reward implementation that maintains advantage distribution integrity while incorporating exploratory guidance. Experiments across three public datasets demonstrate i-MENTOR's effectiveness with a 22.39% improvement on the difficult dataset Countdown-4.

Generative Auto-Bidding with Value-Guided Explorations

Auto-bidding, with its strong capability to optimize bidding decisions within dynamic and competitive online environments, has become a pivotal strategy for advertising platforms. Existing approaches typically employ rule-based strategies or Reinforcement Learning (RL) techniques. However, rule-based strategies lack the flexibility to adapt to time-varying market conditions, and RL-based methods struggle to capture essential historical dependencies and observations within Markov Decision Process (MDP) frameworks. Furthermore, these approaches often face challenges in ensuring strategy adaptability across diverse advertising objectives. Additionally, as offline training methods are increasingly adopted to facilitate the deployment and maintenance of stable online strategies, the issues of documented behavioral patterns and behavioral collapse resulting from training on fixed offline datasets become increasingly significant. To address these limitations, this paper introduces a novel offline Generative Auto-bidding framework with Value-Guided Explorations (GAVE). GAVE accommodates various advertising objectives through a score-based Return-To-Go (RTG) module. Moreover, GAVE integrates an action exploration mechanism with an RTG-based evaluation method to explore novel actions while ensuring stability-preserving updates. A learnable value function is also designed to guide the direction of action exploration and mitigate Out-of-Distribution (OOD) problems. Experimental results on two offline datasets and real-world deployments demonstrate that GAVE outperforms state-of-the-art baselines in both offline evaluations and online A/B tests. The implementation code is publicly available to facilitate reproducibility and further research.

CHIME: A Compressive Framework for Holistic Interest Modeling

Modeling holistic user interests is important for improving recommendation systems but is challenged by high computational cost and difficulty in handling diverse information with full behavior context. Existing search-based methods might lose critical signals during behavior selection. To overcome these limitations, we propose CHIME: A Compressive Framework for Holistic Interest Modeling. It uses adapted large language models to encode complete user behaviors with heterogeneous inputs. We introduce multi-granular contrastive learning objectives to capture both persistent and transient interest patterns and apply residual vector quantization to generate compact embeddings. CHIME demonstrates superior ranking performance across diverse datasets, establishing a robust solution for scalable holistic interest modeling in recommendation systems.

BBQRec: Behavior-Bind Quantization for Multi-Modal Sequential Recommendation

Multi-modal sequential recommendation systems leverage auxiliary signals (e.g., text, images) to alleviate data sparsity in user-item interactions. While recent methods exploit large language models to encode modalities into discrete semantic IDs for autoregressive prediction, we identify two critical limitations: (1) Existing approaches adopt fragmented quantization, where modalities are independently mapped to semantic spaces misaligned with behavioral objectives, and (2) Over-reliance on semantic IDs disrupts inter-modal semantic coherence, thereby weakening the expressive power of multi-modal representations for modeling diverse user preferences. To address these challenges, we propose a Behavior-Bind multi-modal Quantization for Sequential Recommendation (BBQRec for short) featuring dual-aligned quantization and semantics-aware sequence modeling. First, our behavior-semantic alignment module disentangles modality-agnostic behavioral patterns from noisy modality-specific features through contrastive codebook learning, ensuring semantic IDs are inherently tied to recommendation tasks. Second, we design a discretized similarity reweighting mechanism that dynamically adjusts self-attention scores using quantized semantic relationships, preserving multi-modal synergies while avoiding invasive modifications to the sequence modeling architecture. Extensive evaluations across four real-world benchmarks demonstrate BBQRec's superiority over the state-of-the-art baselines.