recommendation strategies

This work addresses the lack of a precise definition of “personalization” in existing algorithmic recourse methods, which hinders systematic evaluation of its impact on effectiveness, cost, and reasonableness. The paper formalizes personalization as individualized actionability by incorporating hard constraints—restricting the set of actionable features—and soft constraints—modeling users’ preferences over the value and cost of recommended actions—within a causal recourse framework. It further introduces a pre-recourse user prompting mechanism to enable personalized recommendations. Experimental results demonstrate that hard constraints substantially reduce both the effectiveness and reasonableness of recourse suggestions. Moreover, significant disparities emerge across social groups in terms of recourse cost and reasonableness, revealing a complex trade-off between personalized design and fairness.

Dynamic optimization of the ε-greedy exploration rate in recommender systems faces practical challenges including batched updates, time-varying traffic, short decision horizons, and minimum exploration constraints; existing heuristic strategies lack theoretical grounding and adaptive capability. This paper proposes the first differentiable optimization framework explicitly modeling Bayesian regret, integrating stochastic gradient descent (SGD) with model predictive control (MPC) to jointly adapt exploration intensity across temporal, batch, and traffic dimensions. The method overcomes limitations of conventional fixed or decaying ε policies, enabling theory-driven online exploration scheduling. Extensive simulations and empirical evaluations on multiple recommendation datasets demonstrate that our approach automatically calibrates exploration strategies, consistently matching or outperforming optimal heuristic baselines in both cumulative reward and cold-start item coverage.

This work addresses the challenges of personalized ranking when users lack knowledge of data attributes or struggle to articulate their preferences explicitly. Existing approaches are limited by reliance on a single candidate item selection strategy, which constrains flexibility and user control. To overcome this, the authors propose a visual analytics framework that integrates model-driven active learning with human-driven item selection, establishing—for the first time—a unified interactive item selection space. This space supports six complementary strategies for expressing list-level preferences and enables iterative learning to produce interpretable rankings. A formative user study (N=10) demonstrates the approach’s effectiveness and reveals trade-offs among accuracy, diversity, novelty, transparency, perceived control, and user satisfaction across different selection strategies.

This paper addresses the challenge of modeling group-level personalized preferences for AI-generated product images in e-commerce. We propose PerFusion—the first diffusion-model-oriented framework for group preference alignment. PerFusion innovatively integrates feature-cross reward modeling with adaptive preference optimization to enable fine-grained comparative preference estimation across multiple candidate images. Built upon a text-to-image diffusion model, the system delivers real-time, high-fidelity image generation, supporting a “sell-then-manufacture” paradigm. Online A/B testing demonstrates that AI-generated products outperform manually designed ones by over 13% in both click-through rate and conversion rate. The framework has been deployed at scale on Alibaba’s platform, significantly reducing inventory and prototyping costs. Key contributions include: (1) the first preference-aligned diffusion framework for e-commerce imagery; (2) a novel reward modeling approach enabling robust multi-candidate preference ranking; and (3) empirical validation of substantial business impact in live production environments.

Industrial recommendation systems often rely on manual, iterative hyperparameter tuning, leading to fragmented historical experimental knowledge that is difficult to reuse. Existing approaches are limited in retrieval accuracy and cross-scenario transferability due to their neglect of hierarchical scenario structures. To address these challenges, this work proposes a closed-loop, self-evolving A/B testing agent that organizes historical strategies into a hierarchical experience tree. It leverages multi-path Tree-RAG to enable goal-aware strategy generation and continuously refines strategies while updating the knowledge base using online A/B test feedback. Evaluated on a short-video e-commerce recommendation system, the method achieves a significant 4.829% increase in GMV, with all guardrail metrics showing positive trends, demonstrating its effectiveness and robustness.

This study addresses the challenge of personalizing large language model (LLM) agents when handling users’ raw, ambiguous queries—a task hindered by difficulties in intent inference, preference extraction, and multi-option decision-making. To this end, the authors introduce the first personalized product search (PPS) evaluation platform grounded in real user behavior, along with the Agent Personalization Benchmark (APeB), and propose a scoring-rule-based evaluation methodology. They further develop history-aware query refinement techniques, notably the Variational Query Refinement Agent (VQRA). Experimental results demonstrate that while mainstream LLMs perform well on explicit queries, their effectiveness is limited during early, ambiguous interaction stages. In contrast, a multi-step agent workflow integrating VQRA substantially improves performance, underscoring the critical role of dedicated modules that leverage historical interaction data for personalization.

This study aims to enhance the digital outreach effectiveness and service quality of Taiwan’s Cultural Memory Bank 2.0 platform. By integrating the Visitor Relationship Management (VRM) framework, user behavior analytics, and Search Engine Optimization (SEO) assessment, the research empirically examines user demographics, browsing patterns, and engagement behaviors while quantifying the platform’s search visibility and organic traffic performance. Key user segments and their behavioral characteristics are identified, and combined with SEO metrics to formulate data-driven strategies for website optimization and social media promotion. The primary contribution lies in the synergistic integration of user personas with SEO indicators, offering actionable pathways to improve both user experience and online visibility for digital cultural platforms.

This work addresses critical limitations in industrial recommendation systems—such as information fragmentation, rigid rule-based designs, and insufficient real-time intent awareness—by introducing DREAM, a novel architecture that operates atop existing models without replacement. DREAM features a programmable and auditable policy control layer, pioneering an agent-based control paradigm grounded in three-tiered intent modeling (L0/L1/L2) and meta-model reasoning. It enables end-to-end autonomous optimization through协同 operation of an intent engine and a meta-engine, integrating on-device signals, edge-cloud trigger chains, and memory-driven hierarchical reasoning (M1–M3). A dual-loop reward mechanism dynamically schedules policies and hyperparameters. Large-scale A/B tests on Taobao’s homepage demonstrate that DREAM boosts IPV by 2.06% and core IPV by 2.39% in reranking alone, with GMV up 0.88%; extending to fine-ranking further improves gains to 3.06% (IPV) and 1.31% (GMV), while PV sustains growth exceeding 1%.