arXiv subjects

Tianyu Zhan

Publications and source records attributed to Tianyu Zhan.

At least 19 recordsLinked to original sources

arXiv2026

Semantic Trimming and Auxiliary Multi-step Prediction for Generative Recommendation

Generative Recommendation (GR) has recently transitioned from atomic item-indexing to Semantic ID (SID)-based frameworks to capture intrinsic item relationships and enhance generalization. However, the adoption of high-granularity SIDs leads to two critical challenges: prohibitive training overhead due to sequence expansion and unstable performance reliability characterized by non-monotonic accuracy fluctuations. We identify that these disparate issues are fundamentally rooted in the Semantic Dilution Effect, where redundant tokens waste massive computation and dilute the already sparse learning signals in recommendation. To counteract this, we propose STAMP (Semantic Trimming and Auxiliary Multi-step Prediction), a framework utilizing a dual-end optimization strategy. We argue that effective SID learning requires simultaneously addressing low input information density and sparse output supervision. On the input side, Semantic Adaptive Pruning (SAP) dynamically filters redundancy during the forward pass, converting noise-laden sequences into compact, information-rich representations. On the output side, Multi-step Auxiliary Prediction (MAP) employs a multi-token objective to densify feedback, strengthening long-range dependency capture and ensuring robust learning signals despite compressed inputs. Unifying input purification and signal amplification, STAMP enhances both training efficiency and representation capability. Experiments on public Amazon and large-scale industrial datasets show STAMP achieves 1.23--1.38$\times$ speedup and 17.2\%--54.7\% VRAM reduction while maintaining or improving performance across multiple architectures.

cs.IR↗

arXiv2026

TSGR: Taobao Search Generative Retrieval

Generative retrieval (GR) has demonstrated strong promise for industrial e-commerce search by training a single autoregressive model to directly generate the Semantic IDs (SIDs) of target items. However, existing GR systems are primarily optimized for semantic matching and remain insensitive to item business value: SID construction is value-unaware, and candidates are ranked without access to item side-info. Consequently, high-value items are often missed or deprioritized at the retrieval stage, limiting downstream business impact. This limitation is particularly critical in industrial settings such as Taobao Search, where business objectives are central to system design. To address this, we propose $\textbf{T}$aobao $\textbf{S}$earch $\textbf{G}$enerative $\textbf{R}$etrieval ($\textbf{TSGR}$), a unified generative retrieval framework that incorporates value awareness into both item representation and candidate ranking. 1) For item representation, TSGR introduces $\textbf{Query-aware Parallel SID (QP-SID)}$, which encodes query-conditioned value orderings into the SID construction by building parallel codebooks derived from query-item statistics, so that higher-value and query-relevant items are assigned better token indices. 2) For candidate ranking, we introduce a $\textbf{Value-aware Ranking Module (VRM)}$ that is built upon and jointly optimized with the GR, enabling a single model to seamlessly serve as both retriever and pre-ranker without a dedicated pre-ranking stage. A progressive training pipeline further aligns the model with semantic relevance, user preferences, and business objectives. Offline experiments show that TSGR achieves a 9.16\% improvement in HR@1000, and online A/B tests further validate its effectiveness, yielding gains of +0.43\% in IPV, +1.12\% in Transaction Count, and +1.64\% in GMV. TSGR has been fully deployed in production.

cs.IR↗

arXiv2026

Synergy Area with FDR-controlled Evaluation (SAFE) to robustly assess safety profile in clinical trials

Safety assessment plays a fundamental role in developing a new drug via clinical trials for ethical considerations. Due to complexity, manual review is typically conducted on the totality of data to draw safety conclusions. There are some existing quantitative methods to facilitate or tailor further medical review, with a controlled error rate and integration of clinical knowledge. In addition to those two key aspects, we emphasize the importance of relying on substantial evidence to draw robust conclusions on safety. Motivated by these three important properties, we propose a two-layer Synergy Area with FDR-controlled Evaluation (SAFE) structural framework to robustly assess the safety profile in clinical trials. In the first layer of SAFE, we investigate each clinically meaningful Synergy Area (SA) based on compelling evidence. In the next layer, the false discovery rate (FDR) is controlled for potential findings across all SAs. Simulation studies show that SAFE properly controls error rates within and across SAs at the nominal level. We further apply the proposed approach to two case studies based on real data from the Historical Trial Data (HTD) Sharing Initiative of the DataCelerate platform. As compared to some direct methods, SAFE demonstrates an appealing feature of screening out extreme data and reaching solid safety conclusions. It can act as either a building block in another framework, or a platform to incorporate additional components.

stat.AP↗

arXiv2026

RASTP: Representation-Aware Semantic Token Pruning for Generative Recommendation with Semantic Identifiers

Generative recommendation systems typically leverage Semantic Identifiers (SIDs), which represent each item as a sequence of tokens that encode semantic information. However, representing item ID with multiple SIDs significantly increases input sequence length, which is a major determinant of computational complexity and memory consumption. While existing efforts primarily focus on optimizing attention computation and KV cache, we propose RASTP (Representation-Aware Semantic Token Pruning), which directly prunes less informative tokens in the input sequence. Specifically, RASTP evaluates token importance by combining semantic saliency, measured via representation magnitude, and attention centrality, derived from cumulative attention weights. Since RASTP dynamically prunes low-information or irrelevant semantic tokens, experiments on three real-world Amazon datasets show that RASTP reduces training time by 26.7\%, while maintaining or slightly improving recommendation performance. The code has been open-sourced at https://github.com/Yuzt-zju/RASTP.

cs.IR↗

arXiv2025

An Integrated and Coherent Framework for Point Estimation and Hypothesis Testing with Concurrent Controls in Platform Trials

A platform trial with a master protocol provides an infrastructure to ethically and efficiently evaluate multiple treatment options in multiple diseases. Given that certain study drugs can enter or exit a platform trial, the randomization ratio is possible to change over time, and this potential modification is not necessarily dependent on accumulating outcomes data. It is recommended that the analysis should account for time periods with different randomization ratios, with possible approaches such as Inverse Probability of Treatment Weighting (IPTW) or a weighted approach by the time period. To guide practical implementation, we specifically investigate the relationship between these two estimators, and further derive an optimal estimator within this class to gain efficacy. Practical guidance is provided on how to construct estimators based on observed data to approximate this unknown weight. The connection between the proposed method and the weighted least squares is also studied. We conduct simulation studies to demonstrate that the proposed method can control type I error rate with a reduced estimation bias, and can also achieve satisfactory power and mean squared error (MSE) with computational efficiency. Another appealing feature of our framework is the ability to provide consistent conclusions for both point estimation and hypothesis testing. This is critical to the interpretation of clinical trial results. The proposed method is further applied to the Accelerating COVID-19 Therapeutic Interventions and Vaccines (ACTIV) platform trial.

stat.ME↗

arXiv2025

Device-Cloud Collaborative Correction for On-Device Recommendation

With the rapid development of recommendation models and device computing power, device-based recommendation has become an important research area due to its better real-time performance and privacy protection. Previously, Transformer-based sequential recommendation models have been widely applied in this field because they outperform Recurrent Neural Network (RNN)-based recommendation models in terms of performance. However, as the length of interaction sequences increases, Transformer-based models introduce significantly more space and computational overhead compared to RNN-based models, posing challenges for device-based recommendation. To balance real-time performance and high performance on devices, we propose Device-Cloud \underline{Co}llaborative \underline{Corr}ection Framework for On-Device \underline{Rec}ommendation (CoCorrRec). CoCorrRec uses a self-correction network (SCN) to correct parameters with extremely low time cost. By updating model parameters during testing based on the input token, it achieves performance comparable to current optimal but more complex Transformer-based models. Furthermore, to prevent SCN from overfitting, we design a global correction network (GCN) that processes hidden states uploaded from devices and provides a global correction solution. Extensive experiments on multiple datasets show that CoCorrRec outperforms existing Transformer-based and RNN-based device recommendation models in terms of performance, with fewer parameters and lower FLOPs, thereby achieving a balance between real-time performance and high efficiency.

cs.IR↗

arXiv2025

Semiparametric Weighted Spline Regression (SWSR) in Confirmatory Clinical Trials with Time-Varying Placebo Effects

In confirmatory Phase 3 clinical trials with recruitment over the years, the underlying placebo effect may follow an unknown temporal trend. Taking a clinical trial on Hidradenitis Suppurativa (HS) as an example, fluctuations or variabilities are common in HS-related endpoints, mainly due to the natural disease characteristics, variations of evaluation from different physicians, and standard of care evolvement. The adjustment of time-varying placebo effects receives some attention in adaptive clinical trials and platform trials, but is usually ignored in traditional non-adaptive designs. However, under the impact of such a time drift, some existing methods may not simultaneously control the type I error rate and achieve satisfactory power. In this article, we propose SWSR (Semiparametric Weighted Spline Regression) to estimate the treatment effect with B-splines to accommodate the time-varying placebo effects nonparametrically. Our method aims to achieve the following three objectives: a proper type I error rate control under varying settings, an overall high power to detect a potential treatment effect, and robustness to unknown time-varying placebo effects. Simulation studies and a case study provide supporting evidence. Those three key features make SWSR an appealing option to be pre-specified for practical confirmatory clinical trials. Supplemental materials, including the R code, additional simulation results and theoretical discussion, are available online.

stat.ME↗

arXiv2025

Collaboration of Large Language Models and Small Recommendation Models for Device-Cloud Recommendation

Large Language Models (LLMs) for Recommendation (LLM4Rec) is a promising research direction that has demonstrated exceptional performance in this field. However, its inability to capture real-time user preferences greatly limits the practical application of LLM4Rec because (i) LLMs are costly to train and infer frequently, and (ii) LLMs struggle to access real-time data (its large number of parameters poses an obstacle to deployment on devices). Fortunately, small recommendation models (SRMs) can effectively supplement these shortcomings of LLM4Rec diagrams by consuming minimal resources for frequent training and inference, and by conveniently accessing real-time data on devices. In light of this, we designed the Device-Cloud LLM-SRM Collaborative Recommendation Framework (LSC4Rec) under a device-cloud collaboration setting. LSC4Rec aims to integrate the advantages of both LLMs and SRMs, as well as the benefits of cloud and edge computing, achieving a complementary synergy. We enhance the practicability of LSC4Rec by designing three strategies: collaborative training, collaborative inference, and intelligent request. During training, LLM generates candidate lists to enhance the ranking ability of SRM in collaborative scenarios and enables SRM to update adaptively to capture real-time user interests. During inference, LLM and SRM are deployed on the cloud and on the device, respectively. LLM generates candidate lists and initial ranking results based on user behavior, and SRM get reranking results based on the candidate list, with final results integrating both LLM's and SRM's scores. The device determines whether a new candidate list is needed by comparing the consistency of the LLM's and SRM's sorted lists. Our comprehensive and extensive experimental analysis validates the effectiveness of each strategy in LSC4Rec.

cs.IR↗

arXiv2025

A general, flexible and harmonious framework to construct interpretable functions in regression analysis

An interpretable model or method has several appealing features, such as reliability to adversarial examples, transparency of decision-making, and communication facilitator. However, interpretability is a subjective concept, and even its definition can be diverse. The same model may be deemed as interpretable by a study team, but regarded as a black-box algorithm by another squad. Simplicity, accuracy and generalizability are some additional important aspects of evaluating interpretability. In this work, we present a general, flexible and harmonious framework to construct interpretable functions in regression analysis with a focus on continuous outcomes. We formulate a functional skeleton in light of users' expectations of interpretability. A new measure based on Mallows's $C_p$-statistic is proposed for model selection to balance approximation, generalizability, and interpretability. We apply this approach to derive a sample size formula in adaptive clinical trial designs to demonstrate the general workflow, and to explain operating characteristics in a Bayesian Go/No-Go paradigm to show the potential advantages of using meaningful intermediate variables. Generalization to categorical outcomes is illustrated in an example of hypothesis testing based on Fisher's exact test. A real data analysis of NHANES (National Health and Nutrition Examination Survey) is conducted to investigate relationships between some important laboratory measurements. We also discuss some extensions of this method.

stat.ME↗

arXiv2024

MergeNet: Knowledge Migration across Heterogeneous Models, Tasks, and Modalities

In this study, we focus on heterogeneous knowledge transfer across entirely different model architectures, tasks, and modalities. Existing knowledge transfer methods (e.g., backbone sharing, knowledge distillation) often hinge on shared elements within model structures or task-specific features/labels, limiting transfers to complex model types or tasks. To overcome these challenges, we present MergeNet, which learns to bridge the gap of parameter spaces of heterogeneous models, facilitating the direct interaction, extraction, and application of knowledge within these parameter spaces. The core mechanism of MergeNet lies in the parameter adapter, which operates by querying the source model's low-rank parameters and adeptly learning to identify and map parameters into the target model. MergeNet is learned alongside both models, allowing our framework to dynamically transfer and adapt knowledge relevant to the current stage, including the training trajectory knowledge of the source model. Extensive experiments on heterogeneous knowledge transfer demonstrate significant improvements in challenging settings, where representative approaches may falter or prove less applicable.

cs.LG↗

arXiv2024

Preliminary Evaluation of the Test-Time Training Layers in Recommendation System (Student Abstract)

This paper explores the application and effectiveness of Test-Time Training (TTT) layers in improving the performance of recommendation systems. We developed a model, TTT4Rec, utilizing TTT-Linear as the feature extraction layer. Our tests across multiple datasets indicate that TTT4Rec, as a base model, performs comparably or even surpasses other baseline models in similar environments.

cs.IR↗

arXiv2024

Semantic Codebook Learning for Dynamic Recommendation Models

Dynamic sequential recommendation (DSR) can generate model parameters based on user behavior to improve the personalization of sequential recommendation under various user preferences. However, it faces the challenges of large parameter search space and sparse and noisy user-item interactions, which reduces the applicability of the generated model parameters. The Semantic Codebook Learning for Dynamic Recommendation Models (SOLID) framework presents a significant advancement in DSR by effectively tackling these challenges. By transforming item sequences into semantic sequences and employing a dual parameter model, SOLID compresses the parameter generation search space and leverages homogeneity within the recommendation system. The introduction of the semantic metacode and semantic codebook, which stores disentangled item representations, ensures robust and accurate parameter generation. Extensive experiments demonstrates that SOLID consistently outperforms existing DSR, delivering more accurate, stable, and robust recommendations.

cs.IR↗

arXiv2023

A Class of Computational Methods to Reduce Selection Bias when Designing Phase 3 Clinical Trials

When designing confirmatory Phase 3 studies, one usually evaluates one or more efficacious and safe treatment option(s) based on data from previous studies. However, several retrospective research articles reported the phenomenon of ``diminished treatment effect in Phase 3'' based on many case studies. Even under basic assumptions, it was shown that the commonly used estimator could substantially overestimate the efficacy of selected group(s). As alternatives, we propose a class of computational methods to reduce estimation bias and mean squared error (MSE) with a broader scope of multiple treatment groups and flexibility to accommodate summary results by group as input. Based on simulation studies and a real data example, we provide practical implementation guidance for this class of methods under different scenarios. For more complicated problems, our framework can serve as a starting point with additional layers built in. Proposed methods can also be widely applied to other selection problems.

stat.ME↗

arXiv2023

Deep Neural Networks Guided Ensemble Learning for Point Estimation

In modern statistics, interests shift from pursuing the uniformly minimum variance unbiased estimator to reducing mean squared error (MSE) or residual squared error. Shrinkage based estimation and regression methods offer better prediction accuracy and improved interpretation. However, the characterization of such optimal statistics in terms of minimizing MSE remains open and challenging in many problems, for example estimating treatment effect in adaptive clinical trials with pre-planned modifications to design aspects based on accumulated data. From an alternative perspective, we propose a deep neural network based automatic method to construct an improved estimator from existing ones. Theoretical properties are studied to provide guidance on applicability of our estimator to seek potential improvement. Simulation studies demonstrate that the proposed method has considerable finite-sample efficiency gain as compared with several common estimators. In the Adaptive COVID-19 Treatment Trial (ACTT) as an important application, our ensemble estimator essentially contributes to a more ethical and efficient adaptive clinical trial with fewer patients enrolled. The proposed framework can be generally applied to various statistical problems, and can be served as a reference measure to guide statistical research.

stat.ME↗

arXiv2022

A practical Response Adaptive Block Randomization (RABR) design with analytic type I error protection

Response adaptive randomization (RAR) is appealing from methodological, ethical, and pragmatic perspectives in the sense that subjects are more likely to be randomized to better performing treatment groups based on accumulating data. However, applications of RAR in confirmatory drug clinical trials with multiple active arms are limited largely due to its complexity, and lack of control of randomization ratios to different treatment groups. To address the aforementioned issues, we propose a Response Adaptive Block Randomization (RABR) design allowing arbitrarily pre-specified randomization ratios for the control and high-performing groups to meet clinical trial objectives. We show the validity of the conventional unweighted test in RABR with a controlled type I error rate based on the weighted combination test for sample size adaptive design invoking no large sample approximation. The advantages of the proposed RABR in terms of robustly reaching target final sample size to meet regulatory requirements and increasing statistical power as compared with the popular Doubly Adaptive Biased Coin Design (DBCD) are demonstrated by statistical simulations and a practical clinical trial design example.

stat.ME↗

arXiv2022

Deep Historical Borrowing Framework to Prospectively and Simultaneously Synthesize Control Information in Confirmatory Clinical Trials with Multiple Endpoints

In current clinical trial development, historical information is receiving more attention as it provides utility beyond sample size calculation. Meta-analytic-predictive (MAP) priors and robust MAP priors have been proposed for prospectively borrowing historical data on a single endpoint. To simultaneously synthesize control information from multiple endpoints in confirmatory clinical trials, we propose to approximate posterior probabilities from a Bayesian hierarchical model and estimate critical values by deep learning to construct pre-specified strategies for hypothesis testing. This feature is important to ensure study integrity by establishing prospective decision functions before the trial conduct. Simulations are performed to show that our method properly controls family-wise error rate (FWER) and preserves power as compared with a typical practice of choosing constant critical values given a subset of null space. Satisfactory performance under prior-data conflict is also demonstrated. We further illustrate our method using a case study in Immunology.

stat.ME↗

arXiv2022

Finite-Sample Two-Group Composite Hypothesis Testing via Machine Learning

In the problem of composite hypothesis testing, identifying the potential uniformly most powerful (UMP) unbiased test is of great interest. Beyond typical hypothesis settings with exponential family, it is usually challenging to prove the existence and further construct such UMP unbiased tests with finite sample size. For example in the COVID-19 pandemic with limited previous assumptions on the treatment for investigation and the standard of care, adaptive clinical trials are appealing due to ethical considerations, and the ability to accommodate uncertainty while conducting the trial. Although several methods have been proposed to control type I error rates, how to find a more powerful hypothesis testing strategy is still an open question. Motivated by this problem, we propose an automatic framework of constructing test statistics and corresponding critical values via machine learning methods to enhance power in a finite sample. In this article, we particularly illustrate the performance using Deep Neural Networks (DNN) and discuss its advantages. Simulations and two case studies of adaptive designs demonstrate that our method is automatic, general and pre-specified to construct statistics with satisfactory power in finite-sample. Supplemental materials are available online including R code and an R shiny app.

stat.ME↗

arXiv2022

Optimizing Graphical Procedures for Multiplicity Control in a Confirmatory Clinical Trial via Deep Learning

In confirmatory clinical trials, it has been proposed to use a simple iterative graphical approach to construct and perform intersection hypotheses tests with a weighted Bonferroni-type procedure to control type I errors in the strong sense. Given Phase II study results or other prior knowledge, it is usually of main interest to find the optimal graph that maximizes a certain objective function in a future Phase III study. In this article, we evaluate the performance of two existing derivative-free constrained methods, and further propose a deep learning enhanced optimization framework. Our method numerically approximates the objective function via feedforward neural networks (FNNs) and then performs optimization with available gradient information. It can be constrained so that some features of the testing procedure are held fixed while optimizing over other features. Simulation studies show that our FNN-based approach has a better balance between robustness and time efficiency than some existing derivative-free constrained optimization algorithms. Compared to the traditional stochastic search method, our optimizer has moderate multiplicity adjusted power gain when the number of hypotheses is relatively large. We further apply it to a case study to illustrate how to optimize a multiple testing procedure with respect to a specific study objective.

stat.ME↗