arXiv subjects

Ke Zhu

Publications and source records attributed to Ke Zhu.

At least 37 records · Page 2Linked to original sources

arXiv2025

Design-based theory for causal inference

Causal inference, as a major research area in statistics and data science, plays a central role across diverse fields such as medicine, economics, education, and the social sciences. Design-based causal inference begins with randomized experiments and emphasizes conducting statistical inference by leveraging the known randomization mechanism, thereby enabling identification and estimation of causal effects under weak model dependence. Grounded in the seminal works of Fisher and Neyman, this paradigm has evolved to include various design strategies, such as stratified randomization and rerandomization, and analytical methods including Fisher randomization tests, Neyman-style asymptotic inference, and regression adjustment. In recent years, with the emergence of complex settings involving high-dimensional data, individual noncompliance, and network interference, design-based causal inference has witnessed remarkable theoretical and methodological advances. This paper provides a systematic review of recent progress in this field, focusing on covariate-balanced randomization designs, design-based statistical inference methods, and their extensions to high-dimensional, noncompliance, and network interference scenarios. It concludes with a comprehensive perspective on future directions for the theoretical development and practical applications of causal inference.

stat.ME↗

arXiv2025

COADVISE: Covariate Adjustment with Variable Selection in Randomized Controlled Trials

Adjusting for covariates in randomized controlled trials can enhance the credibility and efficiency of treatment effect estimation. However, handling numerous covariates and their complex (non-linear) transformations poses a challenge. Motivated by the case study of the Best Apnea Interventions for Research (BestAIR) trial data from the National Sleep Research Resource (NSRR), where the number of covariates (p=114) is comparable to the sample size (N=196), we propose a principled Covariate Adjustment with Variable Selection (COADVISE) framework. COADVISE enables variable selection for covariates most relevant to the outcome while accommodating both linear and nonlinear adjustments. This framework ensures consistent estimates with improved efficiency over unadjusted estimators and provides robust variance estimation, even under outcome model misspecification. We demonstrate efficiency gains through theoretical analysis, extensive simulations, and a re-analysis of the BestAIR trial data to compare alternative variable selection strategies, offering cautionary recommendations. A user-friendly R package, Coadvise, is available to facilitate practical implementation.

stat.ME↗

arXiv2025

Images Speak Louder Than Scores: Failure Mode Escape for Enhancing Generative Quality

Diffusion models have achieved remarkable progress in class-to-image generation. However, we observe that despite impressive FID scores, state-of-the-art models often generate distorted or low-quality images, especially in certain classes. This gap arises because FID evaluates global distribution alignment, while ignoring the perceptual quality of individual samples. We further examine the role of CFG, a common technique used to enhance generation quality. While effective in improving metrics and suppressing outliers, CFG can introduce distribution shift and visual artifacts due to its misalignment with both training objectives and user expectations. In this work, we propose FaME, a training-free and inference-efficient method for improving perceptual quality. FaME uses an image quality assessment model to identify low-quality generations and stores their sampling trajectories. These failure modes are then used as negative guidance to steer future sampling away from poor-quality regions. Experiments on ImageNet demonstrate that FaME brings consistent improvements in visual quality without compromising FID. FaME also shows the potential to be extended to improve text-to-image generation.

cs.CV↗

arXiv2025

X-Intelligence 3.0: Training and Evaluating Reasoning LLM for Semiconductor Display

Large language models (LLMs) have recently achieved significant advances in reasoning and demonstrated their advantages in solving challenging problems. Yet, their effectiveness in the semiconductor display industry remains limited due to a lack of domain-specific training and expertise. To bridge this gap, we present X-Intelligence 3.0, the first high-performance reasoning model specifically developed for the semiconductor display industry. This model is designed to deliver expert-level understanding and reasoning for the industry's complex challenges. Leveraging a carefully curated industry knowledge base, the model undergoes supervised fine-tuning and reinforcement learning to enhance its reasoning and comprehension capabilities. To further accelerate development, we implemented an automated evaluation framework that simulates expert-level assessments. We also integrated a domain-specific retrieval-augmented generation (RAG) mechanism, resulting in notable performance gains on benchmark datasets. Despite its relatively compact size of 32 billion parameters, X-Intelligence 3.0 outperforms SOTA DeepSeek-R1-671B across multiple evaluations. This demonstrates its exceptional efficiency and establishes it as a powerful solution to the longstanding reasoning challenges faced by the semiconductor display industry.

cs.CL↗

arXiv2025

Quantization without Tears

Deep neural networks, while achieving remarkable success across diverse tasks, demand significant resources, including computation, GPU memory, bandwidth, storage, and energy. Network quantization, as a standard compression and acceleration technique, reduces storage costs and enables potential inference acceleration by discretizing network weights and activations into a finite set of integer values. However, current quantization methods are often complex and sensitive, requiring extensive task-specific hyperparameters, where even a single misconfiguration can impair model performance, limiting generality across different models and tasks. In this paper, we propose Quantization without Tears (QwT), a method that simultaneously achieves quantization speed, accuracy, simplicity, and generality. The key insight of QwT is to incorporate a lightweight additional structure into the quantized network to mitigate information loss during quantization. This structure consists solely of a small set of linear layers, keeping the method simple and efficient. More importantly, it provides a closed-form solution, allowing us to improve accuracy effortlessly under 2 minutes. Extensive experiments across various vision, language, and multimodal tasks demonstrate that QwT is both highly effective and versatile. In fact, our approach offers a robust solution for network quantization that combines simplicity, accuracy, and adaptability, which provides new insights for the design of novel quantization paradigms. The code is publicly available at https://github.com/wujx2001/QwT

cs.CV↗

arXiv2025

Doubly Robust Fusion of Many Treatments for Policy Learning

Individualized treatment rules/recommendations (ITRs) aim to improve patient outcomes by tailoring treatments to the characteristics of each individual. However, when there are many treatment groups, existing methods face significant challenges due to data sparsity within treatment groups and highly unbalanced covariate distributions across groups. To address these challenges, we propose a novel calibration-weighted treatment fusion procedure that robustly balances covariates across treatment groups and fuses similar treatments using a penalized working model. The fusion procedure ensures the recovery of latent treatment group structures when either the calibration model or the outcome model is correctly specified. In the fused treatment space, practitioners can seamlessly apply state-of-the-art ITR learning methods with the flexibility to utilize a subset of covariates, thereby achieving robustness while addressing practical concerns such as fairness. We establish theoretical guarantees, including consistency, the oracle property of treatment fusion, and regret bounds when integrated with multi-armed ITR learning methods such as policy trees. Simulation studies show superior group recovery and policy value compared to existing approaches. We illustrate the practical utility of our method using a nationwide electronic health record-derived de-identified database containing data from patients with Chronic Lymphocytic Leukemia and Small Lymphocytic Lymphoma.

stat.ME↗

arXiv2025

Sharp variance estimator and causal bootstrap in stratified randomized experiments

Randomized experiments are the gold standard for estimating treatment effects, and randomization serves as a reasoned basis for inference. In widely used stratified randomized experiments, randomization-based finite-population asymptotic theory enables valid inference for the average treatment effect, relying on normal approximation and a Neyman-type conservative variance estimator. However, when the sample size is small or the outcomes are skewed, the Neyman-type variance estimator may become overly conservative, and the normal approximation can fail. To address these issues, we propose a sharp variance estimator and two causal bootstrap methods to more accurately approximate the sampling distribution of the weighted difference-in-means estimator in stratified randomized experiments. The first causal bootstrap procedure is based on rank-preserving imputation and we prove its second-order refinement over normal approximation. The second causal bootstrap procedure is based on constant-treatment-effect imputation and is further applicable in paired experiments. In contrast to traditional bootstrap methods, where randomness originates from hypothetical super-population sampling, our analysis for the proposed causal bootstrap is randomization-based, relying solely on the randomness of treatment assignment in randomized experiments. Numerical studies and two real data applications demonstrate advantages of our proposed methods in finite samples. The \texttt{R} package \texttt{CausalBootstrap} implementing our method is publicly available.

math.ST↗

arXiv2025

Enhancing Statistical Validity and Power in Hybrid Controlled Trials: A Randomization Inference Approach with Conformal Selective Borrowing

External controls from historical trials or observational data can augment randomized controlled trials when large-scale randomization is impractical or unethical, such as in drug evaluation for rare diseases. However, non-randomized external controls can introduce biases, and existing Bayesian and frequentist methods may inflate the type I error rate, particularly in small-sample trials where external data borrowing is most critical. To address these challenges, we propose a randomization inference framework that ensures finite-sample exact and model-free type I error rate control, adhering to the "analyze as you randomize" principle to safeguard against hidden biases. Recognizing that biased external controls reduce the power of randomization tests, we leverage conformal inference to develop an individualized test-then-pool procedure that selectively borrows comparable external controls to improve power. Our approach incorporates selection uncertainty into randomization tests, providing valid post-selection inference. Additionally, we propose an adaptive procedure to optimize the selection threshold by minimizing the mean squared error across a class of estimators encompassing both no-borrowing and full-borrowing approaches. The proposed methods are supported by non-asymptotic theoretical analysis, validated through simulations, and applied to a randomized lung cancer trial that integrates external controls from the National Cancer Database.

stat.ME↗

arXiv2025

On Data Synthesis and Post-training for Visual Abstract Reasoning

This paper is a pioneering work attempting to address abstract visual reasoning (AVR) problems for large vision-language models (VLMs). We make a common LLaVA-NeXT 7B model capable of perceiving and reasoning about specific AVR problems, surpassing both open-sourced (e.g., Qwen-2-VL-72B) and closed-sourced powerful VLMs (e.g., GPT-4o) with significant margin. This is a great breakthrough since almost all previous VLMs fail or show nearly random performance on representative AVR benchmarks. Our key success is our innovative data synthesis and post-training process, aiming to fully relieve the task difficulty and elicit the model to learn, step by step. Our 7B model is also shown to be behave well on AVR without sacrificing common multimodal comprehension abilities. We hope our paper could serve as an early effort in this area and would inspire further research in abstract visual reasoning.

cs.CV↗

arXiv2025

Rejoinder to Reader Reaction "On exact randomization-based covariate-adjusted confidence intervals" by Jacob Fiksel

We applaud Fiksel (2024) for their valuable contributions to randomization-based inference, particularly their work on inverting the Fisher randomization test (FRT) to construct confidence intervals using the covariate-adjusted test statistic. FRT is advocated by many scholars because it produces finite-sample exact p-values for any test statistic and can be easily adopted for any experimental design (Rosenberger et al., 2019; Proschan and Dodd, 2019; Young, 2019; Bind and Rubin, 2020). By inverting FRTs, we can construct the randomization-based confidence interval (RBCI). To the best of our knowledge, Zhu and Liu (2023) are the first to analytically invert the FRT for the difference-in-means statistic. Fiksel (2024) extended this analytical approach to the covariate-adjusted statistic, producing a monotonic p-value function under certain conditions. In this rejoinder, we propose an analytical approach to invert the FRT for test statistics that yield a non-monotonic p-value function, with the studentized t-statistic as an important special case. Exploiting our analytical approach, we can recover the non-monotonic p-value function and construct RBCI based on the studentized t-statistic. The RBCI generated by the proposed analytical approach is guaranteed to achieve the desired coverage probability and resolve the contradiction between Luo et al. (2021) and Wu and Ding (2021). Simulation results validate our findings and demonstrate that our method is also computationally efficient.

stat.ME↗

arXiv2025

Tensor dynamic conditional correlation model: A new way to pursuit "Holy Grail of investing"

Style investing creates asset classes (or the so-called "styles") with low correlations, aligning well with the principle of "Holy Grail of investing" in terms of portfolio selection. The returns of styles naturally form a tensor-valued time series, which requires new tools for studying the dynamics of the conditional correlation matrix to facilitate the aforementioned principle. Towards this goal, we introduce a new tensor dynamic conditional correlation (TDCC) model, which is based on two novel treatments: trace-normalization and dimension-normalization. These two normalizations adapt to the tensor nature of the data, and they are necessary except when the tensor data reduce to vector data. Moreover, we provide an easy-to-implement estimation procedure for the TDCC model, and examine its finite sample performance by simulations. Finally, we assess the usefulness of the TDCC model in international portfolio selection across ten global markets and in large portfolio selection for 1800 stocks from the Chinese stock market.

q-fin.PM↗

arXiv2025

All You Need in Knowledge Distillation Is a Tailored Coordinate System

Knowledge Distillation (KD) is essential in transferring dark knowledge from a large teacher to a small student network, such that the student can be much more efficient than the teacher but with comparable accuracy. Existing KD methods, however, rely on a large teacher trained specifically for the target task, which is both very inflexible and inefficient. In this paper, we argue that a SSL-pretrained model can effectively act as the teacher and its dark knowledge can be captured by the coordinate system or linear subspace where the features lie in. We then need only one forward pass of the teacher, and then tailor the coordinate system (TCS) for the student network. Our TCS method is teacher-free and applies to diverse architectures, works well for KD and practical few-shot learning, and allows cross-architecture distillation with large capacity gap. Experiments show that TCS achieves significantly higher accuracy than state-of-the-art KD methods, while only requiring roughly half of their training time and GPU memory costs.

cs.CV↗

arXiv2024

Continual SFT Matches Multimodal RLHF with Negative Supervision

Multimodal RLHF usually happens after supervised finetuning (SFT) stage to continually improve vision-language models' (VLMs) comprehension. Conventional wisdom holds its superiority over continual SFT during this preference alignment stage. In this paper, we observe that the inherent value of multimodal RLHF lies in its negative supervision, the logit of the rejected responses. We thus propose a novel negative supervised finetuning (nSFT) approach that fully excavates these information resided. Our nSFT disentangles this negative supervision in RLHF paradigm, and continually aligns VLMs with a simple SFT loss. This is more memory efficient than multimodal RLHF where 2 (e.g., DPO) or 4 (e.g., PPO) large VLMs are strictly required. The effectiveness of nSFT is rigorously proved by comparing it with various multimodal RLHF approaches, across different dataset sources, base VLMs and evaluation metrics. Besides, fruitful of ablations are provided to support our hypothesis. We hope this paper will stimulate further research to properly align large vision language models.

cs.LG↗

arXiv2024

Imputation-based randomization tests for randomized experiments with interference

The presence of interference renders classic Fisher randomization tests infeasible due to nuisance unknowns. To address this issue, we propose imputing the nuisance unknowns and computing Fisher randomization p-values multiple times, then averaging them. We term this approach the imputation-based randomization test and provide theoretical results on its asymptotic validity. Our method leverages the merits of randomization and the flexibility of the Bayesian framework: for multiple imputations, we can either employ the empirical distribution of observed outcomes to achieve robustness against model mis-specification or utilize a parametric model to incorporate prior information. Simulation results demonstrate that our method effectively controls the type I error rate and significantly enhances the testing power compared to existing randomization tests for randomized experiments with interference. We apply our method to a two-round randomized experiment with multiple treatments and one-way interference, where existing randomization tests exhibit limited power.

stat.ME↗

arXiv2024

Computable, obstructed Morse homology for clean intersections

In this paper, we develop a method to compute the Morse homology of a manifold when descending manifolds and ascending manifolds intersect cleanly, but not necessarily transversely. While obstruction bundle gluing defined by Hutchings and Taubes is a computable tool to handle non-transverse intersections, it has only been developed for specific cases. In contrast, most virtual techniques apply to general cases but lack computational efficiency. To address this, we construct minimal semi-global Kuranishi structures for the moduli spaces of Morse trajectories, which generalize obstruction bundle gluing while maintaining its computability feature. Through this construction, we obtain iterated gluing equals simultaneous gluing.

math.SG↗

arXiv2024

Self-Supervised Visual Preference Alignment

This paper makes the first attempt towards unsupervised preference alignment in Vision-Language Models (VLMs). We generate chosen and rejected responses with regard to the original and augmented image pairs, and conduct preference alignment with direct preference optimization. It is based on a core idea: properly designed augmentation to the image input will induce VLM to generate false but hard negative responses, which helps the model to learn from and produce more robust and powerful answers. The whole pipeline no longer hinges on supervision from GPT-4 or human involvement during alignment, and is highly efficient with few lines of code. With only 8k randomly sampled unsupervised data, it achieves 90\% relative score to GPT-4 on complex reasoning in LLaVA-Bench, and improves LLaVA-7B/13B by 6.7\%/5.6\% score on complex multi-modal benchmark MM-Vet. Visualizations shows its improved ability to align with user-intentions. A series of ablations are firmly conducted to reveal the latent mechanism of the approach, which also indicates its potential towards further scaling. Code are available in https://github.com/Kevinz-code/SeVa.

cs.CV↗

arXiv2024

Enhancement of price trend trading strategies via image-induced importance weights

We open up the "black-box" to identify the predictive general price patterns in price chart images via the deep learning image analysis techniques. Our identified price patterns lead to the construction of image-induced importance (triple-I) weights, which are applied to weighted moving average the existing price trend trading signals according to their level of importance in predicting price movements. From an extensive empirical analysis on the Chinese stock market, we show that the triple-I weighting scheme can significantly enhance the price trend trading signals for proposing portfolios, with a thoughtful robustness study in terms of network specifications, image structures, and stock sizes. Moreover, we demonstrate that the triple-I weighting scheme is able to propose long-term portfolios from a time-scale transfer learning, enhance the news-based trading strategies through a non-technical transfer learning, and increase the overall strength of numerous trading rules for portfolio selection.

q-fin.PM↗

arXiv2024

Generalized Maximum Likelihood Estimation for Perspective-n-Point Problem

The Perspective-n-Point (PnP) problem has been widely studied in the literature and applied in various vision-based pose estimation scenarios. However, existing methods ignore the anisotropy uncertainty of observations, as demonstrated in several real-world datasets in this paper. This oversight may lead to suboptimal and inaccurate estimation, particularly in the presence of noisy observations. To this end, we propose a generalized maximum likelihood PnP solver, named GMLPnP, that minimizes the determinant criterion by iterating the GLS procedure to estimate the pose and uncertainty simultaneously. Further, the proposed method is decoupled from the camera model. Results of synthetic and real experiments show that our method achieves better accuracy in common pose estimation scenarios, GMLPnP improves rotation/translation accuracy by 4.7%/2.0% on TUM-RGBD and 18.6%/18.4% on KITTI-360 dataset compared to the best baseline. It is more accurate under very noisy observations in a vision-based UAV localization task, outperforming the best baseline by 34.4% in translation estimation accuracy.

cs.CV↗