Author name cluster

Lan Wang

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

7 papers

2 author rows

AAAI Conference 2026 Conference Paper

Causally-Grounded Dual-Path Attention Intervention for Object Hallucination Mitigation in LVLMs

Liu Yu
Zhonghao Chen
Ping Kuang
Zhikun Feng
Fan Zhou
Lan Wang
Gillian Dobbie

Object hallucination remains a critical challenge in Large Vision-Language Models (LVLMs), where models generate content inconsistent with visual inputs. Existing language-decoder based mitigation approaches often regulate visual or textual attention independently, overlooking their interaction as two key causal factors. To address this, we propose Owl (Bi-mOdal attention reWeighting for Layer-wise hallucination mitigation), a causally-grounded framework that models hallucination process via a structural causal graph, treating decomposed visual and textual attentions as mediators. We introduce VTACR (Visual-to-Textual Attention Contribution Ratio), a novel metric that quantifies the modality contribution imbalance during decoding. Our analysis reveals that hallucinations frequently occur in low-VTACR scenarios, where textual priors dominate and visual grounding is weakened. To mitigate this, we design a fine-grained attention intervention mechanism that dynamically adjusts token- and layer-wise attention guided by VTACR signals. Finally, we propose a dual-path contrastive decoding strategy: one path emphasizes visually grounded predictions, while the other amplifies hallucinated ones -- letting visual truth shine and hallucination collapse. Experimental results on the POPE and CHAIR benchmarks show that Owl achieves significant hallucination reduction, setting a new SOTA in faithfulness while preserving vision-language understanding capability. Our code is available at https://github.com/CikZ2023/OWL

PDF Details DOI

TMLR Journal 2025 Journal Article

Doubly Robust Uncertainty Quantification for Quantile Treatment Effects in Sequential Decision Making

Yang Xu
Chengchun Shi
Shikai Luo
Lan Wang
Rui Song

We consider multi-stage sequential decision making, where the treatment at any stage may depend on the subject’s entire treatment and covariate history. We introduce a general framework for doubly robust uncertainty quantification for the quantiles of cumulative outcomes under a sequential treatment rule. While previous studies focused on mean effects, quantile effects offer unique insights into the distributional properties and are more robust for heavy-tailed outcomes. It is known that, doubly robust inference is significantly more challenging and largely unexplored for quantile treatment effects. More importantly, for mean effects, doubly robust estimation does not ensure doubly robust inference. Our approach first provides a doubly robust estimator for any quantile of interest based on pre-collected data, achieving semi-parametric efficiency. We then propose a novel doubly robust estimator for the asymptotic variance, enabling the construction of a doubly robust confidence interval. To overcome the challenges in parameter-dependent nuisance functions, we leverage deep conditional generative learning techniques. We demonstrate advantages of our approach via both simulation and real data from a short video platform. Additionally, we observe that our proposed approach leads to another mean effect estimator that outperforms existing estimators with heavy-tailed outcomes.

PDF Details

ICLR Conference 2024 Conference Paper

FairerCLIP: Debiasing CLIP's Zero-Shot Predictions using Functions in RKHSs

Sepehr Dehdashtian
Lan Wang
Vishnu Naresh Boddeti

Large pre-trained vision-language models such as CLIP provide compact and general-purpose representations of text and images that are demonstrably effective across multiple downstream zero-shot prediction tasks. However, owing to the nature of their training process, these models have the potential to 1) propagate or amplify societal biases in the training data and 2) learn to rely on spurious features. This paper proposes FairerCLIP, a general approach for making zero-shot predictions of CLIP more fair and robust to spurious correlations. We formulate the problem of jointly debiasing CLIP’s image and text representations in reproducing kernel Hilbert spaces (RKHSs), which affords multiple benefits: 1) Flexibility: Unlike existing approaches, which are specialized to either learn with or without ground-truth labels, FairerCLIP is adaptable to learning in both scenarios. 2) Ease of Optimization: FairerCLIP lends itself to an iterative optimization involving closed-form solvers, which leads to 4×-10× faster training than the existing methods. 3) Sample Efficiency: Under sample-limited conditions, FairerCLIP significantly outperforms baselines when they fail entirely. And, 4) Performance: Empirically, FairerCLIP achieves appreciable accuracy gains on benchmark fairness and spurious correlation datasets over their respective baselines.

Details

ICLR Conference 2021 Conference Paper

Generalization bounds via distillation

Daniel J. Hsu
Ziwei Ji
Matus Telgarsky
Lan Wang

This paper theoretically investigates the following empirical phenomenon: given a high-complexity network with poor generalization bounds, one can distill it into a network with nearly identical predictions but low complexity and vastly smaller generalization bounds. The main contribution is an analysis showing that the original network inherits this good generalization bound from its distillation, assuming the use of well-behaved data augmentation. This bound is presented both in an abstract and in a concrete form, the latter complemented by a reduction technique to handle modern computation graphs featuring convolutional layers, fully-connected layers, and skip connections, to name a few. To round out the story, a (looser) classical uniform convergence analysis of compression is also presented, as well as a variety of experiments on cifar and mnist demonstrating similar generalization performance between the original network and its distillation.

Details

JMLR Journal 2018 Journal Article

Sparse Concordance-assisted Learning for Optimal Treatment Decision

Shuhan Liang
Wenbin Lu
Rui Song
Lan Wang

To find optimal decision rule, Fan et al. (2016) proposed an innovative concordance-assisted learning algorithm which is based on maximum rank correlation estimator. It makes better use of the available information through pairwise comparison. However the objective function is discontinuous and computationally hard to optimize. In this paper, we consider a convex surrogate loss function to solve this problem. In addition, our algorithm ensures sparsity of decision rule and renders easy interpretation. We derive the $L_2$ error bound of the estimated coefficients under ultra-high dimension. Simulation results of various settings and application to STAR*D both illustrate that the proposed method can still estimate optimal treatment regime successfully when the number of covariates is large. [abs] [ pdf ][ bib ] &copy JMLR 2018. ( edit, beta )

PDF Details

JMLR Journal 2016 Journal Article

A Consistent Information Criterion for Support Vector Machines in Diverging Model Spaces

Xiang Zhang
Yichao Wu
Lan Wang
Runze Li

Information criteria have been popularly used in model selection and proved to possess nice theoretical properties. For classification, Claeskens et al. (2880) proposed support vector machine information criterion for feature selection and provided encouraging numerical evidence. Yet no theoretical justification was given there. This work aims to fill the gap and to provide some theoretical justifications for support vector machine information criterion in both fixed and diverging model spaces. We first derive a uniform convergence rate for the support vector machine solution and then show that a modification of the support vector machine information criterion achieves model selection consistency even when the number of features diverges at an exponential rate of the sample size. This consistency result can be further applied to selecting the optimal tuning parameter for various penalized support vector machine methods. Finite-sample performance of the proposed information criterion is investigated using Monte Carlo studies and one real-world gene selection problem. [abs] [ pdf ][ bib ] &copy JMLR 2016. ( edit, beta )

PDF Details

JMLR Journal 2016 Journal Article

An Error Bound for L1-norm Support Vector Machine Coefficients in Ultra-high Dimension

Bo Peng
Lan Wang
Yichao Wu

Comparing with the standard $L_2$-norm support vector machine (SVM), the $L_1$-norm SVM enjoys the nice property of simultaneously preforming classification and feature selection. In this paper, we investigate the statistical performance of $L_1$-norm SVM in ultra-high dimension, where the number of features $p$ grows at an exponential rate of the sample size $n$. Different from existing theory for SVM which has been mainly focused on the generalization error rates and empirical risk, we study the asymptotic behavior of the coefficients of $L_1$-norm SVM. Our analysis reveals that the $L_1$-norm SVM coefficients achieve near oracle rate, that is, with high probability, the $L_2$ error bound of the estimated $L_1$-norm SVM coefficients is of order $O_p(\sqrt{q\log p/n})$, where $q$ is the number of features with nonzero coefficients. Furthermore, we show that if the $L_1$-norm SVM is used as an initial value for a recently proposed algorithm for solving non- convex penalized SVM (Zhang et al., 2016b), then in two iterative steps it is guaranteed to produce an estimator that possesses the oracle property in ultra-high dimension, which in particular implies that with probability approaching one the zero coefficients are estimated as exactly zero. Simulation studies demonstrate the fine performance of $L_1$-norm SVM as a sparse classifier and its effectiveness to be utilized to solve non-convex penalized SVM problems in high dimension. [abs] [ pdf ][ bib ] &copy JMLR 2016. ( edit, beta )

PDF Details