arXiv subjects

Songkai Xue

Publications and source records attributed to Songkai Xue.

10 recordsLinked to original sources

arXiv2026

DeepBias: Adaptive In-depth Probing of Social Biases in LVLMs

While Large Vision-Language Models (LVLMs) demonstrate remarkable capabilities, they remain highly susceptible to embedded social biases. Existing bias evaluation protocols predominantly rely on static datasets, which provide only a superficial assessment, as their fixed test cases cannot adaptively evolve to measure the true depth and limits of model vulnerabilities. We introduce DeepBias, an adaptive framework for the in-depth probing of social biases in LVLMs with carefully designed agents. Our approach operates through a dynamic ''generation-evolution-probing'' loop. First, a generative ProposerAgent synthesizes test data and is iteratively updated via Direct Preference Optimization (DPO) based on the target LVLM's responses, exploring model-specific failure modes. Second, an autonomous skill-driven DiggerAgent rewrites each test data across multiple probing turns, adaptively selecting from a curated skill library of deepening and rewriting strategies. At each turn, this process is conditioned on the model's previous response, enabling progressively deeper biases to be exposed. Furthermore, we build a benchmark named DeepBiasBench using our framework. By employing an ensemble of five diverse state-of-the-art LVLMs as anchors, the benchmark captures vulnerabilities shared across architectures. Comprehensive experiments demonstrate the effectiveness of our framework and show that DeepBias provides a challenging benchmark for in-depth bias evaluation, establishing an evolutionary paradigm for LVLM safety assessment.

cs.CY↗

arXiv2025

Textual Data Bias Detection and Mitigation -- An Extensible Pipeline with Experimental Evaluation

Textual data used to train large language models (LLMs) exhibits multifaceted bias manifestations encompassing harmful language and skewed demographic distributions. Regulations such as the European AI Act require identifying and mitigating biases against protected groups in data, with the ultimate goal of preventing unfair model outputs. However, practical guidance and operationalization are lacking. We propose a comprehensive data bias detection and mitigation pipeline comprising four components that address two data bias types, namely representation bias and (explicit) stereotypes for a configurable sensitive attribute. First, we leverage LLM-generated word lists created based on quality criteria to detect relevant group labels. Second, representation bias is quantified using the Demographic Representation Score. Third, we detect and mitigate stereotypes using sociolinguistically informed filtering. Finally, we compensate representation bias through Grammar- and Context-Aware Counterfactual Data Augmentation. We conduct a two-fold evaluation using the examples of gender, religion and age. First, the effectiveness of each individual component on data debiasing is evaluated through human validation and baseline comparison. The findings demonstrate that we successfully reduce representation bias and (explicit) stereotypes in a text dataset. Second, the effect of data debiasing on model bias reduction is evaluated by bias benchmarking of several models (0.6B-8B parameters), fine-tuned on the debiased text dataset. This evaluation reveals that LLMs fine-tuned on debiased data do not consistently show improved performance on bias benchmarks, exposing critical gaps in current evaluation methodologies and highlighting the need for targeted data manipulation to address manifested model bias.

cs.CL↗

arXiv2025

Diverse Human Value Alignment for Large Language Models via Ethical Reasoning

Ensuring that Large Language Models (LLMs) align with the diverse and evolving human values across different regions and cultures remains a critical challenge in AI ethics. Current alignment approaches often yield superficial conformity rather than genuine ethical understanding, failing to address the complex, context-dependent nature of human values. In this paper, we propose a novel ethical reasoning paradigm for LLMs inspired by well-established ethical decision-making models, aiming at enhancing diverse human value alignment through deliberative ethical reasoning. Our framework consists of a structured five-step process, including contextual fact gathering, hierarchical social norm identification, option generation, multiple-lens ethical impact analysis, and reflection. This theory-grounded approach guides LLMs through an interpretable reasoning process that enhances their ability to understand regional specificities and perform nuanced ethical analysis, which can be implemented with either prompt engineering or supervised fine-tuning methods. We perform evaluations on the SafeWorld benchmark that specially designed for regional value alignment. Experimental results demonstrate our framework significantly improves LLM alignment with diverse human values compared to baseline methods, enabling more accurate social norm identification and more culturally appropriate reasoning. Our work provides a concrete pathway toward developing LLMs that align more effectively with the multifaceted values of global societies through interdisciplinary research.

cs.AI↗

arXiv2024

Distributionally Robust Performative Prediction

Performative prediction aims to model scenarios where predictive outcomes subsequently influence the very systems they target. The pursuit of a performative optimum (PO) -- minimizing performative risk -- is generally reliant on modeling of the distribution map, which characterizes how a deployed ML model alters the data distribution. Unfortunately, inevitable misspecification of the distribution map can lead to a poor approximation of the true PO. To address this issue, we introduce a novel framework of distributionally robust performative prediction and study a new solution concept termed as distributionally robust performative optimum (DRPO). We show provable guarantees for DRPO as a robust approximation to the true PO when the nominal distribution map is different from the actual one. Moreover, distributionally robust performative prediction can be reformulated as an augmented performative prediction problem, enabling efficient optimization. The experimental results demonstrate that DRPO offers potential advantages over traditional PO approach when the distribution map is misspecified at either micro- or macro-level.

cs.LG↗

arXiv2024

Minimax Regret Learning for Data with Heterogeneous Subgroups

Modern complex datasets often consist of various sub-populations with known group information. In the presence of sub-population heterogeneity, it is crucial to develop robust and generalizable learning methods that (1) can enjoy robust performance on each of the training populations, and (2) is generalizable to an unseen testing population. While various min-max formulations have been proposed to achieve (1) in the robust learning literature, their generalization to an unseen testing is less explored. Moreover, a general min-max formulation can be sensitive to the noise heterogeneity, and, in the extreme case, be degenerate such that a single high-noise population dominates. The min-max-regret (MMR) can mitigate these challenges. In this work, we consider a distribution-free robust hierarchical model for the generalization from multiple training populations to an unseen testing population. Under the robust hierarchical model, the empirical MMR can enjoy the regret guarantees on each of the training populations as well as the unseen testing population. We further specialize the general MMR framework to linear regression and generalized linear model, where we characterize the geometry of MMR and its distinction from other robust methods. We demonstrate the effectiveness of MMR through extensive simulation studies and an application to image recognition.

stat.ME↗

arXiv2023

On Uniform Consistency of Spectral Embeddings

In this paper, we study the convergence of the spectral embeddings obtained from the leading eigenvectors of certain similarity matrices to their population counterparts. We opt to study this convergence in a uniform (instead of average) sense and highlight the benefits of this choice. Using the Newton-Kantorovich Theorem and other tools from functional analysis, we first establish a general perturbation result for orthonormal bases of invariant subspaces. We then apply this general result to normalized spectral clustering. By tapping into the rich literature of Sobolev spaces and exploiting some concentration results in Hilbert spaces, we are able to prove a finite sample error bound on the uniform consistency error of the spectral embeddings in normalized spectral clustering.

math.ST↗

arXiv2023

Calibrated Data-Dependent Constraints with Exact Satisfaction Guarantees

We consider the task of training machine learning models with data-dependent constraints. Such constraints often arise as empirical versions of expected value constraints that enforce fairness or stability goals. We reformulate data-dependent constraints so that they are calibrated: enforcing the reformulated constraints guarantees that their expected value counterparts are satisfied with a user-prescribed probability. The resulting optimization problem is amendable to standard stochastic optimization algorithms, and we demonstrate the efficacy of our method on a fairness-sensitive classification task where we wish to guarantee the classifier's fairness (at test time).

stat.ML↗

arXiv2022

How does overparametrization affect performance on minority groups?

The benefits of overparameterization for the overall performance of modern machine learning (ML) models are well known. However, the effect of overparameterization at a more granular level of data subgroups is less understood. Recent empirical studies demonstrate encouraging results: (i) when groups are not known, overparameterized models trained with empirical risk minimization (ERM) perform better on minority groups; (ii) when groups are known, ERM on data subsampled to equalize group sizes yields state-of-the-art worst-group-accuracy in the overparameterized regime. In this paper, we complement these empirical studies with a theoretical investigation of the risk of overparameterized random feature models on minority groups. In a setting in which the regression functions for the majority and minority groups are different, we show that overparameterization always improves minority group performance.

cs.LG↗

arXiv2021

Statistical inference for individual fairness

As we rely on machine learning (ML) models to make more consequential decisions, the issue of ML models perpetuating or even exacerbating undesirable historical biases (e.g., gender and racial biases) has come to the fore of the public's attention. In this paper, we focus on the problem of detecting violations of individual fairness in ML models. We formalize the problem as measuring the susceptibility of ML models against a form of adversarial attack and develop a suite of inference tools for the adversarial cost function. The tools allow auditors to assess the individual fairness of ML models in a statistically-principled way: form confidence intervals for the worst-case performance differential between similar individuals and test hypotheses of model fairness with (asymptotic) non-coverage/Type I error rate control. We demonstrate the utility of our tools in a real-world case study.

stat.ML↗

arXiv2020

Auditing ML Models for Individual Bias and Unfairness

We consider the task of auditing ML models for individual bias/unfairness. We formalize the task in an optimization problem and develop a suite of inferential tools for the optimal value. Our tools permit us to obtain asymptotic confidence intervals and hypothesis tests that cover the target/control the Type I error rate exactly. To demonstrate the utility of our tools, we use them to reveal the gender and racial biases in Northpointe's COMPAS recidivism prediction instrument.

stat.ML↗