Arrow Research search

Author name cluster

Xin Yan

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.

10 papers
1 author row

Possible papers

10

NeurIPS Conference 2025 Conference Paper

Decoupling Contrastive Decoding: Robust Hallucination Mitigation in Multimodal Large Language Models

  • Wei Chen
  • Xin Yan
  • Bin Wen
  • Fan Yang
  • Tingting Gao
  • Di Zhang
  • Long Chen

Although multimodal large language models (MLLMs) exhibit remarkable reasoning capabilities on complex multimodal understanding tasks, they still suffer from the notorious 'hallucination' issue: generating outputs misaligned with obvious visual or factual evidence. Currently, training-based solutions, like direct preference optimization (DPO), leverage paired preference data to suppress hallucinations. However, they risk sacrificing general reasoning capabilities due to the likelihood displacement. Meanwhile, training-free solutions, like contrastive decoding, achieve this goal by subtracting the estimated hallucination pattern from a distorted input. Yet, these handcrafted perturbations (e. g. , add noise to images) may poorly capture authentic hallucination patterns. To avoid these weaknesses of existing methods, and realize ``robust'' hallucination mitigation (\ie, maintaining general reasoning performance), we propose a novel framework: Decoupling Contrastive Decoding (DCD). Specifically, DCD decouples the learning of positive and negative samples in preference datasets, and trains separate positive and negative image projections within the MLLM. The negative projection implicitly models real hallucination patterns, which enables vision-aware negative images in the contrastive decoding inference stage. Our DCD alleviates likelihood displacement by avoiding pairwise optimization and generalizes robustly without handcrafted degradation. Extensive ablations across hallucination benchmarks and general reasoning tasks demonstrate the effectiveness of DCD, \ie, it matches DPO’s hallucination suppression while preserving general capabilities and outperforms the handcrafted contrastive decoding methods.

PDF Details

NeurIPS Conference 2023 Conference Paper

Diffusion Model for Graph Inverse Problems: Towards Effective Source Localization on Complex Networks

  • Xin Yan
  • Hui Fang
  • Qiang He

Information diffusion problems, such as the spread of epidemics or rumors, are widespread in society. The inverse problems of graph diffusion, which involve locating the sources and identifying the paths of diffusion based on currently observed diffusion graphs, are crucial to controlling the spread of information. The problem of localizing the source of diffusion is highly ill-posed, presenting a major obstacle in accurately assessing the uncertainty involved. Besides, while comprehending how information diffuses through a graph is crucial, there is a scarcity of research on reconstructing the paths of information propagation. To tackle these challenges, we propose a probabilistic model called DDMSL (Discrete Diffusion Model for Source Localization). Our approach is based on the natural diffusion process of information propagation over complex networks, which can be formulated using a message-passing function. First, we model the forward diffusion of information using Markov chains. Then, we design a reversible residual network to construct a denoising-diffusion model in discrete space for both source localization and reconstruction of information diffusion paths. We provide rigorous theoretical guarantees for DDMSL and demonstrate its effectiveness through extensive experiments on five real-world datasets.

PDF Details

NeurIPS Conference 2021 Conference Paper

Clustering Effect of Adversarial Robust Models

  • Yang Bai
  • Xin Yan
  • Yong Jiang
  • Shu-Tao Xia
  • Yisen Wang

Adversarial robustness has received increasing attention along with the study of adversarial examples. So far, existing works show that robust models not only obtain robustness against various adversarial attacks but also boost the performance in some downstream tasks. However, the underlying mechanism of adversarial robustness is still not clear. In this paper, we interpret adversarial robustness from the perspective of linear components, and find that there exist some statistical properties for comprehensively robust models. Specifically, robust models show obvious hierarchical clustering effect on their linearized sub-networks, when removing or replacing all non-linear components (e. g. , batch normalization, maximum pooling, or activation layers). Based on these observations, we propose a novel understanding of adversarial robustness and apply it on more tasks including domain adaption and robustness boosting. Experimental evaluations demonstrate the rationality and superiority of our proposed clustering strategy. Our code is available at https: //github. com/bymavis/Adv Weight NeurIPS2021.

PDF Details

AAAI Conference 2019 Conference Paper

AI-Sketcher: A Deep Generative Model for Producing High-Quality Sketches

  • Nan Cao
  • Xin Yan
  • Yang Shi
  • Chaoran Chen

Sketch drawings play an important role in assisting humans in communication and creative design since ancient period. This situation has motivated the development of artificial intelligence (AI) techniques for automatically generating sketches based on user input. Sketch-RNN, a sequence-to-sequence variational autoencoder (VAE) model, was developed for this purpose and known as a state-of-the-art technique. However, it suffers from limitations, including the generation of lowquality results and its incapability to support multi-class generations. To address these issues, we introduced AI-Sketcher, a deep generative model for generating high-quality multiclass sketches. Our model improves drawing quality by employing a CNN-based autoencoder to capture the positional information of each stroke at the pixel level. It also introduces an influence layer to more precisely guide the generation of each stroke by directly referring to the training data. To support multi-class sketch generation, we provided a conditional vector that can help differentiate sketches under various classes. The proposed technique was evaluated based on two large-scale sketch datasets, and results demonstrated its power in generating high-quality sketches.

PDF Details

JMLR Journal 2012 Journal Article

Facilitating Score and Causal Inference Trees for Large Observational Studies

  • Xiaogang Su
  • Joseph Kang
  • Juanjuan Fan
  • Richard A. Levine
  • Xin Yan

Assessing treatment effects in observational studies is a multifaceted problem that not only involves heterogeneous mechanisms of how the treatment or cause is exposed to subjects, known as propensity, but also differential causal effects across sub-populations. We introduce a concept termed the facilitating score to account for both the confounding and interacting impacts of covariates on the treatment effect. Several approaches for estimating the facilitating score are discussed. In particular, we put forward a machine learning method, called causal inference tree (CIT), to provide a piecewise constant approximation of the facilitating score. With interpretable rules, CIT splits data in such a way that both the propensity and the treatment effect become more homogeneous within each resultant partition. Causal inference at different levels can be made on the basis of CIT. Together with an aggregated grouping procedure, CIT stratifies data into strata where causal effects can be conveniently assessed within each. Besides, a feasible way of predicting individual causal effects (ICE) is made available by aggregating ensemble CIT models. Both the stratified results and the estimated ICE provide an assessment of heterogeneity of causal effects and can be integrated for estimating the average causal effect (ACE). Mean square consistency of CIT is also established. We evaluate the performance of proposed methods with simulations and illustrate their use with the NSW data in Dehejia and Wahba (1999) where the objective is to assess the impact of a labor training program, the National Supported Work (NSW) demonstration, on post-intervention earnings. [abs] [ pdf ][ bib ] &copy JMLR 2012. ( edit, beta )

PDF Details