首页/论文/2107.02112

arXiv 2107.021122021-07-05cs.CVcs.MM

Recovering the Unbiased Scene Graphs from the Biased Ones

National University of Singapore（新加坡国立大学）；ByteDance AI Lab（字节跳动人工智能实验室）

Meng-Jiun Chiou, Henghui Ding, Hanshu Yan, Changhu Wang, Roger Zimmermann, Jiashi Feng

备注：: Accepted by ACMMM 2021. Source code will be available at https://github.com/coldmanck/recovering-unbiased-scene-graphs

阅读 PDF ↗

arXiv 页面 ↗HTML 原文 ↗

英文摘要相关论文

英文摘要：

Given input images, scene graph generation (SGG) aims to produce comprehensive, graphical representations describing visual relationships among salient objects. Recently, more efforts have been paid to the long tail problem in SGG; however, the imbalance in the fraction of missing labels of different classes, or reporting bias, exacerbating the long tail is rarely considered and cannot be solved by the existing debiasing methods. In this paper we show that, due to the missing labels, SGG can be viewed as a "Learning from Positive and Unlabeled data" (PU learning) problem, where the reporting bias can be removed by recovering the unbiased probabilities from the biased ones by utilizing label frequencies, i.e., the per-class fraction of labeled, positive examples in all the positive examples. To obtain accurate label frequency estimates, we propose Dynamic Label Frequency Estimation (DLFE) to take advantage of training-time data augmentation and average over multiple training iterations to introduce more valid examples. Extensive experiments show that DLFE is more effective in estimating label frequencies than a naive variant of the traditional estimate, and DLFE significantly alleviates the long tail and achieves state-of-the-art debiasing performance on the VG dataset. We also show qualitatively that SGG models with DLFE produce prominently more balanced and unbiased scene graphs.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

阅读原文

阅读 PDF ↗

arXiv 页面 ↗HTML 原文 ↗

本文内容

英文摘要相关论文

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

继续探索

相关论文

CogTree: Cognition Tree Loss for Unbiased Scene Graph Generation

Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）；Peking University（北京大学）；University of Adelaide（阿德莱德大学）；Beihang University（北京航空航天大学）

Scene graphs are semantic abstraction of images that encourage visual understanding and reasoning. However, the performance of Scene Graph Generation (SGG) is unsatisfactory when faced with biased data in real-world scenarios. Conventional debiasing research mainly studies from the view of balancing data distribution or learning unbiased models and representations, ignoring the correlations among the biased classes. In this work, we analyze this problem from a novel cognition perspective: automatically building a hierarchical cognitive structure from the biased predictions and navigating that …

2021-06-08cs.CLcs.CVcs.MM

从有偏数据集训练无偏扩散模型

Training Unbiased Diffusion Models From Biased Dataset

KAIST（韩国科学技术院）

针对扩散模型受有偏训练数据影响的问题，提出时间依赖重要性重加权与分数校正目标，实现向无偏分布收敛，并在多数据集多种偏差设置下优于基线。

2024-03-02cs.CVcs.LG

无偏监督对比学习

Unbiased Supervised Contrastive Learning

University of Turin（都灵大学）；LTCI, Télécom Paris, IP Paris（巴黎理工学院巴黎电信学院 LTCI 实验室）

针对数据集偏差导致对比学习失效的问题，提出基于间隔的理论框架，推导epsilon-SupInfoNCE损失与FairKL去偏正则项，在多视觉数据集上实现SOTA去偏性能。

2023-05-04cs.CVcs.LGstat.ML

Unravelling the Effect of Image Distortions for Biased Prediction of Pre-trained Face Recognition Models

IIIT-Delhi（德里印度理工学院）；IIT Jodhpur（印度理工学院焦特布尔分校）

Identifying and mitigating bias in deep learning algorithms has gained significant popularity in the past few years due to its impact on the society. Researchers argue that models trained on balanced datasets with good representation provide equal and unbiased performance across subgroups. However, \textit{can seemingly unbiased pre-trained model become biased when input data undergoes certain distortions?} For the first time, we attempt to answer this question in the context of face recognition. We provide a systematic analysis to evaluate the performance of four state-of-the-art deep face re…

2021-08-14cs.CV

Towards Learning an Unbiased Classifier from Biased Data via Conditional Adversarial Debiasing

Friedrich Schiller University Jena（弗里德里希·席勒大学耶拿分校）；Technische Universität Berlin（柏林工业大学）

Bias in classifiers is a severe issue of modern deep learning methods, especially for their application in safety- and security-critical areas. Often, the bias of a classifier is a direct consequence of a bias in the training dataset, frequently caused by the co-occurrence of relevant features and irrelevant ones. To mitigate this issue, we require learning algorithms that prevent the propagation of bias from the dataset into the classifier. We present a novel adversarial debiasing method, which addresses a feature that is spuriously connected to the labels of training images but statistically…

2021-03-10cs.AIcs.CV

从有偏训练中学习无偏场景图生成

Unbiased Scene Graph Generation from Biased Training

Nanyang Technological University（南洋理工大学）；Damo Academy, Alibaba Group（阿里达摩院）；Renmin University of China（中国人民大学）；Tsinghua University（清华大学）

针对场景图生成中的训练偏差问题，提出基于因果推断和总直接效应的去偏框架，在Visual Genome上显著提升无偏性能。

2025-10-26cs.CVcs.LG

更多相关论文收起更多论文

Exploring Biases and Prejudice of Facial Synthesis via Semantic Latent Space

Australian National University（澳大利亚国立大学）

Deep learning (DL) models are widely used to provide a more convenient and smarter life. However, biased algorithms will negatively influence us. For instance, groups targeted by biased algorithms will feel unfairly treated and even fearful of negative consequences of these biases. This work targets biased generative models' behaviors, identifying the cause of the biases and eliminating them. We can (as expected) conclude that biased data causes biased predictions of face frontalization models. Varying the proportions of male and female faces in the training data can have a substantial effect …

2021-08-23cs.CVcs.CYcs.LG

Towards Unbiased Label Distribution Learning for Facial Pose Estimation Using Anisotropic Spherical Gaussian

Purdue University（普渡大学）；Rochester Institute of Technology（罗切斯特理工学院）；Meta AI

Facial pose estimation refers to the task of predicting face orientation from a single RGB image. It is an important research topic with a wide range of applications in computer vision. Label distribution learning (LDL) based methods have been recently proposed for facial pose estimation, which achieve promising results. However, there are two major issues in existing LDL methods. First, the expectations of label distributions are biased, leading to a biased pose estimation. Second, fixed distribution parameters are applied for all learning samples, severely limiting the model capability. In t…

2022-08-19cs.CV

Label noise in segmentation networks : mitigation must deal with bias

École Polytechnique de Montréal（蒙特利尔理工学院）

Imperfect labels limit the quality of predictions learned by deep neural networks. This is particularly relevant in medical image segmentation, where reference annotations are difficult to collect and vary significantly even across expert annotators. Prior work on mitigating label noise focused on simple models of mostly uniform noise. In this work, we explore biased and unbiased errors artificially introduced to brain tumour annotations on MRI data. We found that supervised and semi-supervised segmentation methods are robust or fairly robust to unbiased errors but sensitive to biased errors. …

2021-07-05cs.CVcs.LG

通过可学习数据增强实现模型去偏

Model Debiasing by Learnable Data Augmentation

Istituto Italiano di Tecnologia（意大利技术研究院）；University of Verona（维罗纳大学）

针对未知偏差导致模型学习虚假关联的问题，提出两阶段可学习数据增强流程，先识别有偏样本，再学习混合参数进行正则化，在多类数据集上取得最优分类准确率并提升泛化性。

2024-08-09cs.CVcs.LG

Context-aware Mixture-of-Experts for Unbiased Scene Graph Generation

AIRS, CUHK-Shenzhen（香港中文大学（深圳）人工智能与机器人研究院）；McGill University（麦吉尔大学）

Scene graph generation (SGG) has gained tremendous progress in recent years. However, its underlying long-tailed distribution of predicate classes is a challenging problem. For extremely unbalanced predicate distributions, existing approaches usually construct complicated context encoders to extract the intrinsic relevance of scene context to predicates and complex networks to improve the learning ability of network models for highly imbalanced predicate distributions. To address the unbiased SGG problem, we introduce a simple yet effective method dubbed Context-Aware Mixture-of-Experts (CAME)…

2023-01-01cs.CV

解耦稳定性与可塑性以实现多模态测试时适应

Decoupling Stability and Plasticity for Multi-Modal Test-Time Adaptation

Zhejiang University（浙江大学）

本文提出DASP框架，通过解耦多模态测试时适应中的稳定性与可塑性，解决现有方法在无偏模态中的负迁移和有偏模态中的灾难性遗忘问题。

2026-05-14cs.AIcs.CV

Unbiased Scene Graph Generation using Predicate Similarities

The University of Tokyo（东京大学）

Scene Graphs are widely applied in computer vision as a graphical representation of relationships between objects shown in images. However, these applications have not yet reached a practical stage of development owing to biased training caused by long-tailed predicate distributions. In recent years, many studies have tackled this problem. In contrast, relatively few works have considered predicate similarities as a unique dataset feature which also leads to the biased prediction. Due to the feature, infrequent predicates (e.g., parked on, covered in) are easily misclassified as closely-relate…

2022-10-03cs.CV

基于语义原型学习的全景场景图生成

Panoptic Scene Graph Generation with Semantics-Prototype Learning

National University of Singapore（新加坡国立大学）；The University of Sydney（悉尼大学）；Zhejiang University（浙江大学）；Hangzhou City University（杭州城市大学）

针对全景场景图生成中的谓词标注偏差问题，提出ADTrans框架，通过学习无偏语义原型并自适应筛选、转换有偏标注，显著提升基准模型性能并达到新SOTA。

2024-01-22cs.CV

Cross-Domain Empirical Risk Minimization for Unbiased Long-tailed Classification

Nanyang Technological University（南洋理工大学）；Damo Academy, Alibaba Group（阿里巴巴达摩院）

We address the overlooked unbiasedness in existing long-tailed classification methods: we find that their overall improvement is mostly attributed to the biased preference of tail over head, as the test distribution is assumed to be balanced; however, when the test is as imbalanced as the long-tailed training data -- let the test respect Zipf's law of nature -- the tail bias is no longer beneficial overall because it hurts the head majorities. In this paper, we propose Cross-Domain Empirical Risk Minimization (xERM) for training an unbiased model to achieve strong performances on both test dis…

2021-12-29cs.CVcs.LG