Conditional Gaussian Distribution Learning for Open Set Recognition

Sun, Xin; Yang, Zhenning; Zhang, Chi; Peng, Guohao; Ling, Keck Voon

doi:10.48550/arxiv.2003.08823

Cited by 11 publications

(15 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Approaches as CGDL [39] and C2AE [30] couple the training of the generative model (usually a VAE) with the supervised training for KKC classification. These approaches allow for and end-to-end training of DNNs for OSR image classification tasks, and are currently considered the state-of-the-art for this task.…”

Section: Discussionmentioning

confidence: 99%

“…A recent trend in both Anomaly Detection and OSR for deep image classification has been to incorporate input reconstruction error in supervised DNN training as a way to identify OOD samples [44,30,39]. These approaches fall under the branch of generative OSR, according to the taxonomy by Geng et al [11].…”

Section: Open Set Recognitionmentioning

confidence: 99%

“…Class Conditional AutoEncoder (C2AE) [30], similarly to CROSR, uses the reconstruction error of the input (||x − x||) from an AE and EVT modeling to determine a threshold in order to discern between KKC and UUC samples. Following the same trend of thresholding a certain point in the density function of the reconstruction error from the inputs, Conditional Gaussian Distribution Learning (CGDL) [39] uses a Variational AutoEncoder (VAE) to model the bottleneck representation of the input images according to a vector of gaussian means µ and standard deviations σ in a lower-dimentional high semantic-level space. This modeling allows CGDL to unsupervisedly discriminate between KKCs and UUCs by thresholding the likelihood of the embedding z i generated from a novel sample x i pertaining to the multivariate gaussians N (z i , µ k , σ 2 k ), where k represents the predicted class for sample x i .…”

Section: Open Set Recognitionmentioning

confidence: 99%

“…OpenPCS works similarly to CGDL [39], but with three important differences: 1) we fit gaussian priors using not only the input images X, but also the intermediary activations (i.e. a (L 3 ) , a (L 4 ) , a (L 5 ) , etc); 2) we use a PCA instead of a VAE; and 3) the training is purely supervised and the low dimensional gaussians are fitted only during the validation phase, and not conjointly with the training, as CGDL.…”

Section: Openpcsmentioning

confidence: 99%

See 3 more Smart Citations

Fully Convolutional Open Set Segmentation

Oliveira¹,

Silva²,

Machado³

et al. 2020

Preprint

View full text Add to dashboard Cite

In semantic segmentation knowing about all existing classes is essential to yield effective results with the majority of existing approaches. However, these methods trained in a Closed Set of classes fail when new classes are found in the test phase. It means that they are not suitable for Open Set scenarios, which are very common in real-world computer vision and remote sensing applications. In this paper, we discuss the limitations of Closed Set segmentation and propose two fully convolutional approaches to effectively address Open Set semantic segmentation: OpenFCN and OpenPCS. OpenFCN is based on the wellknown OpenMax algorithm, configuring a new application of this approach in segmentation settings. OpenPCS is a fully novel approach based on feature-space from DNN activations that serve as features for computing PCA and multi-variate gaussian likelihood in a lower dimensional space. Experiments were conducted on the well-known Vaihingen and Potsdam segmentation datasets. OpenFCN showed little-to-no improvement when compared to the simpler and much more time efficient SoftMax thresholding, while being between some orders of magnitude slower. OpenPCS achieved promising results in almost all experiments by overcoming both OpenFCN and SoftMax thresholding. OpenPCS is also a reasonable compromise between the runtime performances of the extremely fast SoftMax thresholding and the extremely slow OpenFCN, being close able to run close to real-time. Experiments also indicate that OpenPCS is effective, robust and suitable for Open Set segmentation, being able to improve the recognition of unknown class pixels without reducing the accuracy on the known class pixels.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Open Set Recognitionmentioning

confidence: 99%

Section: Open Set Recognitionmentioning

confidence: 99%

Section: Openpcsmentioning

confidence: 99%

See 2 more Smart Citations

Fully Convolutional Open Set Segmentation

Oliveira¹,

Silva²,

Machado³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Table 1 shows the experimental results of distinguishing unknown classes with the DNN method on the CIFAR10 dataset. The Area Under the Receiver Operating Characteristic (AUROC) curve [41] is adopted as the evaluation index. The greater AUROC value indicates a stronger ability of the model to distinguish unknown samples.…”

Section: Decision Variablesmentioning

confidence: 99%

Multi-granularity Inter-class Correlation Based Contrastive Learning for Open Set Recognition

Zhu¹,

Zhang²,

Wang³

et al. 2022

International Journal of Software and Informatics

View full text Add to dashboard Cite

In recent years, deep neural networks have continuously achieved breakthroughs in the classification task. However, they will mistakenly give a wrong known class prediction when faced with unknown samples in the testing phase. The open set recognition is a possible way to solve the problem, which requires the model to not only classify the known classes but also distinguish the unknown samples accurately. Most of the existing methods are designed heuristically on the basis of certain assumptions. Despite keeping the performance increasing, they have not analyzed the key factors that affect the task. In this paper, we analyze the commonalities of existing methods by designing a new decision variable experiment and find that the ability of the model to learn representations of known classes is an important factor. Then an open set recognition method is proposed based on the representation learning ability enhancement of the model. Firstly, due to the powerful representation learning capabilities demonstrated by the contrastive learning and the label information contained in the open set recognition task, supervised contrastive learning is introduced to improve the modeling ability of the model for known classes. Secondly, considering that the inter-class correlation is the representation learning at the class level, and the hierarchical structure relationship among the classes is often presented, a loss function of the multi-granularity inter-class correlation is designed. In the way of building the hierarchical structure in the label semantic space and measuring the multi-granularity inter-class correlation, the loss function of multi-granularity inter-class correlation constrains the model to learn the correlation among different known classes to further improve the representation learning ability of the model. Finally, experimental results on multiple standard datasets verify the effectiveness of the proposed method in open set recognition.

show abstract

Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety

Houben

Abrecht

Akila

et al. 2022

Deep Neural Networks and Data for Automated Driving

View full text Add to dashboard Cite

Deployment of modern data-driven machine learning methods, most often realized by deep neural networks (DNNs), in safety-critical applications such as health care, industrial plant control, or autonomous driving is highly challenging due to numerous model-inherent shortcomings. These shortcomings are diverse and range from a lack of generalization over insufficient interpretability and implausible predictions to directed attacks by means of malicious inputs. Cyber-physical systems employing DNNs are therefore likely to suffer from so-called safety concerns, properties that preclude their deployment as no argument or experimental setup can help to assess the remaining risk. In recent years, an abundance of state-of-the-art techniques aiming to address these safety concerns has emerged. This chapter provides a structured and broad overview of them. We first identify categories of insufficiencies to then describe research activities aiming at their detection, quantification, or mitigation. Our work addresses machine learning experts and safety engineers alike: The former ones might profit from the broad range of machine learning topics covered and discussions on limitations of recent methods. The latter ones might gain insights into the specifics of modern machine learning methods. We hope that this contribution fuels discussions on desiderata for machine learning systems and strategies on how to help to advance existing approaches accordingly.

show abstract

Conditional Gaussian Distribution Learning for Open Set Recognition

Cited by 11 publications

References 31 publications

Fully Convolutional Open Set Segmentation

Fully Convolutional Open Set Segmentation

Multi-granularity Inter-class Correlation Based Contrastive Learning for Open Set Recognition

Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety

Contact Info

Product

Resources

About