Neighborhood Contrastive Learning for Novel Class Discovery

Zhong, Zhun; Fini, Enrico; Roy, Subhankar; Luo, Zhiming; Ricci, Elisa; Sebe, Nicu

doi:10.1109/cvpr46437.2021.01072

Cited by 94 publications

(75 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…OpenMix [50] mixes the labelled and unlabelled data to avoid the model from over-fitting for NCD. NCL [49] extracts and aggregates the pairwise pseudo-labels for the unlabelled data with contrastive learning and generates hard negatives by mixing the labelled and unlabelled data in the feature space for NCD. Jia et al [21] propose an end-to-end NCD method for singleand multi-modal data with contrastive learning and winnertakes-all hashing.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Generalized Category Discovery

Vaze¹,

Han²,

Vedaldi³

et al. 2022

Preprint

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

“…Meanwhile, self-supervised contrastive learning has been widely used as pre-training to achieve robust representations in NCD [21,49]. Furthermore, when combined with vision transformers, it generates models which are good nearest neighbour classifiers [5].…”

Section: Our Approachmentioning

confidence: 99%

Generalized Category Discovery

Vaze¹,

Han²,

Vedaldi³

et al. 2022

Preprint

View full text Add to dashboard Cite

“…In this section, we explore whether it is necessary to use the function form based on InfoNCE loss to approximate the probability to the one-hot form. Therefore, we consider binary cross entropy loss (BCE) which is used in [79]. For convenience, we redefine the probabilities p i and pi as p 0 i and p 1 i .…”

Section: The Importance Of Infonce Lossmentioning

confidence: 99%

Probabilistic Contrastive Learning for Domain Adaptation

Li¹,

Zhang²,

Wang³

et al. 2021

Preprint

View full text Add to dashboard Cite

Recent feature contrastive learning (FCL) has shown promising performance in self-supervised representation learning. For domain adaptation, however, FCL cannot show overwhelming gains since the class weights are not involved during optimization, which does not guarantee the produced features to be clustered around the class weights learned from source data. To tackle this issue, we propose a novel probability contrastive learning (PCL) in this paper, which not only produces compact features but also enforces them to be distributed around the class weights. Specifically, we propose to use the output probabilities after softmax to perform contrastive learning instead of the extracted features and remove the 2 normalization in the traditional FCL. In this way, the probability will approximate the onehot form, thereby narrowing the distance between the features and the class weights. Our proposed PCL is simple and effective. We conduct extensive experiments on two domain adaptation tasks, i.e., unsupervised domain adaptation and semi-supervised domain adaptation. The results on multiple datasets demonstrate that our PCL can consistently get considerable gains and achieves the state-of-theart performance. In addition, our method also obtains considerable gains on semi-supervised tasks when labeled data is scarce.

show abstract

“…Given that, we propose Domain-Specific Contrastive Learning (DSCL) to fully mine intra-domain knowledge by only selecting the positive and negative samples from the query's domain for contrastive learning. Furthermore, inspired by the recent negative mining methods (Kalantidis et al 2020;Zhong et al 2021) that use interpolation in the latent space to synthesize more negative samples, we propose Second-Order Nearest Interpolation (SONI) to obtain additional hard negative samples for the query from the target domain. To ensure the synthetic negative samples are reliable and informative, SONI selects a set of nearest negative centroids and then uses each centroid as an anchor to interpolate with another nearest negative centroid that is nearest to it but has a different pseudo label.…”

Section: Introductionmentioning

confidence: 99%

Multi-Centroid Representation Network for Domain Adaptive Person Re-ID

Wu¹,

Huang²,

Yao³

et al. 2022

AAAI

View full text Add to dashboard Cite

Recently, many approaches tackle the Unsupervised Domain Adaptive person re-identification (UDA re-ID) problem through pseudo-label-based contrastive learning. During training, a uni-centroid representation is obtained by simply averaging all the instance features from a cluster with the same pseudo label. However, a cluster may contain images with different identities (label noises) due to the imperfect clustering results, which makes the uni-centroid representation inappropriate. In this paper, we present a novel Multi-Centroid Memory (MCM) to adaptively capture different identity information within the cluster. MCM can effectively alleviate the issue of label noises by selecting proper positive/negative centroids for the query image. Moreover, we further propose two strategies to improve the contrastive learning process. First, we present a Domain-Specific Contrastive Learning (DSCL) mechanism to fully explore intra-domain information by comparing samples only from the same domain. Second, we propose Second-Order Nearest Interpolation (SONI) to obtain abundant and informative negative samples. We integrate MCM, DSCL, and SONI into a unified framework named Multi-Centroid Representation Network (MCRN). Extensive experiments demonstrate the superiority of MCRN over state-of-the-art approaches on multiple UDA re-ID tasks and fully unsupervised re-ID tasks.

show abstract

Neighborhood Contrastive Learning for Novel Class Discovery

Cited by 94 publications

References 10 publications

Generalized Category Discovery

Generalized Category Discovery

Probabilistic Contrastive Learning for Domain Adaptation

Multi-Centroid Representation Network for Domain Adaptive Person Re-ID

Contact Info

Product

Resources

About