RankMI: A Mutual Information Maximizing Ranking Loss

Kemertas, Mete; Pishdad, Leila; Derpanis, Konstantinos G.; Fazly, Afsaneh

doi:10.1109/cvpr42600.2020.01437

Cited by 28 publications

(14 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We follow DGI and use InfoNCE [27] as our learning objective to maximize the hierarchical mutual information. But we find that, compared with the binary cross-entropy loss, the pairwise ranking loss, which has also been proved to be effective in mutual information estimation [18], is more compatible with the recommendation task. We then define the objective function of the self-supervised task as follows:…”

Section: Enhancing Mhcn With Self-supervised Learningmentioning

confidence: 93%

Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social Recommendation

Yin

et al. 2021

Proceedings of the Web Conference 2021

272

119

View full text Add to dashboard Cite

Social relations are often used to improve recommendation quality when user-item interaction data is sparse in recommender systems. Most existing social recommendation models exploit pairwise relations to mine potential user preferences. However, real-life interactions among users are very complicated and user relations can be high-order. Hypergraph provides a natural way to model complex high-order relations, while its potentials for improving social recommendation are under-explored. In this paper, we fill this gap and propose a multi-channel hypergraph convolutional network to enhance social recommendation by leveraging high-order user relations. Technically, each channel in the network encodes a hypergraph that depicts a common high-order user relation pattern via hypergraph convolution. By aggregating the embeddings learned through multiple channels, we obtain comprehensive user representations to generate recommendation results. However, the aggregation operation might also obscure the inherent characteristics of different types of high-order connectivity information. To compensate for the aggregating loss, we innovatively integrate self-supervised learning into the training of the hypergraph convolutional network to regain the connectivity information with hierarchical mutual information maximization. The experimental results on multiple real-world datasets show that the proposed model outperforms the SOTA methods, and the ablation study verifies the effectiveness of the multi-channel setting and the selfsupervised task. The implementation of our model is available via https://github.com/Coder-Yu/RecQ.

show abstract

Section: Enhancing Mhcn With Self-supervised Learningmentioning

confidence: 93%

Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social Recommendation

Yin

et al. 2021

Proceedings of the Web Conference 2021

272

119

View full text Add to dashboard Cite

show abstract

“…For example, LLE was the best when the dimension was 32, and Isomap was the best when the dimension was 1024. When the embedding size was 128, P 2 ML-PA showed about 1% higher recall rate than RankMI [19] and PADS [58]. This proves that P 2 ML can support the scalability of embedding size.…”

Section: B Ablation Studymentioning

confidence: 70%

“…Although P 2 ML-Tri was designed based on basic Trip-semi, it showed 2.0% higher performance in CUB200 than DSML [21], which analyzed class variability in terms of SNR. Also, when the embedding size is 128, P 2 ML-Tri was comparable to RankMI [19] of SOTA performance. In the large-scale SOP dataset, it is worth paying attention to the generalization performance of P 2 ML, which is superior to RankMI.…”

Section: A Performance Evaluationmentioning

confidence: 84%

“…They learned to place points of the same class close to each other, while at the same time placing points of different classes far away. Also, the informationtheoretic approach placed points of the same class close to each other in terms of probability distribution by using Kullback-Leibler divergence (KLD) [18] or mutual information (MI) [19]. For the development trend of the latest DML techniques, refer to [20].…”

Section: Related Workmentioning

confidence: 99%

“…Some representative DML methods and several state-of-theart (SOTA) algorithms were adopted to compare with the proposed method [4,5,19,41]. Table II shows the quantitative results for CUB200-2011, Cars196, and SOP datasets.…”

Section: A Performance Evaluationmentioning

confidence: 99%

See 2 more Smart Citations

Deep Metric Learning With Manifold Class Variability Analysis

Kim

Song

2022

IEEE Trans. Multimedia

View full text Add to dashboard Cite

In deep metric learning (DML) techniques, understanding both the local and global characteristics of embedding space is essential. However, conventional DML techniques have two limitations as follows: First, Euclidean distance-based metrics never imply global information such as class variability because they only depend on the physical distance of samples. Second, they assume that the embedding space is simply a vector space which cannot represent complex data features. Therefore, we propose a novel loss function which can fully utilize characteristics of embedding space by using discriminant analysis and nonlinear mapping. With theoretical analysis, the superior performance of the proposed method is verified for the fine-grained retrieval datasets such as Cars196, CUB200-2011, Stanford online products, and In-shop clothes.

show abstract

Facial Representation Extraction by Mutual Information Maximization and Correlation Minimization

Wang

Sun

2021

Intelligent Information and Database Systems

View full text Add to dashboard Cite

RankMI: A Mutual Information Maximizing Ranking Loss

Cited by 28 publications

References 24 publications

Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social Recommendation

Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social Recommendation

Deep Metric Learning With Manifold Class Variability Analysis

Facial Representation Extraction by Mutual Information Maximization and Correlation Minimization

Contact Info

Product

Resources

About