Adversarial Multimodal Representation Learning for Click-Through Rate Prediction

Li, Xiang; Wang, Chao; Tan, Jiwei; Zeng, Xiaoyi; Ou, Dan; Zheng, Bo

doi:10.1145/3366423.3380163

Cited by 43 publications

(23 citation statements)

References 27 publications

(43 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the recommendation system, the platform calculates CTR through big data. The CTR of recommended products can be calculated through big data and compared with a typical click-through rate: if the CTR is bigger than 0.2%, the recommendation is more effective [22]. The platform can find personalized recommendations suitable for consumers through CTR value.…”

Section: 2description Of C-terminalmentioning

confidence: 99%

Bilateral Positioning Model for the E-commerce Platform

Du¹,

Guo²,

Xiao³

2021

Proceedings of the 2021 International Conference on Economic Development and Business Culture (ICEDBC 2021)

View full text Add to dashboard Cite

As the link between consumers and merchants, e-commerce platforms have problems that need to be improved. In particular, the accuracy and authenticity of the information they provide cannot be guaranteed, which will harm market fairness and consumers' right to know. To improve the presence of the e-commerce platform, the C2B model and big data are used, optimizing the consumer shopping experience and helping businesses to develop in the direction of professionalism. Based on this research concept, this article uses the systematic literature research method and collects professional literature through the search of keywords such as C2B, e-commerce platform, big data, and supply chain. The article summarizes the findings of the literature and finds imperfections, thereby innovating the Bilateral Positioning Model, which is, respectively optimizing the UI interface of the merchant and the consumer, and embedding the supervision module so that consumer evaluation can play an important role in the healthy competition of the platform. While describing the Bilateral Positioning Model, this article also analyzes its feasibility and practicability and theoretically proves the sustainability of the model.

show abstract

Section: 2description Of C-terminalmentioning

confidence: 99%

Bilateral Positioning Model for the E-commerce Platform

Du¹,

Guo²,

Xiao³

2021

Proceedings of the 2021 International Conference on Economic Development and Business Culture (ICEDBC 2021)

View full text Add to dashboard Cite

show abstract

“…Although early fusion methods have low computational complexity, the existence of redundancy reduces the effectiveness of information. The other type of fusion is late fusion (score-level fusion) [10,11]. The extracted representations or unimodal results are fused at the late stage.…”

Section: Introductionmentioning

confidence: 99%

Multimodal Emotion Recognition with Capsule Graph Convolutional Based Representation Fusion

Liu

Chen

Wang

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Due to the more robust characteristics compared to unimodal, audiovideo multimodal emotion recognition (MER) has attracted a lot of attention. The efficiency of representation fusion algorithm often determines the performance of MER. Although there are many fusion algorithms, information redundancy and information complementarity are usually ignored. In this paper, we propose a novel representation fusion method, Capsule Graph Convolutional Network (CapsGCN). Firstly, after unimodal representation learning, the extracted audio and video representations are distilled by capsule network and encapsulated into multimodal capsules respectively. Multimodal capsules can effectively reduce data redundancy by the dynamic routing algorithm. Secondly, the multimodal capsules with their inter-relations and intra-relations are treated as a graph structure. The graph structure is learned by Graph Convolutional Network (GCN) to get hidden representation which is a good supplement for information complementarity. Finally, the multimodal capsules and hidden relational representation learned by CapsGCN are fed to multihead self-attention to balance the contributions of source representation and relational representation. To verify the performance, visualization of representation, the results of commonly used fusion methods, and ablation studies of the proposed CapsGCN are provided. Our proposed fusion method achieves 80.83% accuracy and 80.23% F1 score on eNTERFACE05'.

show abstract

“…Most existing methods are adopted from the image-text embedding methods, which focus on the visual representation of videos. Some researchers [4,5,7,16,31,32,32,35,40,42,43] struggle to find a representative video frame, and then feed it into the imagetext model for video-text retrieval. However, other rich information in the videos effective for video-text retrieval is ignored.…”

Section: Introductionmentioning

confidence: 99%

Multi-Feature Graph Attention Network for Cross-Modal Video-Text Retrieval

Hao

Zhou

et al. 2021

Proceedings of the 2021 International Conference on Multimedia Retrieval

View full text Add to dashboard Cite

Cross-modal retrieval between videos and texts has attracted growing attention due to the rapid growth of user-generated videos on the web. To solve this problem, most approaches try to learn a joint embedding space to measure the cross-modal similarities, while paying little attention to the representation of each modality. Video is more complicated than the commonly used visual feature, since the audio and caption on the screen also contain rich information. Recently, the aggregations of multiple features in videos boost the benchmark of the video-text retrieval system. However, they usually handle each feature independently, which ignores the interchange of high-level semantic relations among these multiple features. Moreover, despite the inter-modal ranking constraint where semantically-similar texts and videos should stay closer, the modality-specific requirement, i.e. two similar videos/texts should have similar representations, is also significant. In this paper, we propose a novel Multi-Feature Graph ATtention Network (MFGATN) for cross-modal video-text retrieval. Specifically, we introduce a multi-feature graph attention module, which enriches the representation of each feature in videos with the interchange of highlevel semantic information among them. Moreover, we elaborately design a novel Dual Constraint Ranking Loss (DCRL), which simultaneously considers the inter-modal ranking constraint and the intra-modal structure constraint to preserve both the crossmodal semantic similarity and the modality-specific consistency * Corresponding author.

show abstract

Adversarial Multimodal Representation Learning for Click-Through Rate Prediction

Cited by 43 publications

References 27 publications

Bilateral Positioning Model for the E-commerce Platform

Bilateral Positioning Model for the E-commerce Platform

Multimodal Emotion Recognition with Capsule Graph Convolutional Based Representation Fusion

Multi-Feature Graph Attention Network for Cross-Modal Video-Text Retrieval

Contact Info

Product

Resources

About