Deep User Match Network for Click-Through Rate Prediction

Huang, Zai; Tao, Mingyuan; Zhang, Bufeng

doi:10.1145/3404835.3463078

Cited by 16 publications

(5 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared with others gating mechanism module (e.g. [37, 40, 42]), our methods is different with previous gating mechanism modules. For example, the [42] proposes gated multi‐scale aggregation module to fuse multi‐scale feature, which introduces gated sum and RNN to produce some weights of different scales.…”

Section: Methodsmentioning

confidence: 82%

See 1 more Smart Citation

Lite‐weight semantic segmentation with AG self‐attention

Liu

Gao

et al. 2023

IET Computer Vision

View full text Add to dashboard Cite

Due to the large computational and GPUs memory cost of semantic segmentation, some works focus on designing a lite weight model to achieve a good trade‐off between computational cost and accuracy. A common method is to combined CNN and vision transformer. However, these methods ignore the contextual information of multi receptive fields. And existing methods often fail to inject detailed information losses in the downsampling of multi‐scale feature. To fix these issues, we propose AG Self‐Attention, which is Enhanced Atrous Self‐Attention (EASA), and Gate Attention. AG Self‐Attention adds the contextual information of multi receptive fields into the global semantic feature. Specifically, the Enhanced Atrous Self‐Attention uses weight shared atrous convolution with different atrous rates to get the contextual information under the specific different receptive fields. Gate Attention introduces gating mechanism to inject detailed information into the global semantic feature and filter detailed information by producing “fusion” gate and “update” gate. In order to prove our insight. We conduct numerous experiments in common semantic segmentation datasets, consisting of ADE20 K, COCO‐stuff, PASCAL Context, Cityscapes, to show that our method achieves state‐of‐the‐art performance and achieve a good trade‐off between computational cost and accuracy.

show abstract

Section: Methodsmentioning

confidence: 82%

“…Highway networks used gates to regulate the passing of information to ease the process of training. GateNet [37] proposed feature embedding and hidden gates to capture the high‐order interaction information. PAD‐Net [38] fused multi‐model features in different auxiliary tasks.…”

Section: Related Workmentioning

confidence: 99%

Lite‐weight semantic segmentation with AG self‐attention

Liu

Gao

et al. 2023

IET Computer Vision

View full text Add to dashboard Cite

show abstract

“…In DIN [2], gated weighting is used to dynamically adjust the weight of different users' historical behavior. In GateNet [10] , it is proposed that the embedding gate should employ a MLP gate at the embedding layer and employ a MLP gate after MLP. It is proposed that information can be shared between deep feature representations and shallow feature representations for simultaneous use in CTR prediction.…”

Section: Gating Mechanismmentioning

confidence: 99%

Reweighted Attention with Auto Feature Select Gate for CTR Prediction

2024

AJCIS

View full text Add to dashboard Cite

CTR (Click-Through-Rate) prediction is an important part of today's recommendation scenarios. Its purpose is to predict whether the user will click on the relevant item.The more mainstream research includes feature interaction and user history behavior interest modeling. However, in some current CTR feature interaction methods, only bit-wise or vector-wise feature interaction is used, or the two-tower approach is used to simply add the output of the two-tower structure at the prediction layer. These methods cannot more comprehensively represent the interaction between features, thus losing a lot of potential feature interaction information.In this paper we propose two novel method:1)Reweighted Attention Network(RAN),which employs vector re-weight after capturing the explicit feature interaction .This module can help model capture the high-order feature potential information more effectively.2)Auto Feature Select Gate(AFSG),which mining potential interactive information of shallow features and higher-order features On the basis of avoiding information loss.Experiments on three public datasets show that our method performs better than the current mainstream CTR prediction models.

show abstract

“…e interest denoising layer is designed to purify the presentation of implicit negative feedback with noise. In order to explore the effectiveness of using like representation to denoise unclick representation by denoising module, we design three different denoising modules for comparison, namely, MBIN-IDN, which removes the interest denoising layer; MBIN DUMN [19], which utilizes a vector project layer to get the vector representation after denoising; and MBIN XDM [20], which gets the representation purified by the confidence layer and triplet loss.…”

Section: Effectiveness Of the Interest Denoising Layermentioning

confidence: 99%

Multifeedback Behavior-Based Interest Modeling Network for Adaptive Click-Through Rate Prediction

Shang

Huang

2022

Mobile Information Systems

View full text Add to dashboard Cite

With the rapid development of the Internet, the recommendation system is becoming more and more important in people’s life. Click-through rate prediction is a crucial task in the recommendation system, which directly determines the effect of the recommendation system. Recently, researchers have found that considering the user behavior sequence can greatly improve the accuracy of the click-through rate prediction model. However, the existing prediction models usually use the user click behavior sequence as the input of the model, which will make it difficult for the model to obtain a comprehensive user interest representation. In this paper, a unified multitype user behavior sequence modeling framework named as MBIN, a.k.a. multifeedback behavior-based Interest modeling network, is proposed to cope with uncertainties in the noisy data. The proposed adaptive model uses deep learning technology, obtains user interest representation through multihead attention, denoises user interest representation using the vector projection method, and fuses the user interests using adaptive dropout technology. First, an interest denoising layer is proposed in the MBIN, which can effectively mitigate the noise problem in user behavior sequences to obtain more accurate user interests. Second, an interest fusion layer is introduced so as to effectively model and fuse various types of interest representations of users to achieve personalized interest fusion. Then, we used auxiliary losses based on behavior sequences to enhance the effect of behavior sequence modeling and improve the effectiveness of user interest characterization. Finally, we conduct extensive experiments based on real-world and large-scale dataset to validate the effectiveness of our approach in CTR prediction tasks.

show abstract

Deep User Match Network for Click-Through Rate Prediction

Cited by 16 publications

References 23 publications

Lite‐weight semantic segmentation with AG self‐attention

Lite‐weight semantic segmentation with AG self‐attention

Reweighted Attention with Auto Feature Select Gate for CTR Prediction

Multifeedback Behavior-Based Interest Modeling Network for Adaptive Click-Through Rate Prediction

Contact Info

Product

Resources

About