SimpleX

Mao, Kelong; Zhu, Jieming; Wang, Jinpeng; Dai, Quanyu; Dong, Zhenhua; Xiao, Xi; He, Xiuqiang

doi:10.1145/3459637.3482297

Cited by 81 publications

(17 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Meanwhile, we compare the model with various types of SOTA models, range from SGL [33], SimpleX [21], SimSGL [36], NCL [19]. We also report the distribution of the learned personalized 𝜏 for Adap-𝜏 and Cu-𝜏 as marked by the orange and blue regions.…”

Section: 12mentioning

confidence: 99%

“…In this subsection, we are curious about how our model compares to those state-of-the-art models. Here, we choose two representative and powerful baseline: SGL [33], SimpleX [21], SimSGL [36], NCL [19], where SimpleX claims it surpass over 11 benchmark datasets and compared with 29 existing CF models in total.…”

Section: Comparison With Sota (Rq3)mentioning

confidence: 99%

“…Temperature has exhibited its capability in numerous fields such as CV and NLP in particular with contrastive learning [3,8]. Motivated by the success in other areas, recommendation combined with contrastive learning has received scant attention in recent research literature [21,28,33,34,38]. Although normalizaiton and 𝜏 are heuristically used by a small amount of work, it still lacks comprehensive exploration in recommendation.…”

Section: Temperature and Normalizationmentioning

confidence: 99%

“…Although normalizaiton and 𝜏 are heuristically used by a small amount of work, it still lacks comprehensive exploration in recommendation. [21] is mostly related work which analyzes the existing loss function and proposes a cosine contrastive loss to achieve hard-mining. However, its loss function filters easy items and needs detailed hyper parameter (margin and weight) searching, which restricted its flexibility in application and different datasets.…”

Section: Temperature and Normalizationmentioning

confidence: 99%

See 3 more Smart Citations

Adap-$τ$: Adaptively Modulating Embedding Magnitude for Recommendation

Chen,

Wu,

et al. 2023

Preprint

View full text Add to dashboard Cite

Recent years have witnessed the great successes of embeddingbased methods in recommender systems. Despite their decent performance, we argue one potential limitation of these methods -the embedding magnitude has not been explicitly modulated, which may aggravate popularity bias and training instability, hindering the model from making a good recommendation. It motivates us to leverage the embedding normalization in recommendation. By normalizing user/item embeddings to a specific value, we empirically observe impressive performance gains (9% on average) on four real-world datasets. Although encouraging, we also reveal a serious limitation when applying normalization in recommendation -the performance is highly sensitive to the choice of the temperature 𝜏 which controls the scale of the normalized embeddings.To fully foster the merits of the normalization while circumvent its limitation, this work studied on how to adaptively set the proper 𝜏. Towards this end, we first make a comprehensive analyses of 𝜏 to fully understand its role on recommendation. We then accordingly develop an adaptive fine-grained strategy Adap-𝜏 for the temperature with satisfying four desirable properties including adaptivity, personalized, efficiency and model-agnostic. Extensive experiments have been conducted to validate the effectiveness of the proposal.

show abstract

Section: 12mentioning

confidence: 99%

Section: Comparison With Sota (Rq3)mentioning

confidence: 99%

Section: Temperature and Normalizationmentioning

confidence: 99%

Section: Temperature and Normalizationmentioning

confidence: 99%

See 2 more Smart Citations

Adap-$τ$: Adaptively Modulating Embedding Magnitude for Recommendation

Chen,

Wu,

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…where  The Calculation of  k will change with different aggregation methods. Here we adopt three common aggregation strategies, including average-pooling, self-attention, and target-attention [29]:…”

mentioning

confidence: 99%

Recommend What to Cache: a Simple Self-supervised Graph-based Recommendation Framework for Edge Caching Network

Sun

Wang

Han³

2023

Preprint

View full text Add to dashboard Cite

The edge caching network based on deep learning can more accurately infer what to cache while driven by data from the user's historical content request, thus greatly relieving the burden of the backbone network. However, the inherent cold-start problem in deep learning may limit the performance of the history-based caching strategies. Due to the mobile and dynamic nature of the wireless network, the base station usually does not have enough data to accurately estimate the user's demand and cache the possible requested data. In this backdrop, we adopt self-supervised learning (SSL) into the caching strategy and propose a Simple Self-supervised Graph-based Recommendation framework for edge caching network (SimSGR). The SSL could alleviate the cold-start problem by generating extra training supervised signals from the raw data. To further optimize the performance of SSL, we propose two new network layers: Mixing and Conversion: the former replaces the data augmentation of the SSL paradigm to avoid destroying the semantic loss, while the latter greatly simplifies the loss function, which is helpful to light the model structure and facilitate deployment on edge caching network. Extensive experiments show that our model yields competitive or better performance than state-of-the-art methods sensitive to augmentation hyperparameters, especially when trained in the cold-start environment.

show abstract