Mitigating Popularity Bias in Recommendation with Unbalanced Interactions: A Gradient Perspective

Ren, Weijieying; Wang, Lei; Liu, Kunpeng; Guo, Ruocheng; Peng, Lim Ee; Fu, Yanjie

doi:10.1109/icdm54844.2022.00054

Cited by 10 publications

(9 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 3 presents the correlation between item popularity and embedding magnitude, demonstrating a significant correlation between these two factors. Our finding aligns with previous work by Ren et al [14], who also observe that popular items tend to have larger embedding magnitudes due to positive gradients acquired during model updates.…”

Section: Test Time Embedding Normalizationsupporting

confidence: 92%

“…We use LightGCN (LGN) [8] trained with BPR and SSM loss as a baseline and backbone model for which TTEN is applied. We compare our approach with four methods, IPS [12], MACR [18], GRAD [14], and BIGNN [11], designed to mitigate the popularity bias in recommender systems. We reproduce all baselines except for GRAD and BiGNN.…”

Section: Methodsmentioning

confidence: 99%

“…Evaulation Protocols. Many approaches that try to mitigate popularity bias [14,18,21] commonly utilize an unbiased test set to assess the performance of the proposed method. This is because the biased test set, which often follows a long-tail distribution, can yield high performance even when the model produces a biased recommendation.…”

Section: Methodsmentioning

confidence: 99%

“…When trained with a proper loss, such as sampled softmax loss [20], we find that the cosine similarity distinguishes preferable and non-preferable items regardless of their popularity. Interestingly, our observation indicates that the well-known models are inherently capable of disentangling popularity and preference without the need for explicit bias mitigation algorithms [11,12,14,18]. The commonly used inner product score function, however, strengthens the popularity bias since the function multiplies cosine similarity and magnitude of embeddings together.…”

Section: Introductionmentioning

confidence: 93%

“…Zheng et al [21] separate user and item embeddings into interest and conformity using a negative sampling strategy based on causality. Ren et al [14] propose a gradient-based method to address popularity bias. They argue that popularity bias arises due to the dominance of popular items in the positive items, resulting in a significantly larger gradient of positive items than negative items.…”

Section: Related Workmentioning

confidence: 99%

See 4 more Smart Citations

Hydrogeochemical anomalies associated with the 2017 MW 5.5 Pohang earthquake in South Korea

Kim¹,

Lee²

2020

Preprint

View full text Add to dashboard Cite

The 2017 MW 5.5 Pohang earthquakes, which were known as triggered by enhanced geothermal system (EGS) stimulations, had significant effects on the groundwater system. This study aimed to identify the hydrogeochemical anomalies and to understand the response mechanisms of groundwater system to the earthquake. For this, the environmental isotopes (222Rn, Sr, 2H, and 18O), major ions, and time-series data (groundwater level, temperature, and electrical conductivity) were analyzed. Principal component analysis (PCA) was also employed. The results from time-series data showed the anomalies in the groundwater wells located near the epicenter. The hydrochemical parameters including stable isotopes data of 2H and 18O showed the different change patterns among groundwater wells before/after the earthquakes, which were related to the distance from epicenter, faults, and seawater. The environmental isotopes, radon and strontium, suggested the possible mechanisms underlying the effects of earthquakes by spatial distributions, such as seawater intrusion, water-rock interactions, shallow and deep aquifer mixing, deep fluid upwelling, and bedrock fracture opening. With this, the main cluster of PCA results was also distributed along these isotope concentration gradients. Our findings proved the usefulness of environmental isotopes and hydrogeochemical parameters to understand the earthquake-related changes in groundwater system. These studied parameters and the adopted methods would be positively applied for other earthquake zones.

show abstract

Section: Test Time Embedding Normalizationsupporting

confidence: 92%