Matches in SemOpenAlex for { <https://semopenalex.org/work/W4322760190> ?p ?o ?g. }
Showing items 1 to 74 of
74
with 100 items per page.
- W4322760190 abstract "Existing research efforts for multi-interest candidate matching in recommender systems mainly focus on improving model architecture or incorporating additional information, neglecting the importance of training schemes. This work revisits the training framework and uncovers two major problems hindering the expressiveness of learned multi-interest representations. First, the current training objective (i.e., uniformly sampled softmax) fails to effectively train discriminative representations in a multi-interest learning scenario due to the severe increase in easy negative samples. Second, a routing collapse problem is observed where each learned interest may collapse to express information only from a single item, resulting in information loss. To address these issues, we propose the REMI framework, consisting of an Interest-aware Hard Negative mining strategy (IHN) and a Routing Regularization (RR) method. IHN emphasizes interest-aware hard negatives by proposing an ideal sampling distribution and developing a Monte-Carlo strategy for efficient approximation. RR prevents routing collapse by introducing a novel regularization term on the item-to-interest routing matrices. These two components enhance the learned multi-interest representations from both the optimization objective and the composition information. REMI is a general framework that can be readily applied to various existing multi-interest candidate matching methods. Experiments on three real-world datasets show our method can significantly improve state-of-the-art methods with easy implementation and negligible computational overhead. The source code will be released." @default.
- W4322760190 created "2023-03-03" @default.
- W4322760190 creator A5031188042 @default.
- W4322760190 creator A5037356694 @default.
- W4322760190 creator A5046802258 @default.
- W4322760190 creator A5052723398 @default.
- W4322760190 creator A5076423724 @default.
- W4322760190 creator A5081953757 @default.
- W4322760190 creator A5089884176 @default.
- W4322760190 creator A5090119712 @default.
- W4322760190 date "2023-02-28" @default.
- W4322760190 modified "2023-09-30" @default.
- W4322760190 title "Rethinking Multi-Interest Learning for Candidate Matching in Recommender Systems" @default.
- W4322760190 doi "https://doi.org/10.48550/arxiv.2302.14532" @default.
- W4322760190 hasPublicationYear "2023" @default.
- W4322760190 type Work @default.
- W4322760190 citedByCount "0" @default.
- W4322760190 crossrefType "posted-content" @default.
- W4322760190 hasAuthorship W4322760190A5031188042 @default.
- W4322760190 hasAuthorship W4322760190A5037356694 @default.
- W4322760190 hasAuthorship W4322760190A5046802258 @default.
- W4322760190 hasAuthorship W4322760190A5052723398 @default.
- W4322760190 hasAuthorship W4322760190A5076423724 @default.
- W4322760190 hasAuthorship W4322760190A5081953757 @default.
- W4322760190 hasAuthorship W4322760190A5089884176 @default.
- W4322760190 hasAuthorship W4322760190A5090119712 @default.
- W4322760190 hasBestOaLocation W43227601901 @default.
- W4322760190 hasConcept C105795698 @default.
- W4322760190 hasConcept C111919701 @default.
- W4322760190 hasConcept C119857082 @default.
- W4322760190 hasConcept C138885662 @default.
- W4322760190 hasConcept C150140777 @default.
- W4322760190 hasConcept C154945302 @default.
- W4322760190 hasConcept C165064840 @default.
- W4322760190 hasConcept C27206212 @default.
- W4322760190 hasConcept C2776135515 @default.
- W4322760190 hasConcept C2778738651 @default.
- W4322760190 hasConcept C33923547 @default.
- W4322760190 hasConcept C41008148 @default.
- W4322760190 hasConcept C43126263 @default.
- W4322760190 hasConcept C557471498 @default.
- W4322760190 hasConcept C97931131 @default.
- W4322760190 hasConceptScore W4322760190C105795698 @default.
- W4322760190 hasConceptScore W4322760190C111919701 @default.
- W4322760190 hasConceptScore W4322760190C119857082 @default.
- W4322760190 hasConceptScore W4322760190C138885662 @default.
- W4322760190 hasConceptScore W4322760190C150140777 @default.
- W4322760190 hasConceptScore W4322760190C154945302 @default.
- W4322760190 hasConceptScore W4322760190C165064840 @default.
- W4322760190 hasConceptScore W4322760190C27206212 @default.
- W4322760190 hasConceptScore W4322760190C2776135515 @default.
- W4322760190 hasConceptScore W4322760190C2778738651 @default.
- W4322760190 hasConceptScore W4322760190C33923547 @default.
- W4322760190 hasConceptScore W4322760190C41008148 @default.
- W4322760190 hasConceptScore W4322760190C43126263 @default.
- W4322760190 hasConceptScore W4322760190C557471498 @default.
- W4322760190 hasConceptScore W4322760190C97931131 @default.
- W4322760190 hasLocation W43227601901 @default.
- W4322760190 hasLocation W43227601902 @default.
- W4322760190 hasOpenAccess W4322760190 @default.
- W4322760190 hasPrimaryLocation W43227601901 @default.
- W4322760190 hasRelatedWork W1980283324 @default.
- W4322760190 hasRelatedWork W2015036509 @default.
- W4322760190 hasRelatedWork W2045871438 @default.
- W4322760190 hasRelatedWork W2348159088 @default.
- W4322760190 hasRelatedWork W2350747448 @default.
- W4322760190 hasRelatedWork W2368095327 @default.
- W4322760190 hasRelatedWork W2809363009 @default.
- W4322760190 hasRelatedWork W2894140549 @default.
- W4322760190 hasRelatedWork W2968745142 @default.
- W4322760190 hasRelatedWork W4360764995 @default.
- W4322760190 isParatext "false" @default.
- W4322760190 isRetracted "false" @default.
- W4322760190 workType "article" @default.