Negative Sampling Strategies for Enhancing Recommendation Systems
Negative sampling is a crucial but often overlooked component in recommendation systems, capable of revealing genuine negative aspects inherent in user preferences and improving model performance.
Enhancing Knowledge-based Recommendation with Common Sense and Large Language Models
Incorporating common sense and large language models enhances knowledge-based recommendation systems, addressing data sparsity and cold start issues.
Challenging the Two Tower Recommendation Model with One Backpropagation
Challenging the traditional two backpropagation strategy in recommendation models with a novel one backpropagation approach for improved performance.
RA-Rec: Efficient ID Representation Alignment Framework for LLM-based Recommendation
Incorporating ID embeddings into LLMs improves recommendation system performance.
InBox: Recommendation with Knowledge Graph using Interest Box Embedding
Introducing InBox, a novel embedding-based model that utilizes knowledge graph entities and relations to improve recommendation systems through box-based interest and concept modeling.
Mamba4Rec: Towards Efficient Sequential Recommendation with Selective State Space Models
State space models offer an efficient solution for sequential recommendation, addressing the effectiveness-efficiency dilemma.
Multi-Tower Multi-Interest Recommendation Framework with User Representation Repel
The author introduces a novel multi-tower multi-interest framework to address challenges in multi-interest learning, enhancing matching performance and industrial adoption.
Federated Recommendation Framework with Hybrid Retrieval and LLMs
The author proposes GPT-FedRec, a novel federated recommendation framework leveraging ChatGPT and a hybrid Retrieval Augmented Generation mechanism to address data sparsity and heterogeneity in FR, achieving superior performance against baseline methods.
NoteLLM: A Retrievable Large Language Model for Note Recommendation
The author introduces the NoteLLM framework, leveraging Large Language Models (LLMs) to enhance item-to-item (I2I) note recommendation by compressing notes and generating hashtags/categories simultaneously.
MENTOR: Multi-level Self-supervised Learning for Multimodal Recommendation
Our MENTOR method addresses label sparsity and modality alignment issues in multimodal recommendation by utilizing self-supervised learning and enhancing specific features of each modality. The approach involves multilevel tasks to align modalities effectively while maintaining historical interaction information.
© 2024 by Linnk AI