Подборка статей 2025 (часть 2)

Фундаментальные модели
- PinFM: Foundation Model for User Activity Sequences at a Billion-scale Visual Discovery Platform
- Large Foundation Model for Ads Recommendation
- Foundation Model for Personalized Recommendation
- Meta’s Generative Ads Model (GEM): The Central Brain Accelerating Ads Recommendation AI Innovation
- Realizing Scaling Laws in Recommender Systems: A Foundation–Expert Paradigm for Hyperscale Model Deployment
- External Large Foundation Model: How to Efficiently Serve Trillions of Parameters for Online Ads Recommendation
- Meta Lattice: Model Space Redesign for Cost-Effective Industry-Scale Ads Recommendations

Мультимодальность и Semantic Ids
- VL-CLIP: Enhancing Multimodal Recommendations via Visual Grounding and LLM-Augmented CLIP Embeddings
- Enhancing Embedding Representation Stability in Recommendation Systems with Semantic ID
- Progressive Semantic Residual Quantization for Multimodal-Joint Interest Modeling in Music Recommendation
- QARM: Quantitative Alignment Multi-Modal Recommendation at Kuaishou
- BiListing: Modality Alignment for listings
- DAS: Dual-Aligned Semantic IDs Empowered Industrial Recommender System
- Sparse Meets Dense: Unified Generative Recommendations with Cascaded Sparse-Dense Representations
- MOON Embedding: Multimodal Representation Learning for E-commerce Search Advertising
- MOON2.0: Dynamic Modality-balanced Multimodal Representation Learning for E-commerce Product Understanding
- LEMUR: Large scale End-to-end MUltimodal Recommendation
- STORE: Semantic Tokenization, Orthogonal Rotation and Efficient Attention for Scaling Up Ranking Models
- Multi-Aspect Cross-modal Quantization for Generative Recommendation
- Personalized Multi Modal Alignment Encoding for CTR-Recommendation in WeChat
- Generative Recommendation with Semantic IDs: A Practitioner’s Handbook
- CoFiRec: Coarse-to-Fine Tokenization for Generative Recommendation
- Generating Long Semantic IDs in Parallel for Recommendation
- A Simple Contrastive Framework Of Item Tokenization For Generative Recommendation
- Learnable Item Tokenization for Generative Recommendation
- SIDE: Semantic ID Embedding for effective learning from sequences
- ActionPiece: Contextually Tokenizing Action Sequences for Generative Recommendation