8 papers
CAMIE: Co-Engagement-Aware Multimodal Item Embeddings for Snap Dynamic Product Ads Retrieval
Xiaodong Liu, Siman Wang, Congfei Zhang +9
Item-to-item (I2I) retrieval is a core primitive in large-scale recommendation and advertising systems. In production Snap Dynamic Product Ads (DPA), I2I retrieval faces two challe…
SetMIR: Multi-Interest Retrieval as Set Prediction
Xiaodong Liu, Congfei Zhang, Hsiang-wei Chao +13
Embedding-based retrieval is at the core of industrial recommender systems, but a single user embedding is often too limited to capture a user's diverse interests. Multi-interest r…
SMART: LLM-Augmented Hybrid Retrieval for Dynamic Product Ads
Congfei Zhang, Jingxiao Ma, Xiaodong Liu +14
Dynamic Product Ads (DPA) require retrieving relevant items from multi-million product catalogs, balancing two competing objectives: retargeting (re-surfacing known interests) and…
EGR: Embedding-Native Generative Retrieval with a Shared LLM
Xiaodong Liu, Congfei Zhang, Hsiang-wei Chao +13
Generative retrieval is increasingly popular in large-scale recommendation and advertising systems, yet current methods introduce practical complications. Semantic-ID methods rely…
CRISP: Pre-LLM Yet Text-Driven Visual Token Pruning for Efficient LVLM Inference
Xu Li, Yi Zheng, Mengyang Zhao +7
Large Vision-Language Models (LVLMs) typically require processing hundreds to thousands of visual tokens, leading to substantial inference overhead. Existing visual token pruning m…
ResPrune: Text-Conditioned Subspace Reconstruction for Visual Token Pruning in Large Vision-Language Models
Xu Li, Yi Zheng, Yuxuan Liang +5
Large Vision-Language Models (LVLMs) rely on dense visual tokens to capture fine-grained visual information, but processing all these tokens incurs substantial computational and me…
Pyramid Token Pruning for High-Resolution Large Vision-Language Models via Region, Token, and Instruction-Guided Importance
Yuxuan Liang, Xu Li, Xiaolei Chen +4
Large Vision-Language Models (LVLMs) have recently demonstrated strong multimodal understanding, yet their fine-grained visual perception is often constrained by low input resoluti…
HERO: Rethinking Visual Token Early Dropping in High-Resolution Large Vision-Language Models
Xu Li, Yuxuan Liang, Xiaolei Chen +4
By cropping high-resolution images into local tiles and encoding them independently, High-Resolution Large Vision-Language Models (HR-LVLMs) have demonstrated remarkable fine-grain…