🏗️ System DesignID:
dssm-two-tower-retrievalDSSM Two-Tower Retrieval
DSSM 双塔向量化召回
🎯Core Definition
DSSM Two-Tower Retrieval is an asymmetric neural architecture that decouples user features and item features into two independent sub-networks; the User Tower embeds user demographics, interaction history, and realtime context into vector , while the Item Tower embeds item attributes into , scoring relevance via dot product ; in production serving, all item embeddings are pre-computed offline and indexed in ANN stores (HNSW/Milvus), requiring only a single User Tower forward pass online followed by millisecond ANN Top-K nearest neighbor search.
💡Use Cases
Multi-channel recall across e-commerce items, sponsored ad candidates, video recommendations, and cross-modal search.
⚡Key Problems Solved
Classical Matrix Factorization fails on cold-start items and multi-field features, while cross-attention networks cannot be pre-computed offline; Two-Tower networks decouple inference completely, enabling sub-10ms recall over 100M items.
🎯5 High-Frequency Exam Points
1
Detail the compute efficiency of In-batch Negative Sampling in Two-Tower training and methods for Popularity Debias correction?
2
Why does the Two-Tower architecture strictly prohibit early low-level cross-features between user and item inputs?
3
How to update User Tower embeddings in real-time streaming pipelines to capture instant user interest shifts?
4
Explain the role and sensitivity of temperature hyperparameter in Two-Tower Softmax cross-entropy loss?
5
Analyze the optimal sampling mixture of Easy Negatives vs Hard Negatives to prevent embedding collapse and boost discrimination?
📖 In-depth Guide:📄 recommendation-system-design →
Updated 2026-08-14🎯
🚀 Start Card Practice ➔Test Your Knowledge: Practice Questions for "DSSM Two-Tower Retrieval"
Single choice pitfall questions with instant feedback and mistake tracking.