PQ-HSA: Reusing Product-Quantized Scores for Hybrid Sparse-Approximate Attention

arXiv:2609.33746 · cs.LG · Submitted 2026-09-27 · Read on arXiv

cs.LG

Submitted: 2026-09-27

Updated: 2026-09-27

Code: https://github.com/KunmingSHAO/pqhsa_release

Terminology

• Decode — 30× in this paper · explained in 2 episodes
• KV cache — 22× in this paper · explained in 3 episodes
• Accuracy — 15× in this paper · explained in 1 episode
• Quantization — 14× in this paper · explained in 12 episodes
• Core — 9× in this paper · explained in 2 episodes
• Sparse Attention — 6× in this paper · explained in 5 episodes
• Transformer — 6× in this paper · explained in 2 episodes
• Sparsity — 5× in this paper · explained in 4 episodes
• Benchmark — 4× in this paper · explained in 2 episodes
• International Conference on Machine Learning — 4× in this paper
• In International Conference on Learning Representations — 3× in this paper
• Aggregation — 2× in this paper · explained in 1 episode
• Generative AI — 2× in this paper · explained in 2 episodes
• Clustering — 1× in this paper · explained in 2 episodes
• Meeting of the Association for Computational Linguistics — 1× in this paper
• Transformer Models — 1× in this paper · explained in 2 episodes

Sources

• GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints
• LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding
• Scatterbrain: Unifying Sparse and Low-rank Attention Approximation
• RetroInfer: A Vector Storage Engine for Scalable Long-Context LLM Inference
• MagicPIG: LSH Sampling for Efficient LLM Generation
• Norm-Explicit Quantization: Improving Vector Quantization for Maximum Inner Product Search
• FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness
• HashAttention: Semantic Sparsity for Faster Inference
• The Faiss library
• The Llama 3 Herd of Models
• Accelerating Large-Scale Inference with Anisotropic Vector Quantization
• Squeezed Attention: Accelerating Long Context Length LLM Inference
• Multipole Attention for Efficient Long Context Reasoning
• RULER: What's the Real Context Size of Your Long-Context Language Models?
• Billion-scale similarity search with GPUs
• Efficient Memory Management for Large Language Model Serving with PagedAttention
• InfiniGen: Efficient Generative Inference of Large Language Models with Dynamic KV Cache Management
• CommVQ: Commutative Vector Quantization for KV Cache Compression
• SnapKV: LLM Knows What You are Looking for Before Generation
• Transformer-VQ: Linear-Time Transformers via Vector Quantization

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks