Research map / Architectures and efficiency / RNNs and normalization
Efficient Large-Scale Approximate Nearest Neighbor Search on OpenCL FPGA
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use the proposed mantissa-exponent codebook compression | IVFPQ on Deep1B · Deep1B | R@10: 0.353 -> 0.366 | Helped |
| Use a parallel fused k-selection kernel instead of the GPU baselines | k-selection · random distance arrays | runtime: 8.59x faster than Faiss and 15.65x faster than fgknn | Helped |
| Use the complete FPGA ANN method | SIFT1M · SIFT1M | Recall@1: 0.8 -> 0.88 | Helped |
| Increase mantissa quantization level (ablation) | proposed codebook compression · quantization evaluation | mAP | Helped |
| Increase the number of PQ centroids rather than the number of sub-quantizers (ablation) | PQ · PQ codebook evaluation | precision | Helped |
| Increase the PQ codebook size (ablation) | proposed compressed-codebook method · PQ codebook evaluation | mAP | Helped |
Closest papers by meaning
- CSPG: Crossing Sparse Proximity Graphs for Approximate Nearest Neighbor Search (NeurIPS 2024 · similarity 0.85)
- LoRANN: Low-Rank Matrix Factorization for Approximate Nearest Neighbor Search (NeurIPS 2024 · similarity 0.84)
- FALQON: Accelerating LoRA Fine-tuning with Low-Bit Floating-Point Arithmetic (NeurIPS 2025 · similarity 0.83)
- LotusFilter: Fast Diverse Nearest Neighbor Search via a Learned Cutoff Table (CVPR 2025 · similarity 0.83)
- ScaleBITS: Scalable Bitwidth Search for Hardware-Aligned Mixed-Precision LLMs (NeurIPS 2026 · similarity 0.82)
- JAQ: Joint Efficient Architecture Design and Low-Bit Quantization with Hardware-Software Co-Exploration (AAAI 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.