AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Polysemous Codes

Matthijs Douze, Hervé Jégou, Florent Perronnin

ECCV 2016 · Learning theory and algorithms › Prediction and decision losses · approach: Approximate nearest neighbor indexing · 48 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (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 changedWhereMeasuredEffect
optimize index assignment (distance loss) before binary comparison (ablation)PQ/binary · SIFT1Mrecall@1: 0.036 -> 0.107Improved
optimize index assignment (distance loss) before binary comparison (ablation)PQ/binary · SIFT1Mrecall@100: 0.129 -> 0.503Improved
optimize index assignment (distance loss) for Hamming filtering (ablation)PQ/dual · SIFT1Mrecall@1: 0.312 -> 0.441Improved
optimize index assignment (distance loss) for Hamming filtering (ablation)PQ/dual · FYCNN1Mrecall@100: 0.522 -> 0.804Improved
use ranking loss instead of distance estimator loss (ablation)PolyD/dual · SIFT1Mrecall@100: 0.987 -> 0.960Hurt
use ranking loss instead of distance estimator loss (ablation)PolyD/binary · FYCNN1Mrecall@100: 0.281 -> 0.222Hurt
add Hamming filtering (dual) on top of polysemous codes versus binary onlyPolyD · SIFT1Mrecall@1: 0.107 -> 0.441Improved
use polysemous dual instead of exhaustive PQ ADCPQ (ADC) · SIFT1Mquery time (ms): 9.01 -> 2.53Improved
use polysemous dual instead of exhaustive PQ ADCPQ (ADC) · SIFT1Mrecall@1: 0.442 -> 0.441No clear effect
replace disidx filtering with polysemous dualPQ/disidx · SIFT1Mrecall@1: 0.071 -> 0.441Improved
add PolyD+ADC filtering to IMI*IMI* 4096^2 16/30k · BIGANN (16 bytes)recall@1: 0.238 -> 0.217Hurt
add PolyD+ADC filtering to IMI*IMI* 4096^2 16/30k · BIGANN (16 bytes, batch)query time (ms): 0.75 -> 0.38Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy