AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Scene Text Recognition with Permuted Autoregressive Sequence Models

Darwin Bautista, Rowel Atienza

ECCV 2022 · Large language models › Language, safety and interpretability · approach: Controlled text generation · 321 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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
train with 6 permutations instead of only left-to-right (K=1) (ablation)PARSeq-S, NAR decoding · Benchmark (94-char)word accuracy: 0.01 -> 92.22Improved
train with 6 permutations instead of only left-to-right (K=1) (ablation)PARSeq-S, cloze refinement · Benchmark (94-char)cloze word accuracy: 71.14 -> 94.81Improved
train with 2 permutations instead of 1 (ablation)PARSeq-S, AR decoding · Benchmark (94-char)AR word accuracy: 93.04 -> 93.48Improved
increase permutations from 6 to 24 (ablation)PARSeq-S, AR decoding · Benchmark (94-char)AR word accuracy: 93.34 -> 92.67Hurt
increase permutations from 6 to 24 (ablation)PARSeq-S, training cost · Benchmark (94-char)training hours: 8.48 -> 13.53Hurt
increase permutations from 6 to 12 (ablation)PARSeq-S, NAR decoding · Benchmark (94-char)NAR word accuracy: 92.22 -> 91.71Hurt
use AR decoding with refinement instead of NAR decodingPARSeq (synthetic train) · Benchmark mean (94-char)word accuracy: 70.5 -> 73.0Improved
train on real data instead of synthetic dataPARSeq_A · Benchmark mean (94-char)word accuracy: 73.0 -> 93.3Improved
use PARSeq_A instead of ABINet (external LM) with synthetic trainingABINet baseline · Benchmark mean (94-char)word accuracy: 66.4 -> 73.0Improved
use PARSeq_A instead of ABINet with real training on harder datasetsABINet baseline · ArT+COCO+Uber total (36-char)word accuracy: 74.6 -> 84.1Improved
add flipped (reversed) permutation pairs to sampling (ablation)PARSeq training · Training stabilitytraining loss smoothnessImproved
add position-query tokens decoupled from context (ablation)PARSeq decoder · PLM traininglearning from PLMImproved

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