AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Fast Abductive Learning by Similarity-based Consistency Optimization

Yu-Xuan Huang, Wang-Zhou Dai, Le-Wen Cai, Stephen Muggleton, Yuan Jiang

NeurIPS 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 18 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
use similarity-based consistency (ABLSim) instead of NGS/DeepProbLog confidence-based searchneuro-symbolic learning from random init · HWF (CIFAR)test accuracy (%): 66.0±14.5 -> 88.4±0.7Improved
ABLSim vs NGS-opt on easy MNIST AdditionABLSim · MNIST Additiontest accuracy (%): 98.5±0.3 -> 98.8±0.1No clear effect
ABLSim vs NGS-opt on CIFAR AdditionABLSim · Addition (CIFAR)test accuracy (%): 88.7±0.8 -> 88.9±0.5No clear effect
increase beam width from 16 to 2048 (ablation)ABLSim · CIFAR-10 Decimal Equation first iterationsuccessful abduction rate: 62% -> 75%Improved
use similarity in consistency measure rather than confidence only (ABL-Conf) (ablation)ABL · CIFAR-10 Decimal Equationtest perception accuracyImproved
add supervised pre-training to ABL-Conf (ablation)ABL-Conf · CIFAR-10 Decimal Equationconvergence speedMixed
weaken the knowledge base by removing rules in ABLSim (ablation)ABLSim-10 · Theft Judicial Sentencing 10% labelsMicro-F1: 0.861 -> 0.851Hurt
ABLSim vs SS-ABL with partial KBSS-ABL · Theft Judicial Sentencing 10% labelsMicro-F1: 0.833 -> 0.851Improved

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