AI Paper Atlas

Research map / Large language models / Agents, code and math

FACT: Learning Governing Abstractions Behind Integer Sequences

Peter Belcak, Ard Kastrati, Flavio Schenker, Roger Wattenhofer

NeurIPS 2022 · Large language models › Agents, code and math · approach: Inductive and abstract reasoning evaluation · 1 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 and test on synthetic instead of OEIS dataRNN · sequence classificationF1 score: 0.37 -> 0.53Improved
train and test on synthetic instead of OEIS dataTransformer · sequence continuationRMSLE: 0.578 -> 0.395Improved
use RNN instead of MLP for next sequence-part predictionRNN · next sequence-part prediction, OEISbinary accuracy: 0.733 -> 0.869Improved
use RNN instead of CNN for next sequence-part predictionRNN · next sequence-part prediction, OEISbinary accuracy: 0.551 -> 0.869Improved
use Transformer instead of RNN for continuationTransformer · sequence continuation, OEISRMSLE: 0.603 -> 0.578Improved
use Transformer instead of CNN for similarityTransformer · sequence similarity, OEIStop-5 RMSE: 0.428 -> 0.267Improved
use CNN instead of Transformer for unmaskingCNN · sequence element unmasking, OEIStop-5 RMSE: 2.811 -> 2.440Improved
use random forest instead of dummy classifierRandom Forest · sequence classification, syntheticF1 score: 0.50 -> 0.51Improved
use RNN instead of random forest for classificationRNN · sequence classification, syntheticF1 score: 0.51 -> 0.53Improved
use XGBoost instead of dummy classifierXGBoost · sequence classification, OEISF1 score: 0.50 -> 0.37Hurt
use linear regressor instead of dummy for continuationLinear Regressor · sequence continuation, OEISRMSLE: 0.923 -> 0.797Improved
use Transformer instead of RNN for similarity on syntheticTransformer · sequence similarity, synthetictop-5 RMSE: 0.438 -> 0.270Improved

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