AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Number Sequence Prediction Problems for Evaluating Computational Powers of Neural Networks

Hyoungwook Nam, Segwang Kim, Kyomin Jung

AAAI 2019 · Architectures and efficiency › Attention and transformers · approach: Transformer expressivity and circuits · 11 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 bidirectional instead of unidirectional LSTM in the attention model (ablation)Attention model · Reverse-order (n=16)test error rate: 42.0% -> 0.0%Improved
Use bidirectional instead of unidirectional LSTM in the attention model (ablation)Attention model · Fibonaccitest error rate: 69.3% -> 60.9%Improved
Add attention to LSTMDigit-level sequence model · Geometrictest error rate: 79.4% -> 62.8%Improved
Add attention to LSTMDigit-level sequence model · Arithmetictest error rate: 77.1% -> 77.0%No clear effect
Add stack memoryDigit-level RNN · Reverse-order (n=16)test error rate: 28.4% -> 0.0%Improved
Add external memory with NTMDigit-level RNN · Reverse-order (n=16)test error rate: 28.4% -> 0.0%Improved
Add stack or external memoryDigit-level RNN · Arithmetictest error rate: 77.1% -> 63.8%; 77.1% -> 65.7%Improved
Add stack or external memoryDigit-level RNN · Fibonaccitest error rate: 80.5% -> 69.4%; 80.5% -> 68.1%Improved
Increase CNN depth for number-level sequences (ablation)Number-level CNN · Number-level recurrence tasksprediction error rateImproved
Increase CNN depth for a mixture of primary sequence rules (ablation)Number-level CNN · Mixed primary sequencesprediction error rateNo clear effect
Use deeper models for compound sequence rules (ablation)Number-level CNN · Compound-rule sequencesprediction error rateImproved
Use ReLU rather than gate activation (ablation)Number-level CNN · Number-level sequencesaccuracyNo clear effect

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