AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Training Linear Finite-State Machines

Arash Ardakani, Amir Ardakani, Warren Gross

NeurIPS 2020 · Architectures and efficiency › RNNs and normalization · approach: LSTM design and memory · 4 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 FSM-based network instead of SC-based TCAD'18MNIST MLP · MNIST testerror rate (%): 3 -> 1.28Improved
use FSM-based network instead of SC-based TCOMP'18MNIST MLP · MNIST testerror rate (%): 2.05 -> 1.28Improved
use FSM-based network instead of SC-based TVLSI'17MNIST MLP · MNIST testerror rate (%): 2.01 -> 1.28Improved
shrink FSM-based network from 784-250-250-10 to 784-70-70-10FSM-based network · MNIST testerror rate (%): 1.28 -> 1.66Hurt
use stream length 128 instead of 256 versus SC baselinesFSM-based network · MNIST teststream length: 256 -> 128Improved
use two states instead of more states (ablation)FSM-based network · MNIST testerror rateImproved
increase stream length beyond 64 (ablation)FSM-based network · MNIST testerror rateNo clear effect
increase stream length and number of states (ablation)three-layer FSM Gabor network · 2D Gabor filterMSEImproved
replace LSTM with 4-state FSM-based modelchar-level language model · Penn TreebankBPC: 1.45 -> 1.52Hurt
replace LSTM with 4-state FSM-based modelchar-level language model · War & PeaceBPC: 1.83 -> 1.89Hurt
replace LSTM with 4-state FSM-based modelchar-level language model · Linux KernelBPC: 1.85 -> 1.93Hurt
replace LSTM with FSM-based modelchar-level language model · Penn Treebanknumber of operations: 8.1M -> 1.1MImproved

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