List your company
ResearchFormerly WizWand

Image Classification on MNIST standard (test)

99.7AccuracyDF blockwise
56.987268.076179.16590.2539Nov 7, 2020Oct 8, 2021Sep 9, 2022Aug 10, 2023Jul 11, 2024Jun 11, 2025May 13, 2026
Updated 5mo ago

Evaluation results

MethodLinks
DF blockwise# Layers=102025.0199.7——
ASGE# Layers=112025.0199.59——
MF cumulative*# Layers=42025.0199.58——
MF final*# Layers=42025.0199.58——
DF layerwise# Layers=102025.0199.53——
BP*# Layers=42025.0199.52——
DeeperForward# Layers=42025.0199.5——
FAUST# Layers=52025.0199.5——
Channel-wise FF# Layers=42025.0199.42——
LogicTreeNet-LModel=CNN, Gate count=1.27 M2026.0299.35——
PearsonNoise Type=Sparse, Noise Level=Low, Bias Correction=false2020.1199.24——
JeffreyNoise Type=Sparse, Noise Level=Low, Bias Correction=false2020.1199.24——
LogicTreeNet-MModel=CNN, Gate count=566 K2026.0299.23——
BitLogicModel=FFN, Gate count=384 K2026.0299.15——
SchrödingerHidden dimension=64, Depth=3, Batch size=16, Dropout=0.1, Global mean pooling=true, Epochs=200, Optimizer=Adam, Learning rate=3 × 10−4, Number of random seeds=52026.0599.13——
CNNBatch size=16, Dropout=0.1, Epochs=200, Optimizer=Adam, Learning rate=3 × 10−4, Number of random seeds=5, Overall width/depth scale=Matched to GNNs2026.0599.07——
CaFo# Layers=32025.0199.04——
FA*# Layers=42025.0198.99——
LILogicNet-LModel=FFN, Gate count=32 K2026.0298.95——
MPNNHidden dimension=64, Depth=3, Batch size=16, Dropout=0.1, Global mean pooling=true, Epochs=200, Optimizer=Adam, Learning rate=3 × 10−4, Number of random seeds=52026.0598.95——
DTP# Layers=62025.0198.93——
OursP(s=1)=0.5, Base loss=MAE, Backbone=CNN2026.0598.86——
SNN (rate coding)Network=784-800-10, Coding=SNN (rate coding), BP scheme=surrogate-gradient, Spike type=multi, Leakage=leaky, τV=τ, τI=τ2022.1198.84——
Trans-PRODENP(s=1)=0.5, Base loss=CCE, Backbone=CNN2026.0598.73——
OursP(s=1)=0.7, Base loss=MAE, Backbone=CNN2026.0598.73——
OursP(s=1)=0.3, Base loss=MAE, Backbone=CNN2026.0598.71——
SCFF# Layers=52025.0198.7——
JeffreyNoise Type=Random, Noise Level=0.7, Bias Correction=false2020.1198.67——
Trans-MCL (MAE)P(s=1)=0.5, Base loss=MAE, Backbone=CNN2026.0598.66——
Trans-PRODENP(s=1)=0.3, Base loss=CCE, Backbone=CNN2026.0598.64——
SNN (rate coding)Network=784-800-10, Coding=SNN (rate coding), BP scheme=surrogate-gradient, Spike type=multi, Leakage=leaky, τV=τ, τI=τ2022.1198.64——
Trans-PRODENP(s=1)=0.7, Base loss=CCE, Backbone=CNN2026.0598.62——
DMINoise Type=Sparse, Noise Level=Low, Bias Correction=false2020.1198.59——
DFA*# Layers=42025.0198.53——
DiffLogic Net (largest)Model=FFN, Gate count=384 K2026.0298.47——
LogicTreeNet-SModel=CNN, Gate count=147 K2026.0298.46——
LILogicNet-MModel=FFN, Gate count=8 K2026.0298.45——
timing-based backpropagation for multi-spike SNNsNetwork=784-400-10, Coding=SNN (temporal coding), BP scheme=timing-based, Spike type=multi, Leakage=leaky, τV=τ, τI=2τ2022.1198.43——
ANNNetwork=784-800-10 (+ dropout), Coding=ANN2022.1198.4——
GINHidden dimension=64, Depth=3, Batch size=16, Dropout=0.1, Global mean pooling=true, Epochs=200, Optimizer=Adam, Learning rate=3 × 10−4, Number of random seeds=52026.0598.33——
PEPITA# Layers=22025.0198.29——
Trans-MCL (EXP)P(s=1)=0.5, Base loss=EXP, Backbone=CNN2026.0598.27——
timing-based backpropagation for multi-spike SNNsNetwork=784-400-10, Coding=SNN (temporal coding), BP scheme=timing-based, Spike type=multi, Leakage=non-leaky*, τV=τ, τI=2τ2022.1198.23——
Trans-MCL (EXP)P(s=1)=0.7, Base loss=EXP, Backbone=CNN2026.0598.12——
Trans-MCL (EXP)P(s=1)=0.3, Base loss=EXP, Backbone=CNN2026.0598.1——
SNN (temporal coding)Network=784-500-10, Coding=SNN (temporal coding), BP scheme=timing-based, Spike type=single, Leakage=non-leaky, τV=∞, τI=∞2022.1197.99——
timing-based backpropagation for SNNsNetwork=784-400-10, Coding=SNN (temporal coding), BP scheme=timing-based, Spike type=single, Leakage=leaky, τV=τ, τI=2τ2022.1197.99——
LILogicNet-SModel=FFN, Gate count=4 K2026.0297.96——
SNN (temporal coding)Network=784-340-10, Coding=SNN (temporal coding), BP scheme=timing-based, Spike type=single, Leakage=leaky, τV=τ, τI=τ2022.1197.96——
timing-based backpropagation for SNNsNetwork=784-400-10, Coding=SNN (temporal coding), BP scheme=timing-based, Spike type=single, Leakage=non-leaky*, τV=τ, τI=2τ2022.1197.96——
FLCNoise Type=Sparse, Noise Level=Low, Bias Correction=false2020.1197.76——
DiffLogic Net (small)Model=FFN, Gate count=48 K2026.0297.69——
SNN (temporal coding)Network=784-800-10, Coding=SNN (temporal coding), BP scheme=timing-based, Spike type=single, Leakage=non-leaky, τV=∞, τI=τ2022.1197.55——
SNN (temporal coding)Network=256-246-10, Coding=SNN (temporal coding), BP scheme=surrogate-gradient, Spike type=multi, Leakage=leaky, τV=τ, τI=τ2022.1197.5——
BLCNoise Type=Sparse, Noise Level=Low, Bias Correction=false2020.1197.37——
DRTP# Layers=22025.0197.32——
PLNoise Type=Sparse, Noise Level=Low, Bias Correction=false2020.1197.21——
SNN (temporal coding)Network=784-350-10, Coding=SNN (temporal coding), BP scheme=timing-based, Spike type=single, Leakage=leaky, τV=τ, τI=2τ2022.1197.2——
GATHidden dimension=64, Depth=3, Batch size=16, Dropout=0.1, Global mean pooling=true, Epochs=200, Optimizer=Adam, Learning rate=3 × 10−4, Number of random seeds=52026.0595.94——
BitLogicModel=CNN, Gate count=253.4 K2026.0295.72——
ChebConvHidden dimension=64, Depth=3, Batch size=16, Dropout=0.1, Global mean pooling=true, Epochs=200, Optimizer=Adam, Learning rate=3 × 10−4, Number of random seeds=52026.0595.72——
Trans-MCL (MAE)P(s=1)=0.3, Base loss=MAE, Backbone=CNN2026.0595.49——
Trans-MCL (MAE)P(s=1)=0.7, Base loss=MAE, Backbone=CNN2026.0595.47——
CENoise Type=Sparse, Noise Level=Low, Bias Correction=false2020.1195.23——
Trans-PLL-AvgP(s=1)=0.7, Base loss=MAE, Backbone=CNN2026.0594.85——
Trans-PLL-AvgP(s=1)=0.5, Base loss=MAE, Backbone=CNN2026.0594.78——
Trans-PLL-AvgP(s=1)=0.3, Base loss=MAE, Backbone=CNN2026.0593.8——
GCNHidden dimension=64, Depth=3, Batch size=16, Dropout=0.1, Global mean pooling=true, Epochs=200, Optimizer=Adam, Learning rate=3 × 10−4, Number of random seeds=52026.0592.09——
PearsonNoise Type=Sparse, Noise Level=High, Bias Correction=false2020.1158.63——
ADGMData augmentation=false2017.05——962
CAEN=1002015.06—13.47—
CAEN=6002015.06—6.3—
CAEN=10002015.06—4.77—
CAEN=30002015.06—3.22—
CatGANData augmentation=false2017.05——19,110
CDBN2014.04—0.82—
Conv. GPInput Dimension=25-dim2019.05—2.1—
Conv. Maxout + Dropout2014.04—0.45—
ConvnetN=1002015.06—22.98—
ConvnetN=6002015.06—7.86—
ConvnetN=10002015.06—6.45—
ConvnetN=30002015.06—3.35—
ConvNet2014.04—0.53—
FMDiscriminator architecture=small, Data augmentation=false2017.05——936.5
GCGPInput Dimension=24-dim2019.05—1.7—
HSC2014.04—0.77—
K-NN-IDM2014.04—0.54—
K-NN-SCM2014.04—0.63—
Ladder networkData augmentation=false2017.05——10,637
LDANet-1Filter size (k1, k2)=7, 72014.04—0.98—
LDANet-2Filter size (k1, k2)=7, 72014.04—0.62—
MTCN=1002015.06—12.03—
MTCN=6002015.06—5.13—
MTCN=10002015.06—3.64—
MTCN=30002015.06—2.57—
PCANet-1Filter size (k1, k2)=7, 72014.04—0.94—
PCANet-1 (k1 = 13)Filter size (k1, k2)=13, 72014.04—0.62—
PCANet-2Filter size (k1, k2)=7, 72014.04—0.66—
PL-DAEN=1002015.06—10.49—
PL-DAEN=6002015.06—5.03—