Inventor · disambiguated record
Eric A. Sather
Also filed as: SATHER ERIC · SATHER ERIC A
40 granted patents·3 pending applications·300 citations·filing 2002–2024
97Inventor score
Top patents by PatentIndex Score
43 records- 0198US11610154B1Preventing overfitting of hyperparameters during training of networkPERCEIVE CORP·Filed 2020·Granted Mar 21, 2023·22 cites·18 claims
- 0298US10592732B1Probabilistic loss function for training network with tripletsPERCEIVE CORP·Filed 2018·Granted Mar 17, 2020·40 cites·17 claims
- 0397US12136039B1Optimizing global sparsity for neural networkPERCEIVE CORP·Filed 2020·Granted Nov 5, 2024·12 cites·19 claims
- 0497US11847567B1Loss-aware replication of neural network layersPERCEIVE CORP·Filed 2020·Granted Dec 19, 2023·11 cites·21 claims
- 0596US12061988B1Decomposition of ternary weight tensorsPERCEIVE CORP·Filed 2020·Granted Aug 13, 2024·6 cites·24 claims
- 0696US11163986B2Using batches of training items for training a networkPERCEIVE CORP·Filed 2020·Granted Nov 2, 2021·4 cites·20 claims
- 0795US12045725B1Batch normalization for replicated layers of neural networkPERCEIVE CORP·Filed 2020·Granted Jul 23, 2024·5 cites·22 claims
- 0895US11900238B1Removing nodes from machine-trained network based on introduction of probabilistic noise during trainingPERCEIVE CORP·Filed 2020·Granted Feb 13, 2024·6 cites·14 claims
- 0995US11741369B2Using batches of training items for training a networkPERCEIVE CORP·Filed 2021·Granted Aug 29, 2023·3 cites·20 claims
- 1095US11604973B1Replication of neural network layersPERCEIVE CORP·Filed 2019·Granted Mar 14, 2023·11 cites·20 claims
- 1195US11531879B1Iterative transfer of machine-trained network inputs from validation set to training setPERCEIVE CORP·Filed 2019·Granted Dec 20, 2022·16 cites·20 claims
- 1295US11151695B1Video denoising using neural networks with spatial and temporal featuresPERCEIVE CORP·Filed 2019·Granted Oct 19, 2021·13 cites·20 claims
- 1394US12112254B1Optimizing loss function during training of networkPERCEIVE CORP·Filed 2020·Granted Oct 8, 2024·5 cites·20 claims
- 1492US10671888B1Using batches of training items for training a networkPERCEIVE CORP·Filed 2018·Granted Jun 2, 2020·6 cites·22 claims
- 1591US11537870B1Training sparse networks with discrete weight valuesPERCEIVE CORP·Filed 2018·Granted Dec 27, 2022·8 cites·17 claims
- 1690US8352296B2Managing real time meeting room statusMICROSOFT CORP·Filed 2008·Granted Jan 8, 2013·71 cites·10 claims
- 1788US12061981B1Decomposition of weight tensors in network with value quantizationPERCEIVE CORP·Filed 2020·Granted Aug 13, 2024·2 cites·22 claims
- 1887US11113603B1Training network with discrete weight valuesPERCEIVE CORP·Filed 2017·Granted Sep 7, 2021·5 cites·23 claims
- 1985US11017295B1Device storing ternary weight parameters for machine-trained networkPERCEIVE CORP·Filed 2017·Granted May 25, 2021·4 cites·19 claims
- 2083US12462350B2Circuit for executing stateful neural networkAMAZON TECH INC·Filed 2024·Granted Nov 4, 2025·0 cites·20 claims
- 2183US8789001B1System and method for using fabric-graph flow to determine resource costsSATHER ERIC A·Filed 2013·Granted Jul 22, 2014·8 cites·22 claims
- 2282US11429861B1Device storing multiple sets of parameters for machine-trained networkPERCEIVE CORP·Filed 2017·Granted Aug 30, 2022·3 cites·20 claims
- 2381US12248880B2Using batches of training items for training a networkAMAZON TECH INC·Filed 2023·Granted Mar 11, 2025·0 cites·20 claims
- 2480US2025028945A1Executing replicated neural network layers on inference circuitPERCEIVE CORP·Filed 2024·Application pending·0 cites
- 2579US11995533B1Executing replicated neural network layers on inference circuitPERCEIVE CORP·Filed 2019·Granted May 28, 2024·2 cites·19 claims
- 2679US11494657B2Quantizing neural networks using approximate quantization functionPERCEIVE CORP·Filed 2019·Granted Nov 8, 2022·2 cites·20 claims
- 2778US11586902B1Training network to minimize worst case surprisePERCEIVE CORP·Filed 2018·Granted Feb 21, 2023·3 cites·20 claims
- 2878US7385420B1Repeatable block producing a non-uniform routing architecture in a field programmable gate array having segmented tracksACTEL CORP·Filed 2006·Granted Jun 10, 2008·6 cites·13 claims
- 2978US2024193426A1Quantizing Neural Networks Using Shifting and ScalingPERCEIVE CORP·Filed 2023·Application pending·0 cites
- 3077US12299555B2Training network with discrete weight valuesAMAZON TECH INC·Filed 2022·Granted May 13, 2025·0 cites·18 claims
- 3176US9300912B2Software based whiteboard capture solution for conference room meetingsTANEJA PUNEET·Filed 2008·Granted Mar 29, 2016·13 cites·20 claims
- 3274US11868871B1Circuit for executing stateful neural networkPERCEIVE CORP·Filed 2019·Granted Jan 9, 2024·1 cites·20 claims
- 3374US11620495B1Neural networks with spatial and temporal featuresPERCEIVE CORP·Filed 2019·Granted Apr 4, 2023·1 cites·20 claims
- 3474US9047431B2System and method for using fabric-graph flow to determine resource costsTABULA INC·Filed 2014·Granted Jun 2, 2015·2 cites·22 claims
- 3572US12175368B2Training sparse networks with discrete weight valuesPERCEIVE CORP·Filed 2022·Granted Dec 24, 2024·0 cites·18 claims
- 3670US12367661B1Weighted selection of inputs for training machine-trained networkAMAZON TECH INC·Filed 2022·Granted Jul 22, 2025·0 cites·20 claims
- 3768US7075334B1Repeatable block producing a non-uniform routing architecture in a field programmable gate array having segmented tracksACTEL CORP·Filed 2005·Granted Jul 11, 2006·3 cites·3 claims
- 3861US11847568B2Quantizing neural networks using shifting and scalingPERCEIVE CORP·Filed 2019·Granted Dec 19, 2023·0 cites·20 claims
- 3957US6891396B1Repeatable block producing a non-uniform routing architecture in a field programmable gate array having segmented tracksACTEL CORP·Filed 2002·Granted May 10, 2005·6 cites·8 claims
- 4052US7579869B2Repeatable block producing a non-uniform routing architecture in a field programmable gate array having segmented tracksACTEL CORP·Filed 2008·Granted Aug 25, 2009·0 cites·18 claims
- 4151US2015324508A1System and Method for Using Fabric-Graph Flow to Determine Resource CostsALTERA CORP·Filed 2015·Application pending·0 cites
- 4246US12165066B1Training network to maximize true positive rate at low false positive ratePERCEIVE CORP·Filed 2018·Granted Dec 10, 2024·0 cites·18 claims
- 4344US11995537B1Training network with batches of input instancesPERCEIVE CORP·Filed 2018·Granted May 28, 2024·0 cites·18 claims
Identity basis: PatentsView inventor disambiguation (2025Q4-odp release). How scoring works →