Zobrazeno 1 - 10
of 79
pro vyhledávání: '"Bar-Hillel, Aharon"'
We propose a novel formulation of deep networks that do not use dot-product neurons and rely on a hierarchy of voting tables instead, denoted as Convolutional Tables (CT), to enable accelerated CPU-based inference. Convolutional layers are the most t
Externí odkaz:
http://arxiv.org/abs/2304.11706
Autor:
Mann, Hanan, Bar Hillel, Aharon, Lev-Tzion, Raffi, Greenfeld, Shira, Kariv, Revital, Lederman, Natan, Matz, Eran, Dotan, Iris, Turner, Dan, Lerner, Boaz
Publikováno v:
In Artificial Intelligence In Medicine November 2023 145
Most of industrial robotic assembly tasks today require fixed initial conditions for successful assembly. These constraints induce high production costs and low adaptability to new tasks. In this work we aim towards flexible and adaptable robotic ass
Externí odkaz:
http://arxiv.org/abs/1809.10699
Publikováno v:
In Ultrasonics February 2022 119
Akademický článek
Tento výsledek nelze pro nepřihlášené uživatele zobrazit.
K zobrazení výsledku je třeba se přihlásit.
K zobrazení výsledku je třeba se přihlásit.
Publikováno v:
In Computers and Electronics in Agriculture October 2021 189
We study classifiers operating under severe classification time constraints, corresponding to 1-1000 CPU microseconds, using Convolutional Tables Ensemble (CTE), an inherently fast architecture for object category recognition. The architecture is bas
Externí odkaz:
http://arxiv.org/abs/1602.04489
Publikováno v:
In Computers and Electronics in Agriculture September 2020 176
Autor:
Mor, Etai, Bar-Hillel, Aharon
Publikováno v:
In Ultrasonics April 2020 103
Collecting large labeled data sets is a laborious and expensive task, whose scaling up requires division of the labeling workload between many teachers. When the number of classes is large, miscorrespondences between the labels given by the different
Externí odkaz:
http://arxiv.org/abs/0903.1125