scholarly journals FORSA: Exploiting Filter Ordering to Reduce Switching Activity for Low Power CNNs

2021 ◽  
Author(s):  
Yasmin Halawani ◽  
Baker Mohammad
Keyword(s):  
Digital Circuits ◽  
Matrix Multiplication ◽  
Temporal Distribution ◽  
Cross Coupling ◽  
Model Structure ◽  
Switching Activity ◽  
Data Set ◽  
Dynamic Power ◽  
Coupling Capacitance ◽  
Point Representation

<div>Switching activity in digital circuits depends on the temporal distribution of the data participating in the operation which directly influences the interconnect, dynamic power and timing of the system. This paper proposes an efficient method for reducing both power and latency of matrix-multiplication operations found in many applications like convolution neural networks (CNNs). The approach takes advantage of the unique characterizes of CNN with input stationary for efficient multiply-add operation. Since most application use reduce accuracy for MAC, the proposed work assumed 8-bit fixed point representation. As a demonstrator, CIFAR-10 data set has been used for end to end analysis of the filters on a 3-ConV with 2-FC model structure. The filters’ were re-ordered to reduce the switching behaviour between successive weight fetching. This directly impacts the dynamic power consumption and miraculously makes the classification activity reduces cross-coupling capacitance which helps improve timing and noise.</div>

scholarly journals FORSA: Exploiting Filter Ordering to Reduce Switching Activity for Low Power CNNs

2021 ◽  
Author(s):  
Yasmin Halawani ◽  
Baker Mohammad
Keyword(s):  
Digital Circuits ◽  
Matrix Multiplication ◽  
Temporal Distribution ◽  
Cross Coupling ◽  
Model Structure ◽  
Switching Activity ◽  
Data Set ◽  
Dynamic Power ◽  
Coupling Capacitance ◽  
Point Representation

<div>Switching activity in digital circuits depends on the temporal distribution of the data participating in the operation which directly influences the interconnect, dynamic power and timing of the system. This paper proposes an efficient method for reducing both power and latency of matrix-multiplication operations found in many applications like convolution neural networks (CNNs). The approach takes advantage of the unique characterizes of CNN with input stationary for efficient multiply-add operation. Since most application use reduce accuracy for MAC, the proposed work assumed 8-bit fixed point representation. As a demonstrator, CIFAR-10 data set has been used for end to end analysis of the filters on a 3-ConV with 2-FC model structure. The filters’ were re-ordered to reduce the switching behaviour between successive weight fetching. This directly impacts the dynamic power consumption and miraculously makes the classification activity reduces cross-coupling capacitance which helps improve timing and noise.</div>

Automated Region-Based Vehicle Conflict Detection Using Computer Vision Techniques

2015 ◽  
Vol 2528 (1) ◽  
pp. 49-59
Author(s):  
Jinling Li ◽  
Yuhao Liu ◽  
Ahmed Tageldin ◽  
Mohamed H. Zaki ◽  
Greg Mori ◽  
...  
Keyword(s):  
Computer Vision ◽  
Three Dimensional ◽  
Data Set ◽  
Time To Collision ◽  
Video Footage ◽  
Precise Calculation ◽  
Vehicle Trajectories ◽  
Traffic Conflict ◽  
Point Representation ◽  
Feature Based

An approach for vehicle conflict analysis based on three-dimensional (3-D) vehicle detection is presented. Techniques for quantitative conflict measurements often use a point trajectory representation for vehicles. More accurate conflict measurement can be facilitated with a region-based vehicle representation instead. This paper describes a computer vision approach for extracting vehicle trajectories from video sequences. The method relied on a fusion of background subtraction and feature-based tracking to provide a three-dimensional (3-D) cuboid representation of the vehicle. Standard conflict measures, including time to collision and postencroachment time, were computed with the use of the 3-D cuboid vehicle representations. The use of these conflict measures was demonstrated on a challenging data set of video footage. Results showed that the region-based representation could provide more precise calculation of traffic conflict indicators compared with approaches based on a point representation.

Subword Switching Activity Minimization to Optimize Dynamic Power Consumption

2009 ◽  
Vol 26 (4) ◽  
pp. 68-77
Author(s):  
M.C. Molina ◽  
R. Ruiz-Sautua ◽  
A. Del Barrio ◽  
J.M. Mendias
Keyword(s):  
Power Consumption ◽  
Switching Activity ◽  
Dynamic Power

Novel SAT-Based Peak Dynamic Power Estimation for Digital Circuits

2009 ◽  
Vol 5 (4) ◽  
pp. 429-438 ◽  
Author(s):  
K. Shyamala ◽  
J. Vimalkumar ◽  
V. Kamakoti
Keyword(s):  
Digital Circuits ◽  
Power Estimation ◽  
Dynamic Power

scholarly journals Using treemaps for variable selection in spatio-temporal visualisation

2008 ◽  
Vol 7 (3-4) ◽  
pp. 210-224 ◽  
Author(s):  
Aidan Slingsby ◽  
Jason Dykes ◽  
Jo Wood
Keyword(s):  
Variable Selection ◽  
Visual Analysis ◽  
Temporal Distribution ◽  
Temporal Data ◽  
Data Set ◽  
Raster Maps ◽  
Categorical Attributes ◽  
Interactive Interfaces ◽  
Road Maps ◽  
Spatio Temporal

We demonstrate and reflect upon the use of enhanced treemaps that incorporate spatial and temporal ordering for exploring a large multivariate spatio-temporal data set. The resulting data-dense views summarise and simultaneously present hundreds of space-, time-, and variable-constrained subsets of a large multivariate data set in a structure that facilitates their meaningful comparison and supports visual analysis. Interactive techniques allow localised patterns to be explored and subsets of interest selected and compared with the spatial aggregate. Spatial variation is considered through interactive raster maps and high-resolution local road maps. The techniques are developed in the context of 42.2 million records of vehicular activity in a 98 km2 area of central London and informally evaluated through a design used in the exploratory visualisation of this data set. The main advantages of our technique are the means to simultaneously display hundreds of summaries of the data and to interactively browse hundreds of variable combinations with ordering and symbolism that are consistent and appropriate for space- and time-based variables. These capabilities are difficult to achieve in the case of spatio-temporal data with categorical attributes using existing geovisualisation methods. We acknowledge limitations in the treemap representation but enhance the cognitive plausibility of this popular layout through our two-dimensional ordering algorithm and interactions. Patterns that are expected (e.g. more traffic in central London), interesting (e.g. the spatial and temporal distribution of particular vehicle types) and anomalous (e.g. low speeds on particular road sections) are detected at various scales and locations using the approach. In many cases, anomalies identify biases that may have implications for future use of the data set for analyses and applications. Ordered treemaps appear to have potential as interactive interfaces for variable selection in spatio-temporal visualisation.

scholarly journals Comparison of mixed layer heights from airborne high spectral resolution lidar, ground-based measurements, and the WRF-Chem model during CalNex and CARES

2013 ◽  
Vol 13 (5) ◽  
pp. 13721-13772 ◽  
Author(s):  
A. J. Scarino ◽  
M. D. Obland ◽  
J. D. Fast ◽  
S. P. Burton ◽  
R. A. Ferrare ◽  
...  
Keyword(s):  
Los Angeles ◽  
Mixed Layer ◽  
Spectral Resolution ◽  
Temporal Distribution ◽  
Carbonaceous Aerosol ◽  
High Spectral Resolution ◽  
Data Set ◽  
Field Campaigns ◽  
Langley Research Center ◽  
High Spectral Resolution Lidar

Abstract. The California Research at the Nexus of Air Quality and Climate Change (CalNex) and Carbonaceous Aerosol and Radiative Effects Study (CARES) field campaigns during May and June 2010 provided a data set appropriate for studying characteristics of the planetary boundary layer (PBL). The NASA Langley Research Center (LaRC) airborne High Spectral Resolution Lidar (HSRL) was deployed to California onboard the NASA LaRC B-200 aircraft to aid in characterizing aerosol properties during these two field campaigns. Measurements of aerosol extinction (532 nm), backscatter (532 and 1064 nm), and depolarization (532 and 1064 nm) profiles during 31 flights, many in coordination with other research aircraft and ground sites, constitute a diverse data set for use in characterizing the spatial and temporal distribution of aerosols, as well as the depth and variability of the daytime mixed layer (ML), which is a subset within the PBL. This work illustrates the temporal and spatial variability of the ML in the vicinity of Los Angeles and Sacramento, CA. ML heights derived from HSRL measurements are compared to PBL heights derived from radiosonde profiles, ML heights measured from ceilometers, and simulated PBL heights from the Weather Research and Forecasting Chemistry (WRF-Chem) community model. Comparisons between the HSRL ML heights and the radiosonde profiles in Sacramento result in a correlation coefficient value (R) of 0.93 (root-mean-square (RMS) difference of 157 m and bias difference (HSRL – radiosonde) of 57 m). HSRL ML heights compare well with those from the ceilometer in the LA Basin with an R of 0.89 (RMS difference of 108 m and bias difference (HSRL – Ceilometer) of −9.7 m) for distances of up to 30 km between the B-200 flight track and the ceilometer site. Simulated PBL heights from WRF-Chem were compared with those obtained from all flights for each campaign, producing an R of 0.58 (RMS difference of 604 m and a bias difference (WRF-Chem – HSRL) of −157 m) for CalNex and 0.59 (RMS difference of 689 m and a bias difference (WRF-Chem – HSRL) of 220 m) for CARES. Aerosol backscatter simulations are also available from WRF-Chem and are compared to those from HSRL to examine differences among the methods used to derive ML heights.

Sign in / Sign up

Export Citation Format

Share Document