2020-2022 YAMAGATA UNIVERSITY Research Seeds Collection
47/90

Cantilever typeOrganic EL typePd/Ge Schottoky Diode45IllustrationFig. 1 Modeling a neural circuit as a logic circuit Fig. 2.(a) Threshold gate; and (b) Threshold circuitInputWeightOutputThresholdInputThreshold gateYamagata University Graduate School of Science and Engineering Tel :+81-238-26-3282 Research InterestFax:+81-238-26-3299ElectronicsE-mail ・ sumio@ieee.orgTel ・ +81-238-26-3282 Fax ・ +81-238-26-3299Content: A logic circuit is a computational model, and consists of Alogiccircuitisacomputationalmodel,andconsistsofbasicbasic elements called logic gates which output one or zero. elementscalledlogicgateswhichoutputoneorzero.WhilealogicWhile a logic circuit plays central role in standard computers, circuitplayscentralroleinstandardcomputers,wecanalsomodelawe can also model a neural network in the brain as a logic neuralnetworkinthebrainasalogiccircuit,sinceitalsoconsistsofcircuit, since it also consists of basic elements called basicelementscalledneuronswhichtransmitsaspikeornot(Fig.1).neurons which transmits a spike or not (Fig. 1). We can thus Wecanthusinvestigateneuralnetworksintermsofcomputationalinvestigate neural networks in terms of computational complexitytheory.Athresholdlogiccircuitisasimpletheoreticalcomplexity theory. A threshold logic circuit is a simple modelofaneuralnetwork,andconsistsofthresholdgatestheoretical model of a neural network, and consists of computinglinearthresholdfunctions(Fig.2).threshold gates computing linear threshold functions (Fig. 2).Weareinterestedinthecomputationalpowerofthresholdcircuits. We are interested in the computational power of threshold Inparticular,westudywhattypesofinformationprocessingiscircuits. In particular, we study what types of information efficientlycarriedoutbythresholdcircuits,wheretheefficiencyisprocessing is efficiently carried out by threshold circuits, measuredbythenumberofgates,computationtime,energywhere the efficiency is measured by the number of gates, consumption,etc.computation time, energy consumption, etc. Appealingpoint:tounderstandinformationprocessingintheOurresearchaimisbrainfromtheviewpointoftheoreticalcomputerscience. Our research aim is to understand information processing in the brain from the viewpoint of theoretical computer science.Yamagata UniversityGraduate School of Science and Engineering Research Interest :Computational complexityYamagata University Graduate School of Science and Engineering Research InterestComputational complexityE-mail :uchizawa@yz.yamagata-uac.jpTel :+81-238-26-3310E-mail ・ uchizawa@yz.yamagata-u.ac.jpTel ・ +81-238-26-3310Yamagata UniversityGraduate School of Science and Engineering Research Interest :ElectronicsE-mail:sumio@ieee.orgContentSpecial objectivesOutputContentContent:Hydrogengasispromisingasanenergyinthenextgeneration. Hydrogen gas is promising as an energy in the next However,becausehydrogengasisexplosive,hydrogengassensorsgeneration. However, because hydrogen gas is explosive, areindispensabletousehydrogengassafely.Westudyvarioushydrogen gas sensors are indispensable to use hydrogen gas hydrogensensors:safely. We study various hydrogen sensors:(1)Hydrogengassensorsusingpalladium.Cantilevertypehydrogen(1) Hydrogen gas sensors using palladium. Cantilever type gassensorsutilizeexpansionofpalladiumbyabsorptionofhydrogenhydrogen gas sensors utilize expansion of palladium by gas.OrganicELhydrogengassensorsusingPdcathodeemitslightabsorption of hydrogen gas. Organic EL hydrogen gas whenitisonlyinhydrogengasatmosphere.sensors using Pd cathode emits light when it is only in (2)Micromachinedhydrogengassensor.hydrogen gas atmosphere.(3)Pd/GeSchottokybarrierdiodehydrogengassensor.(2) Micromachined hydrogen gas sensor.(3) Pd/Ge Schottoky barrier diode hydrogen gas sensor.Hydrogen Gas Sensor using PdMicromachined Hydrogen Gas SensorHydrogen Gas Sensor using SemiconductorComputational Complexity Theory for Neural NetworksAssociate Professor Kei UchizawaAssociate Professor Kei UchizawaAssociate Professor SumioOkuyamaComputational Complexity Theory for Neural NetworksDevelopment of Hydrogen Gas Sensors and MicromachinedGas SensorsDevelopment of Hydrogen Gas Sensors and Micromachined Gas SensorsAssociate Professor Sumio Okuyama

元のページ  ../index.html#47

このブックを見る