PropertyValue
?:abstract
  • Abstract: This work proposes a time-mode translinear principle as a process and voltage scalable technique for nonlinear analog computation that is independent of transistor physics. The principle utilizes the exponential relationships between a variable of interest and time in a first order system to realize nonlinear analog computation. A voltage-mode analog multiplier using this principle is introduced. The circuit is then reformulated as a purely time-mode multiplier to reduce the power consumption of the circuit itself as well as the analog-to-digital interface. Next, an asynchronous, differential implementation is proposed for further reducing the power consumption and boosting the dynamic range. Other useful linear and nonlinear circuit functions are also explored. Finally, the design of an application system, an analog neuromorphic image saliency algorithm, is described to motivate the new technique for pattern recognition hardware. (xsd:string)
?:author
?:comment
  • http://hdl.handle.net/10427/013614. (ISSP) (xsd:string)
?:dataSource
  • ISSP-Bibliography (xsd:string)
?:dateModified
  • 2017 (xsd:gyear)
?:datePublished
  • 2017 (xsd:gyear)
?:duplicate
?:fromPage
  • 182 (xsd:string)
is ?:hasPart of
?:inLanguage
  • english (xsd:string)
is ?:mainEntity of
?:name
  • A Time-Mode Translinear Principle and its Application to Analog Hardware for Pattern Recognition and Machine Learning (xsd:string)
?:publicationType
  • phdthesis (xsd:string)
?:sourceInfo
  • 182, 2017 (xsd:string)
  • Bibsonomy (xsd:string)
?:studyGroup
  • International Social Survey Programme (ISSP) (xsd:string)
?:tags
  • 2017 (xsd:string)
  • FDZ_IUP (xsd:string)
  • ISSP (xsd:string)
  • ISSP_input2019 (xsd:string)
  • ISSP_pro (xsd:string)
  • checked (xsd:string)
  • english (xsd:string)
  • isspbib2019 (xsd:string)
  • lieferungISSP (xsd:string)
  • phdthesis (xsd:string)
  • wa (xsd:string)
?:toPage
  • 182 (xsd:string)
rdf:type
?:url