Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Roadmap to Neuromorphic Computing with Emerging Technologies

Adnan Mehonic, Daniele Ielmini, Kaushik Roy, Onur Mutlu, Shahar Kvatinsky, Teresa Serrano-Gotarredona, Bernabe Linares-Barranco, Sabina Spiga, Sergey Savelev, Alexander G Balanov, Nitin Chawla, Giuseppe Desoli, Gerardo Malavena, Christian Monzio Compagnoni, Zhongrui Wang, J Joshua Yang, Ghazi Sarwat Syed, Abu Sebastian, Thomas Mikolajick, Beatriz Noheda, Stefan Slesazeck, Bernard Dieny, Tuo-Hung, Hou, Akhil Varri, Frank Bruckerhoff-Pluckelmann, Wolfram Pernice, Xixiang Zhang, Sebastian Pazos, Mario Lanza, Stefan Wiefels, Regina Dittmann, Wing H Ng, Mark Buckwell, Horatio RJ Cox, Daniel J Mannion, Anthony J Kenyon, Yingming Lu, Yuchao Yang, Damien Querlioz, Louis Hutin, Elisa Vianello, Sayeed Shafayet Chowdhury, Piergiulio Mannocci, Yimao Cai, Zhong Sun, Giacomo Pedretti, John Paul Strachan, Dmitri Strukov, Manuel Le Gallo, Stefano Ambrogio, Ilia Valov, Rainer Waser

Abstract

The roadmap is organized into several thematic sections, outlining current computing challenges, discussing the neuromorphic computing approach, analyzing mature and currently utilized technologies, providing an overview of emerging technologies, addressing material challenges, exploring novel computing concepts, and finally examining the maturity level of emerging technologies while determining the next essential steps for their advancement.

Roadmap to Neuromorphic Computing with Emerging Technologies

Abstract

The roadmap is organized into several thematic sections, outlining current computing challenges, discussing the neuromorphic computing approach, analyzing mature and currently utilized technologies, providing an overview of emerging technologies, addressing material challenges, exploring novel computing concepts, and finally examining the maturity level of emerging technologies while determining the next essential steps for their advancement.
Paper Structure (112 sections, 3 equations, 22 figures)

This paper contains 112 sections, 3 equations, 22 figures.

Table of Contents

  1. 1. Introduction to the roadmap
  2. 1.1. Taxonomy & Motivation
  3. 2. Computing challenges
  4. 2.1. Digital computing
  5. 3. Neuromorphic computing basics and its evolution
  6. 3.1 What is neuromorphic computing/engineering
  7. 3.2 Different neuromorphic technologies and state of the art
  8. 3.3. Possible future computational primitives for neuromorphic computing
  9. 4. Mature Technologies
  10. 4.1 SRAM
  11. 4.2 Flash memories
  12. 5. Emerging Technologies (computing approaches)
  13. 5.1 Resistive switching and memristor
  14. 5.2 Phase change materials
  15. 5.3 Ferroelectric materials
  16. ...and 17 more sections

Figures (22)

  • Figure 1: a) Increase in computing power demands to run state-of-the-art AI models. b) The cost associated with training AI models. Adapted and reproduced from \ref{['csl:1']} .
  • Figure 2: (a) 7KHz spiral observed in a classic phosphor oscilloscope set in X/Y mode. (b) DVS output event stream when observing the oscilloscope in (a).
  • Figure 3: (a) Fast speed poker deck browsing: events are collected about every 20ms to build a frame to display on a computer screen. (b) Slow speed playback at 77us per reconstructed frame. (c) Poker symbol tracked and displayed on the right, and recognition output on the left. (d) Event-driven CNN to classify four poker symbols. (e) {x,y,t} display during 20ms showing camera events together with recognition events during a change of card with a recognition of less than 2ms.
  • Figure 4: Schematic evolution of the main hardware technologies of interest for neuromorphic computing (the indicates decades represent only a time frame). Triangular symbols mark the refernce period for early stage studies or starting interest in the technology development. From bottom to top of the figure, the listed technologies are today at higher maturity level and more advanced at system integration level.
  • Figure 5: Examples of materials systems currently employed in memristive technologies. The list of materials is not exhastive andinclude only some of the most used ones. For the NVM devices(top line), the main active material is indicated, but each device includes also various types of material electrodes depending on the technology.
  • ...and 17 more figures