! E L M E N T SE "#! $!%&' &( "#! )!*(+, !"#$%&!'(&!"' &)$ $#*+,&!*" *- &)$ .$""/ .0$%% ("1 &)$ $22(",$+ 3),03) 2,01$0% ALSO FEATURING • The Digital Pandemic • The Force on the Front!"# $%&#'(')&$)!# '#*#)'+" ,-$'%). -/ 0-*!-% +-..#(#-. +/ ' 0 2 0 21 · 1 &* 2 % ! 1 7 · / -- 2 ! 1 !" #$% &'(% )& !"(*%'+!",-. -'*,% ()/01#'#!)"'- 2%/'"2+ '"2 #$% !/0%"2!", $'-# #) /))*%’+ -'3, #$% +%/!()"21(#)* !"21+#*. $'+ 4%%" &)*(%2 #) *%-%5'-1'#% #$% #*'2!- #!)"'- ()/01#!", 0'*'2!,/. (%"#*'- #) #$!+ *%-%5'-1'#!)" $'+ 4%%" #$% ")5%- 2%5%-- )0/%"# )& "%1*)/)*0$!( ()/01#!",–'" '00*)'($ #$'#, '# !#+ ()*%, +%%6+ #) *%0-!('#% #$% 4*'!" !" +!-!()". 2%+0!#% ($'--%",%+ )" #$% '-,)*!#$/!( &*)"#, "%1*)/)*0$!( ()/01#!", 0*)/!+%+ ' /'++!5%-. 0'*'--%-, %&&!(!%"#, '"2 +('-'4-% ()/01#'#!)"'- +)-1- #!)" 3!#$ -'*,% !/0-!('#!)"+ )" #$% 2'!-. -!5%+ )& ()"+1/%*+. #$% &1#1*% )& #$% #%($- ")-),., $)3%5%*, !+ 1"(%*#'!". 3!#$ #$% *!+% )& $!,$ 0%*&)*/'"(% '"2 71'"#1/ ()/- 01#!", '+ 0*)/!+!", '-#%*"'#!5%+, #$% +%/!()"21(#)* !"21+#*. '# -'*,% /1+# ()"+!2%* #$% %8#%"# #) 3$!($ "%1*)/)*0$!( ()/01#!", ('" %/%*,% '+ ' 5!'4-% '"2 &%'+!4-% +)-1#!)" !" #$% ()/!", .%'*+. !# !+ 5!#'-, #$%*%&)*%, #) 1"2%*+#'"2 #$% #$%)*%#!('- 1"2%*0!""!",+ )& #$% "%1*)/)*0$!( '00*)'($ '"2 0*%2!(# #$% -!6%-!$))2 )& !#+ !/- 0-%/%"#'#!)" 3!#$!" #$% "%8# 2%('2%. !"#$%&%$'()* *%&'#+)!, Explaining how Projected SNN Training will Largely Impact our Interactions with Technology 3!-- *!$%*2 ed on the algorithmic front. In terms of research and re- sources, in October of 2020, Intel announced a three-year contract with Sandia National Laboratories (a government nuclear research lab) to advance neuromorphic computing beyond edge applications to progressively complex compu- tational problems. The result will be the development of increasingly large neuromorphic systems with the intent of integrating over one billion neurons. Creating a system with such neural capacity exhibits moderate uncertainty in the short-term. Given a) the progress Intel has made when it comes to scaling Loihi and b) the di3culty of integrating an increasingly large number of neurons, the project has an uncertainty factor of around two on the low- and high- end (i.e., the system will be developed within one half to two times the expected implementation time frame). Neu- romorphic computing will also experience strong research in edge applications (particularly energy e3cient robotics) as members of the INRC continue to deploy Loihi in real- time audio and video processing tests.11 While Gartner predicts that traditional computing will hit a wall in 2025 (specifically mentioning neuromorphic computing as the leading technology for a shift in the com- puting paradigm), it is unlikely that a robust commercial product will be available within the next five years due to software limitations. Unlike traditional ANNs, SNNs lack a learning algorithm when it comes to parameter optimi- zation: it is impossible to use gradient descent given the non-di4erentiable nature of spiking neurons (Soni).12 Yann LeCun, the head of AI at Facebook, is a leading skep- tic when it comes to neuromorphic implementation, writ- ing that it is “premature to build a chip” when SNNs face a fundamental training challenge.13 There is also skepticism regarding the performance of the underlying SNNs; spe- cifically, skeptics point to the fact that SNNs do not outper- form ANNs (e.g. Convolutional Neural Networks) when it comes to traditional tests of image recognition (Pfei4er and Pfeil).14 Medium Term (5-10 Years) While SNNs currently face learning challenges, it is expected that within the next five to ten years, researchers will develop novel and e3cient approaches to SNN training. Current advancements have largely been focused on implementing a form of transfer learning (i.e., a process by which parameters are optimized in one “place” and transferred to a new “place”). In March of 2019, Terry Sejnowski of The Salk Institute of La Jolla and his team were able to train a standard recurrent neural network (RNN) via gradient descent methods and transfer the learned parameters to an SNN (Tiernan, “Neuromorphic computing finds new life”). While parameter transferring is a step in the right direction, Sejnowski admits that the next step—learning the SNN itself—is still “in the early days,” but that there is “going to be another big shift, which will probably occur within the next five to ten years.” Given both the di3culty and current progress of SNN training, there is moderate uncertainty regarding how quickly development will occur on the algorithmic front, with a general uncertainty factor of one and one-half to two times on the high-end. If an e3cient learning rule is established, it is possible we will begin to see commercially viable neuromorphic chips within the decade, allowing for complex, energy e3cient tasks on the “edge” (e.g., smartphones). Such commercial implementation, however, is highly dependent on the economic feasibility of widespread production, an aspect of neuromorphic chip development that is still in its infancy. Long Term (10+ Years) Although the future of SNN training and the economic practicality of neuromorphic hardware is uncertain in the near-term, it is possible that beyond 2030 we begin to wit- ness a shift in the current computing paradigm. Such a shift would be characterized by the wide-spread use of neuromorphic chips in energy-constrained edge devices (e.g., mobile devices, computers, robotics, etc.). Sejnowski predicts that once the software is developed and the hard- ware is “su3ciently cheap,” the implementation of such chips will be “ubiquitous...like sensors in phones.” Abu Sebastian, Principal Research Sta4 Member at IBM Zu- rich, predicts that neuromorphic computing will play a vi- tal role in the future feasibility of autonomous vehicles, making the point that you cannot “collect a frame, pass it )'!1/1+1). *2%)3*#* 0)*#& -% /#''--#.#+!'1+ !$%%#. ,$%+!1-%* (!"#$%&'( ") *+,+-&.+/ !"--"0') 35 !"#$%&%$'()* *%&'#+)!, through to a deep neural net, and wait for the response when you’re traveling down a freeway at 70 miles an hour” (Greengard). The long-term draw of neuromorphic com- puting, is energy-e3cient, on-board processing that is ap- plicable to a wide range of technologies, from autonomous vehicles to independently operating interstellar spacecraft. Neuromorphic chips are also poised to play a large role in future medical devices and artificial body parts that can benefit from increasingly fast, on-site processing. While neuromorphic architecture could become ubiqui- tous beyond 2030, it is important to note the false dichoto- my between standard computer architecture and neuro- morphic hardware.15 The promise of neuromorphic chips lies in the creation of an adaptive system capable of the e3cient processing of highly noisy, increasingly complex, spatio-temporal data, not the high-speed processing of pre- dictable, deterministic processes (Greengard). It is short- sighted, therefore, to view neuromorphic computing as a fundamental alternative to traditional computing: the two technologies are complementary to each other, excelling in tasks suited to their specific design. It is also important to note the uncertainty surrounding the long-term ubiquity of neuromorphic computing, namely, that there is a three- way race between neuromorphic computing, high perfor- mance computing, and quantum computing (Vorhies).16 Given the advancements of competing fields, it is possible the technology is outperformed before it is commercially viable (either by existing competitors or a novel approach yet to be developed). The future of computing, then, is largely undecided, with the only certainty being that our current computing paradigm will shift in response to the impending halt to Gordon Moore’s 1965 prediction. IMPLICATIONS Consumers and AI Integration Despite uncertainties regarding the path of computing, a future characterized by the widespread adoption of neuro- morphic technology would have significant implications on the way consumers interact with digital devices. Peter Suma, co-CEO of Applied Brain Research, imagines a world in which neuromorphic technology enables an ex- treme integration of artificial intelligence into our daily lives (Ferry). Suma describes a future in which a technol- ogy like Siri transcends basic voice commands (i.e., a Siri that listens and sees all of your conversations and interac- tions). If you were to ask Siri what idea your friend Melissa came up with regarding your wife’s birthday gift, the e3- cient, continuous, on-board processing of data made pos- sible by neuromorphic computing would enable Siri to re- call the conversation with your friend the week prior, bringing up both the idea and a variety of similar, person- alized ideas. The salient point in Suma’s eyes is that the e3ciency of neuromorphic chips would enable the local storage of information, resolving widespread privacy con- cerns regarding consumer data. From autonomous vehi- cles to smart homes and edge devices, Suma’s point is clear: the promise of neuromorphic technology is a future characterized by the extreme integration of artificial intel- ligence into our daily lives. Semiconductor Industry Research scientist at Hewlett Packard Enterprise, Suhas Kumar, summarizes the state of computing as a “huge rush to find something” that can continue the improve- ment in computer science that we have witnessed the past half century. Semiconductor manufacturers, therefore, are faced with a critical choice of whether to invest in risky chip research to compete in a future characterized by ad- vanced computing technologies. While IBM has seen suc- cess with its TrueNorth system, Intel has exhibited the strongest commitment to neuromorphic research, evi- denced by its continued progress in creating increasingly scalable neuromorphic systems. Intel’s competitors are also faced with the challenge of determining the type of research they wish to undertake (i.e., whether they believe high performance or quantum computing will ultimately outperform the neuromorphic approach). Regardless, the challenge is clear: the computing industry will witness a shift in the coming decades, and the companies that lead this shift will capitalize on the widespread adoption of their novel technology. “The promise of neuromorphic chips lies in the cration of an adaptive system capable of the e4cient processing of highly noisy, increasingly complex, spatio-temporal data” 36 !"!#!$%& : : &'()$* 2021 Large Technology Firms Beyond the semiconductor industry, large technology companies (e.g., Google, Facebook, Amazon, etc.) are faced with the decision to embark upon internal research regarding computing hardware. As deep learning process- es continue to surpass the capabilities of existing technol- ogy, companies with large workloads and available capital are faced with the decision of whether to invest internally in chip development. In February of 2019, Facebook’s Yann LeCun briefly described “internal activities” the firm was taking to address the growing computational divide, referencing Google’s Tensor Processing Unit as another example of the type of internal e4orts large technology firms are capable of taking (Tiernan, “Facebook’s Yann Le- Cun”). While the future is uncertain, it is clear there is a demand for change, and the feasibility of neuromorphic hardware should figure prominently in the minds of firm leaders who are considering undertaking or continuing in- ternal development activities. CONCLUSION From computer science and biology to mathematics, phys- ics, and electrical engineering, neuromorphic computing is a cross-disciplinary challenge that requires a fundamen- tal rethinking of the way computers operate. Neuromor- phic research, however, is progressing: firms like Intel are leading the charge in chip development, and the success of their e4orts will have large-scale implications on consum- ers, the semiconductor industry, and technology firms alike. It is vital, therefore, that business leaders consider the possibilities and threats of this novel technology, un- derstanding that the future computing paradigm will be decided by innovative firms who are rethinking current processes. !'$'&"!- 1. The term neuromorphic means “taking the form of the brain” (Fulton III). Intel (a leader in neuromorphic research) summa- rizes neuromorphic technology as “chips that function less like traditional computers and more like the human brain” (“Intel Newsroom–Neuromorphic Computing”). 2. It can be argued that the original notion of neuromorphic computing was first proposed in Alan Turing’s 1948 paper, Intel- ligent Machinery. However, the concept is typically attributed to Carver Mead’s 1989 Analog VLSI and Neural Systems, a paper in which Mead argued that chips with an increasingly dense col- lection of transistors could best communicate via a replication of the brain’s neural wiring. 3. The downside of traditional computer architecture is typically referred to as the von Neumann bottleneck—the notion that despite increases in processing speed, processors are forced to remain idle while data is transferred to and from the memory of a chip, resulting in increased latency. 4. Neuromorphic chips employ analog circuitry to transfer elec- trical signals between “neurons.” The idea is that the system will be able to modulate the amount of electricity flowing between nodes, mimicking the fact that brain signals naturally have vary- ing degrees of strength. 5. A key emphasis of neuromorphic computing is the selective mapping of neural connections. In order to operate at maximum e3ciency, the brain uses only the specific neurons and synapses necessary to perform a given task. Neuromorphic systems, therefore, seek to emulate this e3ciency by strategically and selectively forming connections between neighboring neurons. 6. Each “spike” in an SNN is a single-bit impulse that is analo- gous to an action potential in a naturally occurring neuron. A given node in the network is capable of spiking only if a state variable exceeds a given threshold. 7. In fully-connected ANNs, all nodes compute and send their output to the next layer at each time step, even if nothing has sig- nificantly changed, making the overall network computationally expensive. Conversely, SNNs employ neurons that process input separately and are only connected to local neighbors, implying that the entire layer does not need to be calculated for informa- tion to proceed to the next layer. 8. In 1965 Intel’s co-founder Gordon Moore predicted that the number of transistors that could be placed on an integrated circuit would double every two years while costs remain constant / decline. 9. Loihi comprises 128 neuromorphic cores, 131,000 “neurons”, and 130 million “synapses” (connections). The chip is named after an active submarine volcano o4 the coast of Hawaii that is set to emerge one day. The idea is that neuromorphic computing is analogously “emerging” and will eventually break the “surface” of the current computing environment. 10. In December 2018 Intel announced the creation of Kapoho Bay, Intel’s smallest neuromorphic system consisting of two Loihi chips. In July of 2019 Intel was able to scale Loihi into their 64-chip Pohoiki Beach, representing the neural capacity of 8 mil- lion neurons. Most recently, Intel debuted its 768-chip Pohoiki Springs with a collective neural capacity of 100 million neurons. The chips reside in a chassis the size of five standard servers and is provided to members of the Intel Neuromorphic Research Community via a cloud-based system. Intel hopes it will emerge as a “tool for researchers to develop and characterize new neuro- inspired algorithms for real-time processing, problem solving, adaptation and learning” (“Intel Newsroom–Pohoiki Springs”). 11. In July of 2020 the National University of Singapore ran an event-driven, visual-tactile perception test on Intel’s Loihi chip 37 !"#$%&%$'()* *%&'#+)!, as well as various GPU systems to compare power consump- tion (Russel). The Loihi test chip slightly outperformed the GPU systems (in inferences per second) at significantly lower power consumption. It is important to note that Loihi’s strong performance is largely the result of the entire system (in this case, robotic sensors, data formats, algorithms, and the Loihi architecture) being “re-engi- neered in an event-based paradigm” (Mike Davies, Director of Neuromorphic Research at Intel. See Russel). Put simply, when the system is consistent with Loihi’s event-based architecture, neuromorphic hardware outperforms traditional GPU’s when it comes to energy e3ciency. 12. Existing deep learning algorithms depend on stochastic gradient descent and error backpropagation to e3ciently “learn” a given ANN. Since SNNs operate in discontinuous, non-di4er- entiable spikes, it is impossible to apply existing learning rules directly to SNN training (Pfei4er and Pfeil). 13. At the 2019 International Solid State Circuits Conference in San Francisco, LeCun heavily criticized neuromorphic comput- ing in his opening keynote address. While Intel’s Mike Davies agrees that there is a lot of progress to be made on the algorith- mic front, which he claims is “holding back the field,” Yann’s criticism resulted in a fireback from Davies citing the e3cacy of neuromorphic chips in a December report produced by Applied Brain Research of Waterloo, Ontario. Davies’s fireback was met with a detailed Facebook post by LeCun outlining the issues he sees with neuromorphic hardware, specifically citing the lack of an e3cient training algorithm. 14. It is important to point out that SNNs are not optimized for performance on existing AI Benchmarks (e.g., ImageNet). Just how the brain is not optimized (but capable of) classifying an image that is quickly flashed on the retina, SNNs struggle with the typical frame-based test of software accuracy. The evaluation of SNNs, therefore, requires benchmark testing that emphasizes the functionality of spiking networks—i.e., “making decisions based on continuous input streams while moving in the real world” (Pfei4er and Pfeil). 15. Dan Hutcheson, CEO of VLSI Research (an independent market analysis and consulting firm that tracks the semiconduc- tor industry), describes the false dichotomy as follows: “Today’s computers are very good at what they do. They will continue to outperform neuromorphic computing systems for conventional processing tasks. The technologies are complementary and so they will coexist.” Adam Stieg, associate director of the California NanoSystems Institute at the University of California at Los An- geles, further describes how “conventional von Neumann-based computing systems” perform very well with “high-speed, pre- dictable, deterministic processes,” but struggle with increasing complexity. The promise of neuromorphic computing, therefore, is the opening up of “an entirely new and unexplored area of computing”—one that allows us to “do things with computers that we couldn’t have imagined in the past” (See Greengard). 16. High-performance computing (HPC) can be generalized as a process of optimizing chip architecture for existing deep learn- ing algorithms (Vorhies). The majority of attention is currently placed on HPC, with large technology firms beginning to enter the chip manufacturing field by developing their own propri- etary chips (e.g., Google’s Tensor Processing Unit) (Tiernan, “Facebook’s Yann LeCun”). Conversely, quantum computing is analogous to neuromorphic computing in that it represents a rethinking of existing computer architecture. Rather than operat- ing in bits (0 or 1), quantum computing employs quantum bits (qubits) that can be set to 0, 1, or both simultaneously (Cho). +!(!+!'5!- Best, Jo. “Neuromorphic Computing Could Solve the Tech Industry’s Looming Crisis.” ZDNet, 5 Oct. 2020, ` www.zdnet.com/article/neuromorphic-computing-could- solve-the-tech-industrys-looming-crisis/. Blain, Loz. “Intel’s Neuromorphic Loihi Chip Is Rapidly Learning to Discern Smells.” New Atlas, 17 Mar. 2020, ne- watlas.com/science/intel-neuromorphic-loihi-smells/. Cho, Adrian. “The Biggest Flipping Challenge in Quan- tum Computing.” Science, The American Association For The Advancement of Science, 9 July 2020, www.sci- encemag.org/news/2020/07/biggest-flipping-challenge- quantum-computing. Davies, Mike, et al. “2018 CES: Neuromorphic Computing Mimics the Human Brain.” Youtube, Intel Newsroom, 8 Jan. 2018, www.youtube.com/ watch?v=gXNCz26UhyY&feature=emb_title. Ferry, Je4. “The Future of AI Is Neuromorphic.” Circuit Insight, www.circuitinsight.com/programs/55035.html. Fulton III, Scott. “What Neuromorphic Engineering Is, and Why It’s Triggered an Analog Revolution.” ZDNet, 8 Feb. 2019, www.zdnet.com/article/what-neuromorphic- engineering-is-and-why-its-triggered-an-analog-revolu- tion/. 38 !"!#!$%& : : &'()$* 2021 Greengard, Samuel. “Neuromorphic Chips Take Shape.” Communications of the ACM, 1 Aug. 2020, cacm.acm. org/magazines/2020/8/246356-neuromorphic-chips- take-shape/fulltext. Hamblen, Matt. “Intel Shares Pohoiki Springs Chassis with AI Researchers.” FierceElectronics, 18 Mar. 2020, www.fierceelectronics.com/electronics/intel-shares-pohoi- ki-springs-chassis-ai-researchers. Irving, Michael. “Artificial Synapses Fill the Gaps for Brainier Computer Chips.” New Atlas, 23 Jan. 2018, newat- l a s . c o m / n e u r o m o r p h i c - c h i p - m i t / 5 3 0 8 1 / ? i t m _ source=newatlas. Jhonsa, Eric. “Intel Teams with Uncle Sam to Research Chip Packages, ‘Neuromorphic’ Systems.” TheStreet, 2 Oct. 2020, www.thestreet.com/investing/intel-teams- with-uncle-sam-to-research-chip-packages-neuromorphic- systems-loihi-foveros-emib. LeCun, Yann. ISSC Keynote Address. Facebook, 23 Feb. 2019, https://m.facebook.com/story.php?story_fbid=1015 5808247852143&id=722677142. Molter, Tim. “A Brief History of Neuromorphic Comput- ing.” Knowm.org, 4 Apr. 2015, knowm. org/a-brief-history-of-neuromorphic-computing/. Muir, Dylan. “Putting the Neural Back into Networks.” Me- dium, 15 Jan. 2020, medium.com/neuromorphic-intelli- gence/putting-the-neural-back-into-networks-f0fd- b19a5efd. Norman, Jeremy M. “Carver Mead Writes the First Book on Neuromorphic Computing.” Carver Mead Writes the First Book on Neuromorphic Computing : History of In- formation, 25 Oct. 2020, www.historyofinformation.com/ detail.php?id=3900. Panetta, Kasey. “Gartner Top 10 Strategic Predictions for 2021 and Beyond.” Smarter With Gartner, 21 Oct. 2020, www.gartner.com/smarterwithgartner/gartner-top- 10-strategic-predictions-for-2021-and-beyond/. Pfei4er, Michael, and Thomas Pfeil. “Deep Learning With Spiking Neurons: Opportunities and Challenges.” Fron- tiers in neuroscience vol. 12 774. 25 Oct. 2018, doi:10.3389/ fnins.2018.00774. Ray, Tiernan. “Facebook’s Yann LeCun Says ‘Internal Ac- tivity’ Proceeds on AI Chips.” ZDNet, 18 Feb. 2019, www. zdnet.com/article/facebooks-yann-lecun-says-internal-ac- tivity-proceeds-on-ai-chips/. Ray, Tiernan. “Intel’s Neuro Guru Slams Deep Learning: ‘It’s Not Actually Learning’.” ZDNet, 23 Feb. 2019, www. zdnet.com/article/intels-neuro-guru-slams-deep-learning- its-not-actually-learning/. Ray, Tiernan. “Neuromorphic Computing Finds New Life in Machine Learning.” ZDNet, 1 July 2019, www.zdnet. com/article/neuromorphic-computing-finds-new-life-in- machine-learning/. Rouse, Margaret. “What Is Neuromorphic Computing?” SearchEnterpriseAI, TechTarget, 11 Feb. 2020, searchen- terpriseai.techtarget.com/definition/neuromorphic-com- puting. Russell, John. “Get a Grip: Intel Neuromorphic Chip Used to Give Robotics Arm a Sense of Touch.” HPCwire, 17 July 2020, www.hpcwire.com/2020/07/15/get-a-grip-intel- neuromorphic-chip-used-to-give-robotics-arm-a-sense-of- touch/. Soni, Devin. “Spiking Neural Networks, the Next Genera- tion of Machine Learning.” Medium, Towards Data Sci- ence, 16 July 2019, https://towardsdatascience.com/spik- ing-neural-networks-the-next-generation-of-machine- learning-84e167f4eb2b Vorhies, William. “The Three Way Race to the Future of AI. Quantum vs. Neuromorphic vs. High Performance Computing.” Data Science Central, Tech Target, 14 Nov. 2017, www.datasciencecentral.com/profiles/blogs/the- three-way-race-to-the-future-of-ai-quantum-vs-neuromor- phic-vs. Wiggers, Kyle. “Intel Debuts Pohoiki Springs, a Powerful Neuromorphic Research System for AI Workloads.” Ven- tureBeat, 18 Mar. 2020, venturebeat.com/2020/03/18/in- tel-debuts-pohoiki-springs-a-powerful-neuromorphic-re- search-system-for-ai-workloads/. “Neuromorphic Computing - Next Generation of AI.” In- tel–Neuromorphic Computing, Intel, 2020, www.intel. com/content/www/us/en/research/neuromorphic-com- puting.html. 39 !"#$%&%$'()* *%&'#+)!, “Intel and Sandia National Labs Collaborate on Neuromor- phic Computing.” Intel Newsroom, Intel, 2 Oct. 2020, newsroom.intel.com/news/intel-sandia-national-labs- neuromorphic-computing/. “Intel Scales Neuromorphic Research System to 100 Mil- lion Neurons.” Intel Newsroom, Intel, 18 Mar. 2020, newsroom.intel.com/news/intel-scales-neuromorphic-re- search-system-100-million-neurons/. “Intel’s Pohoiki Beach, a 64-Chip Neuromorphic System, Delivers Breakthrough Results in Research Tests.” Intel Newsroom, Intel, 15 July 2019, newsroom.intel.com/ news/intels-pohoiki-beach-64-chip-neuromorphic-sys- tem-delivers-breakthrough-results-research-tests/. “Loihi - Intel.” WikiChip, 15 July 2019, en.wikichip.org/ wiki/intel/loihi. 40 !"!#!$%& : : &'()$* 2021 */-" &( 6+"7&+8 !"# $%!!&# '( !"# )'**# +'*,+: !"# )!'-. '( /,&&,%* *+(%01#%2 © Creative Centenaries. (n.d.). The Battle Of The Somme comic: The story of William McFadzean | CCEA. CEA. Retrieved January 19, 2022, from https://ccea.org.uk/learning-resources/understanding-1916/creative-centena- ries/battle-somme-comic-story-william 13 15 28 28 28 28 29 29 (,34-# 1 © originally tweeted by User:@realDonaldTrump (account suspended) on 26 May, 2020. #%)!#- 5-'+&%*%!,'2 '( 1916 © originally uploaded to the English Wikipedia by w:User:Jtdirl (https://commons.wikimedia.org/wiki/ File:Easter_Proclamation_of_1916.png), „Easter Proclamation of 1916“, marked as public domain, more details on Wikimedia Commons: https://commons.wikimedia.org/wiki/Template:PD-old (,34-# 2 © GDI report: Ad-funded COVID-19 Disinformation: Money, Brands, and Tech (https://disinformationindex. org/wp-content/uploads/2020/07/GDI_Ad-funded-COVID-19-Disinformation-1.pdf), „Share of ad revenues generated by COVID Disinfo Sites“. (,34-# 3, 5%2#& % © originally published in Bursztyn, Rao, Roth, Yanagaziwa-Drott: Misinformation During a Pandemic (https:// www.nber.org/papers/w27417), „RF estimates of e4ect of di4erential viewership on cases and deaths (extend- ed)“. (,34-# 3, 5%2#& $ © originally published in Bursztyn, Rao, Roth, Yanagaziwa-Drott: Misinformation During a Pandemic (https:// www.nber.org/papers/w27417), „2SLS estimates of e4ect of di4erential viewership on cases and deaths (ex- tended)“. (,34-# 4, 5%2#& % © originally published in Bursztyn, Rao, Roth, Yanagaziwa-Drott: Misinformation During a Pandemic (https:// www.nber.org/papers/w27417), „Implied COVID-19 Curves, Panel A: Estimates on Cases“. (,34-# 4, 5%2#& $ © originally published in Bursztyn, Rao, Roth, Yanagaziwa-Drott: Misinformation During a Pandemic (https:// www.nber.org/papers/w27417), „Implied COVID-19 Curves, Panel B: Estimates on Deaths“. 35 38 43 45 %-!,(,+,%& ).2%5)#) $%)#0 '2 (#--'-#&#+!-,+ !422#& 642+!,'2) © Mike Davies Intel Labs, Intel Corporation et. al: (https://commons.wikimedia.org/wiki/File:Artificial_syn- apses_based_on_FTJs.png), https://creativecommons.org/licenses/by-sa/4.0/legalcode !%$&# ,,, ,2(#-#2+# 5'/#- %20 5'/#- 4!,&,1%!,'2 © Tasbolat Taunyazov, Weicong Sng, Hian Hian See, Brian Lim, Jethro Kuan, Abdul Fatir Ansari, Benjamin C.K. Tee, and Harold Soh: (http://www.roboticsproceedings.org/rss16/p020.pdf), „Table III: Inference Speed and Power Utilization“ $&%+"# &%*'2! %! "#+&%, *'2!%2%, /,!" "#- )!40#2!), 1893 © Unknown authorUnknown author (https://commons.wikimedia.org/wiki/File:Blanche_Lamont_with_stu- dents.jpg), „Blanche Lamont with students“, marked as public domain, more details on Wikimedia Commons: https://commons.wikimedia.org/wiki/Template:PD-US !"#'0'-# 04--%2!’) *43)"'! © originally published in Report of the trial of William Henry Theodore Durrant (https://archive.org/details/ reportoftrialofw00durr/page/10/mode/2up), marked as public domain. 60 61 61 61 3-%5" ,&&4)!-%!,23 !"# )",(!) ,2 5-,+# %20 74%2!,!. )455&. © originally published in 31.11 E3ciency and Deadweight Loss (https://saylordotorg.github.io/text_economics- theory-through-applications/s35-11-e3ciency-and-deadweight-loss.html), „Figure 31.8 Deadweight Loss“ 3-%5") )"'/,23 !"# 24*$#- %20 8%&4# '( *#-3#-) %20 %+74,),!,'2) © originally published in Statista: Mergers and acquisitions in the United States (https://www.statista.com/ study/65232/mergers-and-acquisitions-in-the-united-states). 3-%5") )"'/,23 3&'$%& 74#-. )"%-# $. )##&+! )#%-+" #23,2#) © originally published in Statista: Global market share of search engines 2010-2022 (https://www.statista.com/ statistics/216573/worldwide-market-share-of-search-engines). 3-%5") )"'/,23 0'*#)!,+ 4) )#%-+" 74#-. )"%-# $. )#&#+! )#%-+" #23,2#) © originally published in Statista: Market share of search engines in the United States 2008-2022 (https:// www.statista.com/statistics/267161/market-share-of-search-engines-in-the-united-states). */-" &( 6+"7&+8 +'-5'-%!# !#+"2'&'3. ,2('-*%!,'2 )#-8,+#) © unpublished tabulation from Corptech Database 62 62 71 75 78 80 82 (,34-# 1 © Samadi, Moradi, Moradinazar, Mostafai, Pasdar: Dietary pattern in relation to the risk of Alzheimer’s disease: a systematic review (https://link.springer.com/article/10.1007/s10072-019-03976-3), „Figure 2: Healthy and unhealthy diet in relation to AD.“ !"# $-#%90'/2 '( '5,2,'2 -#3%-0,23 $-#%9,23 45 ‘$,3 !#+"’ © Statista : Majority of Americans in Favor of Breaking up Big Tech (https://www.statista.com/chart/19440/ survey-responses-breaking-up-big-tech) 5'!#2!,%& ),32%!'-,#), ),32%!'-,#), %20 5%-!,#) !' !"# %,,$ © L.tak (https://commons.wikimedia.org/wiki/File:Asian_Infrastructure_Investment_Bank_map.svg), https://creativecommons.org/licenses/by-sa/4.0/legalcode +",2%’) '23',23 !#--,!'-,%& 0,)54!#)] © O3ce of the Secretary of Defense (https://commons.wikimedia.org/wiki/File:China’s_Disputed_Territories. PNG), „China’s Disputed Territories“, marked as public domain, more details on Wikimedia Commons: https:// commons.wikimedia.org/wiki/Template:PD-US 4) 0#5%-!*#2! '( 64)!,+# %*'23 '!"#-) %22'42+#0 23 +-,*,2%& +"%-3#) © U.S. Department of Justice (https://commons.wikimedia.org/wiki/File:U.S._Commerce_Secretary_Wilbur_ Ross,_Acting_Attorney_General_Matthew_Whitaker,_Homeland_Security_Secretary_Kirstjen_Nielsen,_FBI_ Director_Christopher_Wray_Announces_23_Criminal_Charges_Against_China’s_Huawei_&_Wanzhou_ Meng.jpg), „U.S. Commerce Secretary Wilbur Ross, Acting Attorney General Matthew Whitaker, Homeland Security Secretary Kirstjen Nielsen, FBI Director Christopher Wray Announces 23 Criminal Charges Against China’s Huawei & Wanzhou Meng“, marked as public domain, more details on Wikimedia Commons: https:// commons.wikimedia.org/wiki/Template:PD-US %*#-,+% 3&'$%& !-%0# 0#(,+,! © Statista: U.S. Trade Deficit in Goods Reaches Record High, (https://www.statista.com/chart/17281/us-trade- balance). 98 +"%-&#) +'43"&,2, &#%0#- '( !"# +"-,)!,%2 (-'2! © Craine, Detroit (https://commons.wikimedia.org/wiki/File:CharlesCouglinCraineDetroitPortrait.jpg), „CharlesCouglinCraineDetroitPortrait“, marked as public domain, more details on Wikimedia Commons: https://commons.wikimedia.org/wiki/Template:PD-US 100 3#-*%2 %*#-,+%2 $420 -%&&., 2#/ .'-9, *%0,)'2 )74%-# 3%-#2, (#$ 1939 © Department of Defense. Department of the Army. O3ce of the Chief Signal O3cer. (09/18/1947 - 02/28/1964); ARC Identifier 36068 / Local Identifier 111-OF-2 1943 (https://commons.wikimedia.org/wiki/ File:Nazi_Strike_2.jpg), „Nazi Strike 2“, marked as public domain, more details on Wikimedia Commons: https://commons.wikimedia.org/wiki/Template:PD-US