Computational Thinking - Учебное пособие (1176923), страница 11
Текст из файла (страница 11)
Is it possible to overcome the challenge of smarter-thanhuman intelligence?8. Does the future need us?5. Translate from Russian into English:В 1950г. британский математик и криптограф АланТьюринг выпустил статью «Вычислительные машины иразум», заложив краеугольный камень в отрасль, котораязанимается разработкой искусственного интеллекта.Тьюринг предложил решить проблему практически, а нетеоретически, - с помощью теста.58Если человек в ходе переписки с программой не можетопределить, человек его собеседник или машина, тоследует признать эту программу разумной, что бы это низначило.В этом году в конкурсе должны были участвовать шестьпрограмм: Alice, Brother Jerome, Elbot, Eugene Goostman,Jabberwacky и Ultra Hal. Самой умной была признанапрограмма Elbot – ей удалось одурачить трех из 12 судей,то есть 25%.Тест Тьюринга проходят программы, которые имитируютподдержание беседы, а не ведут ее.
В разговорах с нимибесполезно искать новую для себя информацию. Вбольшинстве случаев они пытаются свести конкретныевопросы к общим ответам. Такой была одна из первыхпрограмм, попытавшихся пройти тест Тьюринга, ELIZA.Она реагировала на ключевые слова и высказывала своемнение по данному вопросу. Ее суждения не выходилиза рамки, жестко заданные программистом.С тех пор искусство имитации продвинулось далековперед. Одна из программ, например, в разговорепостоянно упоминала тетю Соню из Одессы, другая —намеренно шутила о себе как о роботе, чтобы сбитьлюдей с толку. Сам автор Elbot не считает, что егопрограмма хоть как-то может мыслить. Он сравнивает еебеседы с фокусами, у которых есть секрет, но не более.6.
Develop the following ideas in writing an essay (120 – 150words):1. The greatest danger of AI is that people conclude too earlythat they undertstand it.2. One of the most critical points about AI is that it mightincrease in intelligence extremely fast.3. Our alternatives boil down to becoming smarter orbecoming extinct.59Nanoscience: Facts vs. FictionWords and Phrases:Dissipation– рассеиваниеRobust– надёжный, устойчивый к ошибкамResilience- устойчивость(к внешнимвоздействиям)Intrinsically– по сути, в сущностиSpeculation– предположение, догадка,Constraint– ограничениеTo shrink– уменьшаться, сокращатьсяA fabric– строение, структураBandwidth– ширина спектра; диапазонрабочих частотHeterogeneous– неоднородныйTo be embedded– быть внедреннымPerformance gap– разница в исполненииMerging- слияние, объединениеIn the interim– пока что, тем временемSoft-error– программная ошибкаOverhead– издержки, затратыSynergy– взаимное усилениеError-prone– склонный к ошибкамTo garble- искажатьPlausibility– правдоподобиеIntestine- кишечникFeasible- правдоподобныйThere has been a great deal of public and scientific interest inthe so-called nanotechnology revolution.
Nanotechnology canbe defined as the manipulation, precision placement,measurement, modeling, or manufacture of sub-100nanometer (nm) scale matter (lnm equals 1/1,000,000,000*of m). This manipulation of matter at the nano level willgreatly influence most areas of our life, such as manufacturing,engineering, health, pharmaceuticals, and (of special interesthere) information technology. There is a wide range of60nanoparticles of different types and different propertiescurrently in production that may be applied in a variety ofways. It is envisaged that nanoparticle types found useful willbe further developed by large-scale production.Architectures for Silicon Nanoelectronics and BeyondThe semiconductor industry faces serious problems with powerdensity, interconnect scaling, defects and variability,performance and density overkill, design complexity, andmemory-bandwidth limitations.
A candidate to replacecomplementarymetal-oxidesemiconductor(CMOS)technology, nanoelectronics could address some of thesechallenges, but it also introduces new problems. Molecularscale computing will likely allow additional orders-ofmagnitude improvements in device density and complexity.The effective use of nanotechnology will require not justsolutions to increased density, but total system solutions. Wecan't develop an architecture without a sense of the applicationsit will execute.
And any paradigm shift in applications andarchitecture will have a profound effect on the design processand tools required.We define nanoarcbitecture as the organization of basiccomputational structures composed of nanoscale devicesassembled into a system that computes something useful.Nanoarchitecture will enable radically different computationalmodels, and, due to its potential for large capacity, might alsoprovide superior capabilities in some areas.There are two paths to follow: evolutionary and revolutionary.Evolutionary path. Silicon semiconductor technology willcontinue to shrink.
But there's an increasing performance gapbetween device technology and its ability to deliver performancein proportion to device density. Performance, in terms ofmillions of instructions per second per watt, isn't keeping upwith the increase in millions of devices per chip. There's also agap between device density and our ability to design new chipsthat use every device on the chip and guarantee they'redesigned correctly.
Power consumption and heat dissipationpresent additional challenges.61Revolutionary path. Knowing that the end of Moore's lawscaling is in sight for traditional silicon technology, many haveembarkedonrevolutionarynanoelectronicsresearch.Researchers are studying carbon nanotube transistors, carbonnanotube memory devices, molecular electronics, spintronics,quantum-computing devices, magnetic memory devices, andoptoelectronics.Unfortunately, we won't use many of these devices until it'sabsolutely necessary to consider a replacement technology. So,how should we use these revolutionary nanoelectronic devicesin the interim, especially when these devices haven'tdemonstrated sufficient reliability and large enough signal-tonoise ratio to guarantee reliable digital computation?In addition to massive CMOS-scaling efforts, many researchersare pursuing molecular, optical, or quantum devices that theycould integrate with CMOS-based digital logic to producehybrid systems.Hybrid architectures will allow integration of sensing andprocessing functions in ways analogous to biologicalsystems.
Living beings can perform complex real-timefunctionswithremarkableease,unmatchedinperformance by the most powerful man-made computers.Inspired by biology, cellular sensor-processor architecturesappear promising for hybrid nanodevices.The availability of very dense conventional silicontechnology, along with non-conventional, nanoscale storageor memory technology based on phase-changing materials,also makes fascinating hybrid architectures possible.Integration of logic and memory will allow large-scale arraycomputing with increased local storage.
This is, incidentally,an important operating principle of neural circuits. Relevantwork in this area includes the Intelligent RAM project atthe University of California, Berkeley.In addition to logic and memory, hybrid architectures mightallow for integration of sensing with logic and memory. Oneimportant class of applications would be vision systems, inwhich each photo detector would be embedded with its own62circuitry in a cellular architecture. Each element in such anarchitecture would resemble a "neuron" in a retina-like array,where the system performs basic image-processing functionson the incoming image flow.Similarly, other biologically inspired applications includespeech or auditory processing.
The human auditory system canrecognize speech even if it's garbled, embedded in noise, ormixed with other voices. Here, in addition to dense memoriesand massively parallel processing, nanoelectronics offers thepotential for combining the sensing of sound and theprocessing of speech into a single computationalnanoarchitecture.Risks of NanotechnologyWith all the excitement over the potential of nanotechnology,has any research addressed the potential toxicity nanoparticlesmay have on our health and environment?Nanofiction. In his novel, Prey, Michael Crichton writes of aswarm of nanobots released in the Nevada desert, terrorizingthe very scientists who developed them.
Unfortunately, thisdeepening vision of "grey goo" coating our landscape as thenanobots replicate uncontrollably is the picture some peoplehave of the risk from nanotechnology. Such a picture, thoughtitillating, is a scientific fantasy, more in the tradition of KingKong than the realms of scientific plausibility. The actual risksfrom nanotechnology, as seen by those of us working in thisfield, are less dramatic, but nonetheless, are potentially real.Nanofacts.The greatest risk to human health and to the environmentlies in this rapid expansion of different types of nanoparticlesunder development and the potential for their production.
Itbrings with it the possibility of large-scale human exposurepredominantly in the work place as these nanoparticles areincorporated into products of every conceivable type from ITto food. Additionally, the particles can be released during wearand destruction throughout the life of a product. Suchdiversity in use (with numerous potential scenarios forexposure) means that nanoparticles are likely to make contact63with the body via the lungs, intestines, and skin.The lungs are an obvious and critical route of entry to thebody, especially during the manufacturing process where dustclouds can be generated. The intestines provide a route ofentry for nanoparticles contained both within processed foodsand in mucus cleared from the lungs.
Entry of nanoparticlesthrough the skin via cosmetics is also feasible. Recent evidencesuggests that nanoparticles landing in the nose can moveupward into the base of the brain. The effect this has on thebrain and nervous system is under investigation.Medical Uses of NanoparticlesThe very properties that make nanoparticles useful for newapplications are also the very properties that can increase theirharmfulness.For example, it is possible to alter the surface of ananoparticle to direct it toward a specific organ such as theliver or brain, thus allowing drugs to be targeted to a specificspot rather than working through the entire body.
This is notonly more efficient for the patient, but it results in fewer sideeffects. It suggests the way in which nanoparticles travelaround the body and ultimately settle. And where do theparticles go after they have fulfilled their delivery function?What effect do they have on the various organs they target?Nanotubes are like extended buckyballs (a molecule of carbon60); that is, they are very thin (a few tens of nm), but can bevery long (in the order of mm). They are extremely strong,able to maintain their structure and not break down in thelung. Nanotubes are similar to fibers like asbestos and glass,and as such could be very harmful to the lungs causing scarringand cancer.
Therefore, adequate toxicological studies ofnanotubes must be conducted to fully understand thepotential risks.The scientists have recently suggested we must recognize a newscience of Nanotoxicology to specifically examine the productsof the nanotechnology revolution for their likely adverseeffects. We are not suggesting that further advancement ofnanotechnology, nor the development of nanoparticles, should64be halted; in fact, this field promises great benefit, especiallyfor the IT and communications industries.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
