Modern information technology (it) relies on division of labour. Photons carry data around the world and electrons process them. But, before optical fibres, electrons did both—and some people hope to complete the transition by having photons process data as well as carrying them.

現(xiàn)代信息技術(shù)依賴勞動分工。光子傳輸全球的數(shù)據(jù)，電子處理數(shù)據(jù)。但在光纖問世之前，兩項任務(wù)都由電子完成。有人希望實現(xiàn)徹底的轉(zhuǎn)變，使數(shù)據(jù)的傳輸和處理都由光子來完成。

Unlike electrons, photons (which are electrically neutral) can cross each others’ paths without interacting, so glass fibres can handle many simultaneous signals in a way that copper wires cannot. An optical computer could likewise do lots of calculations at the same time. Using photons reduces power consumption, too. Electrical resistance generates heat, which wastes energy. The passage of photons through transparent media is resistance-free.

不同于電子，光子（具有電中性）可以在不相互作用的情況下穿過彼此的路徑，所以玻璃纖維能處理許多同步信號，而銅線做不到。同樣，光學(xué)計算機能同時進行大量的計算。利用光子還能減少功耗，電阻發(fā)熱會有能量損失，而光子在透明介質(zhì)中傳輸是沒有電阻的。

For optical computing to happen, though, the well-established architecture of digital electronic processing would have to be replaced by equivalent optical components. Or maybe not. For some people are working on a novel optical architecture that uses analogue rather than digital computing (that is, it encodes data as a continuous signal rather than as discrete “bits”). At the moment, this architecture is best suited to solving one particular class of problems, those of a branch of maths called linear algebra. But that is a potentially huge market, for linear algebra is fundamental to, among other matters, artificial neural networks, and they, in turn, are fundamental to machine learning—and thus artificial intelligence (ai).

但要想實現(xiàn)光計算，成熟的數(shù)字電子運算架構(gòu)必須由等效的光學(xué)器件來代替。也許沒這個必要，因為有人正在研發(fā)一種新型光學(xué)架構(gòu)，它使用模擬而非數(shù)字計算（將數(shù)據(jù)編碼成連續(xù)的信號而不是離散的“比特”）。目前，這種架構(gòu)最適合解決一類問題：稱為線性代數(shù)的數(shù)學(xué)分支。但這是潛在的巨大市場，因為線性代數(shù)是人工神經(jīng)網(wǎng)絡(luò)的基礎(chǔ)，而人工神經(jīng)網(wǎng)絡(luò)是機器學(xué)習(xí)——進而成為人工智能的基礎(chǔ)。

The power of the matrix

矩陣的力量

Linear algebra manipulates matrices. These are grids of numbers (representing coefficients of simultaneous equations) that can be added and multiplied a bit like individual numbers. Among the things which can be described by matrices are the equations governing the behaviour of electromagnetic radiation (such as light) that were discovered in the 19th century by James Clerk Maxwell. Light’s underlying Maxwellian nature makes it easy, using appropriate modulating devices, to encode matrix data into light beams and then manipulate those data.

線性代數(shù)計算矩陣。矩陣是數(shù)字網(wǎng)格（代表聯(lián)立方程式的系數(shù)）有點類似于相加和相乘的單個數(shù)字。矩陣可用來描述的事物包括19世紀(jì)詹姆斯·克拉克·麥克斯韋發(fā)現(xiàn)的控制電磁輻射行為（例如光）的方程組。光具有潛在的麥克斯韋特性，所以利用適當(dāng)?shù)恼{(diào)制設(shè)備易于將矩陣數(shù)據(jù)編碼成光束，然后運算這些數(shù)據(jù)。

Artificial neural networks are programs that represent layers of nodes, the connections between which represent numbers in matrices. The values of these change in response to incoming signals in a way that results in matrix multiplication. The results are passed on to the next layer for another round of processing, and so on, until they arrive at a final output layer, which synthesises them into an answer. The upshot is to allow a network to recognise and learn about patterns in the input data.

人工神經(jīng)網(wǎng)絡(luò)是表示多層節(jié)點的程序，節(jié)點之間的連接表示矩陣數(shù)據(jù)，它們的數(shù)值隨著輸入信號發(fā)生變化，結(jié)果導(dǎo)致矩陣相乘。計算結(jié)果傳遞給下一層，進行新一輪運算，依此類推，直到傳遞至最終輸出層，將這些計算結(jié)果合成一個答案，最終使神經(jīng)網(wǎng)絡(luò)能夠識別和學(xué)習(xí)輸入數(shù)據(jù)中的模式。

The idea of turning neural networks optical is not new. It goes back to the 1990s. But only now has the technology to make it commercially viable come into existence. One of the people who has observed this transition is Demetri Psaltis, an electrical engineer then at the California Institute of Technology (Caltech) and now at the Swiss Federal Institute of Technology in Lausanne. He was among the first to use optical neural networks for face recognition.

神經(jīng)網(wǎng)絡(luò)向光學(xué)轉(zhuǎn)變不是新理念。它可以追溯到20世紀(jì)90年代，但直到現(xiàn)在才出現(xiàn)使它具有商業(yè)可行性的技術(shù)。杰梅特里·普薩爾蒂斯目睹了轉(zhuǎn)變過程，他當(dāng)時在加州理工學(xué)院擔(dān)任電氣工程師，現(xiàn)在任職于洛桑聯(lián)邦理工學(xué)院，他率先將光神經(jīng)網(wǎng)絡(luò)應(yīng)用于人臉識別。
原創(chuàng)翻譯：龍騰網(wǎng) http://www.top-shui.cn 轉(zhuǎn)載請注明出處


The neural networks of Dr Psaltis’s youth were shallow. They had but one or two layers and a few thousand nodes. These days, so-called deep-learning networks can have more than 100 layers and billions of nodes. Meanwhile, investments by the telecoms industry—the part of it that ships data around through all those optical fibres—have made it possible to fabricate and control optical systems far more complex than those of the past.

神經(jīng)網(wǎng)絡(luò)在普薩爾蒂斯的青年時期比較簡單，只有一兩層和幾千個節(jié)點，現(xiàn)在所謂的深度學(xué)習(xí)網(wǎng)絡(luò)擁有100多層和幾十億個節(jié)點。同時，電信行業(yè)的投資——利用光纖傳輸數(shù)據(jù)的那部分業(yè)務(wù)——使人們得以制造和控制遠(yuǎn)比過去更復(fù)雜的光學(xué)設(shè)備。

That is the technological push. The financial pull derives from shedding the cost of the vast amount of electricity consumed by modern networks as they and the quantities of data they handle get bigger and bigger.

這是技術(shù)上的動力。隨著現(xiàn)代網(wǎng)絡(luò)及其處理的數(shù)據(jù)量越來越大，耗電量十分巨大，而經(jīng)濟上的吸引力是降低電力的成本。

Most efforts to build optical neural networks have not abandoned electrons entirely—they pragmatically retain electronics where appropriate. For example, Lightmatter and Lightelligence, two firms in Boston, Massachusetts, are building hybrid “modulators” that multiply matrices together by manipulating an optically encoded signal according to numbers fed back electronically. This gains the benefit of parallelism for the optical input (which can be 100 times what electronics would permit) while using more conventional kit as what Nicholas Harris, Lightmatter’s founder, describes as the puppet master.

建造光神經(jīng)網(wǎng)絡(luò)的大多數(shù)行動沒有完全拋棄電子——他們務(wù)實地在適當(dāng)?shù)奈恢帽Ａ綦娮釉＠?，位于美國馬薩諸塞州波士頓的兩家公司Lightmatter和Lightelligence正在制造混合型“調(diào)制器”，它們根據(jù)電子反饋的數(shù)據(jù)來處理光學(xué)編碼信號，從而實現(xiàn)矩陣相乘。好處是使用比較常規(guī)的配套元件就能達到光纖輸入量（是電子元件輸入量的100倍），Lightmatter公司創(chuàng)始人尼古拉斯·哈里斯稱其為傀儡主人 。
原創(chuàng)翻譯：龍騰網(wǎng) http://www.top-shui.cn 轉(zhuǎn)載請注明出處


The modulators themselves are made of silicon. Though this is not the absolute best material for light modulation, it is by far the best-developed for electronics. Using silicon allows hybrid chips to be made with equipment designed for conventional ones—perhaps even affording it a new lease of life. For, as Maurice Steinman, vice-president of engineering at Lightelligence, observes, though the decades’ long rise in the performance of electronics is slowing down, “we’re just at the beginning of generational scaling on optics”.

這些調(diào)制器本身是由硅制成的。硅不是光調(diào)制的絕佳材料，但它是迄今針對電子元件開發(fā)的最佳材料。如果使用硅材料，設(shè)計用于制造常規(guī)芯片的設(shè)備可以制造混合芯片——甚至可能使它重獲新生。因為正如Lightelligence公司的工程副總裁莫里斯·斯泰因曼所說的，電子元件持續(xù)數(shù)十年的性能提升正在放緩，“但光學(xué)元件的更新?lián)Q代才剛剛開始”。

Ground zero

起點

Ryan Hamerly and his team at the Massachusetts Institute of Technology (the organisation from which Lightelligence and Lightmatter were spun out) seek to exploit the low power consumption of hybrid optical devices for smart speakers, lightweight drones and even self-driving cars. A smart speaker does not have the computational and energetic chops to run deep-learning programs by itself. It therefore sends a digitised version of what it has heard over the internet to a remote server, which does the processing for it. The server then returns the answer.

瑞恩·哈姆利和他的團隊來自麻省理工學(xué)院（Lightelligence和Lightmatter公司的前身），他們設(shè)法將混合光學(xué)設(shè)備的低功耗特性應(yīng)用于智能音箱、輕型無人機、甚至無人駕駛汽車。智能音箱不具備運行深度學(xué)習(xí)程序的運算能力和能量，所以從互聯(lián)網(wǎng)接收數(shù)據(jù)后發(fā)送至遠(yuǎn)端服務(wù)器來運算數(shù)據(jù)，然后將運算結(jié)果反饋給智能音箱。

All this takes time, though, and is insecure. An optical chip put in such a speaker could perform the needed linear algebra there and then, with low power consumption and without having to transfer potentially sensitive data elsewhere.

但這一過程需要時間，并且不安全。智能音箱內(nèi)置的光學(xué)芯片可以進行必要的線性代數(shù)運算，同時做到低功耗，避免了將潛在敏感的數(shù)據(jù)傳輸?shù)絼e的地方。

Other researchers, including Ugur Tegin, at Caltech, reckon optical computing’s true benefit is its ability to handle large data sets. At the moment, for example, image-recognition systems are trained on low-resolution pictures, because high-res versions are too big for them to handle efficiently, if at all. As long as there is an electronic component to the process, there is limited bandwidth. Dr Tegin’s answer is to forgo electronics altogether and use an all-optical machine.

加州理工學(xué)院的烏格·泰吉納等其他研究人員認(rèn)為，光學(xué)運算的真正益處是能夠處理大型數(shù)據(jù)集。例如，目前的圖像識別系統(tǒng)被訓(xùn)練識別低分辨率圖片，因為高分辨率圖片對它們來說太大了，即使能夠處理，效率也很低。只要有電子元件參與處理這一過程，帶寬就會受限。泰吉納博士的解決辦法是徹底放棄電子元件，使用全光學(xué)設(shè)備。

This has, however, proved tricky—for what allows neural networks to learn pretty well any pattern thrown at them is the use, in addition to all the linear processing, of a non-linear function in each of their nodes. Employing only linear functions would mean that only linear patterns could be learned.

但事實證明難以做到——因為要想讓神經(jīng)網(wǎng)絡(luò)學(xué)會幾乎所有的模式，那么除了所有的線性運算，還需要使用每個節(jié)點中的非線性函數(shù)。如果只用線性函數(shù)，意味著神經(jīng)網(wǎng)絡(luò)只能學(xué)會線性模式。

Fortunately, although light does behave mostly in a linear fashion, there is an exception. This, Dr Tegin explains, is when an extremely short and intense pulse of it is shone through a so-called multi-mode fibre, which exploits multiple properties of light to enhance its ability to carry parallel signals. In these circumstances, the pulse’s passage changes the properties of the material itself, altering the behaviour of the passing light in a non-linear manner.

可喜的是，雖然光以線性模式為主，但也有例外。泰吉納博士解釋說，當(dāng)利用所謂的多模光纖傳輸極短和極強的光脈沖時就會出現(xiàn)例外，多模光纖利用光的多種特性來增強其傳輸并行信號的能力。在這種情況下，光脈沖的傳輸改變了材料自身的特性，以非線性的方式改變了光的行為。

Dr Tegin exploited this feature in what is, save its final output layer, an all-optical network. He describes this in a paper published last year in Nature Computational Science. He is able to keep all of the information in an optical form right up until its arrival at the last layer—the one where the answer emerges. Only then is it converted into electronic form, for processing by the simpler and smaller electronic network which makes up this layer.

泰吉納博士將這種特性應(yīng)用于全光網(wǎng)絡(luò)，最終輸出層除外。去年，他在《自然計算科學(xué)》雜志上發(fā)表的論文中闡述了這一原理。他能以光學(xué)形式保存所有的數(shù)據(jù)，直到傳輸至最后一層，即得出運算結(jié)果的那層。直到這時數(shù)據(jù)轉(zhuǎn)化為電子形式，由構(gòu)成該層的比較簡單和小型的電子網(wǎng)絡(luò)進行運算。

Meanwhile, at the University of California, Los Angeles, Aydogan Ozcan is taking yet another approach to all-optical matrix processing. In a paper published in Science in 2018, he and his collaborators describe how to create optical devices that do it without involving electrons at all.

同時在洛杉磯的加州大學(xué)，埃道甘·奧茲坎正在采取另一種方法實現(xiàn)全光學(xué)矩陣運算。在2018年《科學(xué)》雜志上發(fā)表的論文中，他與合作伙伴描述了如何在完全不使用電子元件的情況下制造出這樣的光學(xué)設(shè)備。

The magic here lies in the use of thin sheets of specially fabricated glass, each the size of a postage stamp, laid on top of each other in stacks analogous to the layers of an artificial neural network. Together, these sheets diffract incoming light in the way that such a neural network would process a digital image.

神奇之處是使用特制的薄玻璃片，每片都有郵票大小，相互層疊在一起，就像人工神經(jīng)網(wǎng)絡(luò)的多層結(jié)構(gòu)一樣。這些玻璃片會以神經(jīng)網(wǎng)絡(luò)處理數(shù)字圖像的方式衍射輸入光。

In this case, the optics work passively, like the lens of a camera, rather than receiving active feedback. Dr Ozcan says that provides security benefits. The system never captures images or sends out the raw data—only the inferred result. There is a trade-off, though. Because the sheets cannot be reconfigured they must, if the inference algorithm changes, be replaced.

在這種情況下，光學(xué)器件就像照相機鏡頭一樣被動工作，而不是接收主動反饋。奧茲坎博士說這能帶來安全方面的益處，該系統(tǒng)絕不會采集圖像或發(fā)送原始數(shù)據(jù)，而只會發(fā)送推導(dǎo)的結(jié)果。但這樣做也有弊端，由于無法重新配置玻璃片，所以如果推理算法發(fā)生變化，必須更換玻璃片。

How far optical computing of this sort will get remains to be seen. But ai based on deep learning is developing fast, as recent brouhaha about Chatgpt, a program that can turn out passable prose (and even poetry) with only a little prompting, shows. Hardware which can speed up that development still more is thus likely to find favour. So, after decades in the doldrums, the future of optical computing now looks pretty bright.

這種光學(xué)計算能走多遠(yuǎn)仍有待于觀察。但是，基于深度學(xué)習(xí)的人工智能發(fā)展很快，最近火爆的Chatgpt就是明證，只需略加提示，這種程序就能創(chuàng)作出差強人意的散文（甚至詩歌）。能夠進一步加快其發(fā)展的硬件可能因此受到青睞。所以經(jīng)過幾十年的低迷階段后，現(xiàn)在看來光計算的前景非常光明。