12 results back to index
Machine Translation by Thierry Poibeau
AltaVista, augmented reality, call centre, Claude Shannon: information theory, cloud computing, combinatorial explosion, crowdsourcing, easy for humans, difficult for computers, en.wikipedia.org, Google Glasses, information retrieval, Internet of things, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, natural language processing, Necker cube, Norbert Wiener, RAND corporation, Robert Mercer, Skype, speech recognition, statistical model, technological singularity, Turing test, wikimedia commons
The problem of multiple translations is well known and can be observed even when one tries to translate iteratively between the same two languages (for example, from English to French and then back to English). The prototypical example is probably this Biblical sentence: “The spirit is willing, but the flesh is weak.” The story goes that this was translated into Russian and then translated back into English as “The whiskey is strong, but the meat is rotten.” Yet this is indeed an apocryphal example.1 The longevity of this invented example is due to the comical nature of the resulting translation, but it also illustrates the fact that multiplying translation steps amounts to gradually straying away from the original text until an incomprehensible translation is obtained.
In fact, the report concluded, in terms of costs, a human translator was more affordable than machine translation. At the time, human translators allowed for better and faster translations, as there was no need for additional editing (correcting a text translated entirely by machine often took longer than a direct translation carried out by an experienced translator). The report only considered translations from Russian to English, and, as a result, concluded that the need for Russian to English translation was limited. The largest “consumers” of Russian translation would do better to learn the language itself, the authors suggested.
He imagined a system of 200 primitives capable of representing the function of a word in a sentence, in order to generate the correct translation in the target language (Smirnov-Trojanskij was interested in Russian, where nouns and adjectives are inflected to reflect their function in the sentence). The analyst had to specify whether the word to be translated was the subject or the object, whether the verb was in the present or imperfect tense, and so on. The machine then took over, selecting the correct word form for the translation. The invention focused on a workspace, rather than on a simple device: Smirnov-Trojanskij’s system was designed in such a way that a translator could first simply look for translation elements at word level with the help of the device.
Nerds on Wall Street: Math, Machines and Wired Markets by David J. Leinweber
AI winter, algorithmic trading, asset allocation, banking crisis, barriers to entry, Bear Stearns, Big bang: deregulation of the City of London, business cycle, butter production in bangladesh, butterfly effect, buttonwood tree, buy and hold, buy low sell high, capital asset pricing model, citizen journalism, collateralized debt obligation, corporate governance, Craig Reynolds: boids flock, creative destruction, credit crunch, Credit Default Swap, credit default swaps / collateralized debt obligations, Danny Hillis, demand response, disintermediation, distributed generation, diversification, diversified portfolio, Emanuel Derman, en.wikipedia.org, experimental economics, financial innovation, fixed income, Gordon Gekko, implied volatility, index arbitrage, index fund, information retrieval, intangible asset, Internet Archive, John Nash: game theory, Kenneth Arrow, load shedding, Long Term Capital Management, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, market fragmentation, market microstructure, Mars Rover, Metcalfe’s law, moral hazard, mutually assured destruction, Myron Scholes, natural language processing, negative equity, Network effects, optical character recognition, paper trading, passive investing, pez dispenser, phenotype, prediction markets, quantitative hedge fund, quantitative trading / quantitative ﬁnance, QWERTY keyboard, RAND corporation, random walk, Ray Kurzweil, Renaissance Technologies, risk free rate, risk tolerance, risk-adjusted returns, risk/return, Robert Metcalfe, Ronald Reagan, Rubik’s Cube, Savings and loan crisis, semantic web, Sharpe ratio, short selling, Silicon Valley, Small Order Execution System, smart grid, smart meter, social web, South Sea Bubble, statistical arbitrage, statistical model, Steve Jobs, Steven Levy, Tacoma Narrows Bridge, the scientific method, The Wisdom of Crowds, time value of money, too big to fail, transaction costs, Turing machine, two and twenty, Upton Sinclair, value at risk, Vernor Vinge, yield curve, Yogi Berra, your tax dollars at work
Anthony Oettinger, a pioneer in machine translation at Harvard going back to the 1950s, told a story of an early English-Russian-English system sponsored by U.S. intelligence agencies. The English “The spirit is willing but the flesh is weak” went in, was translated to Russian, which was then sent in again to be translated back into English. The result: “The vodka is ready but the meat is rotten.” Tony got out of the machine translation business. 30. This modern translator is found at www.systransoft.com. I tried Oettinger’s example again, 50 years later. The retranslation of the Russian back to English this time was “The spirit is of willing of but of the flesh is of weak.” 31.
Language Technology: Your Tax Dollars at Work The U.S. government has been busy spending your money on technologies to do this kind of content extraction and analysis for years.When the Gr eatest Hits of Computation in Finance 55 0.04 Cumulative Abnormal Return 0.02 0.00 0.02 0.04 research first started, the language researchers were most interested in was Russian. Harvard’s Tony Oettinger, who led the research, tells of inputting the English sentence “The spirit is willing but the flesh is weak” into an English-to-Russian translation program. The computer’s Russian translation was then fed back into a Russian-to-English translation program. The resulting retranslation was “The vodka is ready but the meat is rotten.” Language technology has improved markedly, but still has a long way to go.
So far, English is the language for almost all of these systems. Machine translation, in general, has been difficult,29 but for literal, as opposed to artistic, content, as is found in most business and financial stories, it can do a passable job. Systran offers a translation system that you can experiment with online.30 The “as the world turns” time zone effect means that many stories will appear first in international sources in languages other than English. The proliferation of cross-listed or economically equivalent securities means there are often trading opportunities in countries that will be learning the news, via translation, later on in the news cycle.
Found in Translation: How Language Shapes Our Lives and Transforms the World by Nataly Kelly, Jost Zetzsche
airport security, Berlin Wall, Celtic Tiger, crowdsourcing, Donald Trump, glass ceiling, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, randomized controlled trial, Ray Kurzweil, Skype, speech recognition, Steve Jobs, the market place
One famous example of machine translation gone awry is actually an urban legend. As the story goes, the sentence “The spirit is willing, but the flesh is weak” was plugged into a machine translation system to be rendered into Russian. Allegedly, the computer produced “The vodka is strong, but the meat is rotten” in Russian. This tale has never been substantiated, but it’s not completely inconceivable. The story probably serves a good purpose as a warning that generic machine translation cannot and should not be blindly trusted. Parlez-Vous C++? Anyone who’s taken a language course in school knows how hard it is to learn a foreign language.
But can prose this pitch perfectly balanced be translated? Here’s one rendering: We can see its similarity, but it’s difficult to evaluate its success without being able to read Russian—or at least the Cyrillic alphabet. But our Russian friends tell us that they feel the same kind of shivers and tingle of excitement when they read it.2 In other words, even such intimate language is translatable. The passage still sounds erotic, even when translated. (Vladimir Nabokov originally wrote Lolita in English. He had the privilege of growing up trilingually in an aristocratic Russian family, and he also personally translated Lolita into Russian.)
In Yashkova’s case, before interpreting for Expedition 6 space walks, she had to go through the entire space walk training along with three astronauts (two American, one Russian) for a period of nine months. “Space walks are very complex,” she points out. “There are hundreds of connectors and very complex tasks that the crew members carry out while in orbit.” Because of the complexity, the role of language is extremely important. “Mission-critical procedures used by the crew members have to be translated,” Yashkova explains. As an interpreter, Yashkova rarely does written translation work—most of that is handled by a small team of translators who specialize in the written translation side, ensuring that all of the terminology complies with industry standards.
Overcomplicated: Technology at the Limits of Comprehension by Samuel Arbesman
algorithmic trading, Anton Chekhov, Apple II, Benoit Mandelbrot, Chekhov's gun, citation needed, combinatorial explosion, Danny Hillis, David Brooks, digital map, discovery of the americas, en.wikipedia.org, Erik Brynjolfsson, Flash crash, friendly AI, game design, Google X / Alphabet X, Googley, HyperCard, Ian Bogost, Inbox Zero, Isaac Newton, iterative process, Kevin Kelly, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, mandelbrot fractal, Minecraft, Netflix Prize, Nicholas Carr, Parkinson's law, Ray Kurzweil, recommendation engine, Richard Feynman, Richard Feynman: Challenger O-ring, Second Machine Age, self-driving car, software studies, statistical model, Steve Jobs, Steve Wozniak, Steven Pinker, Stewart Brand, superintelligent machines, Therac-25, Tyler Cowen: Great Stagnation, urban planning, Watson beat the top human players on Jeopardy!, Whole Earth Catalog, Y2K
For a clear example of the necessary complexity of a machine model for language, we need only look at how computers are used to translate one language into another. Take this great, though apocryphal, story: During the Cold War, scientists began working on a computational method for translating between English and Russian. When they were ready to test their system, they chose a rather nuanced sentence as their test case: “The spirit is willing, but the flesh is weak.” They converted it into Russian, and then ran the resulting Russian translation back again through the machine into English. The result was something like “The whiskey is strong, but the meat is terrible.” Machine translation, as this computational task is more formally known, is not easy.
And what about regional phrases, like the Pittsburghese utterance “The car needs washed” (skipping over “to be”)? The rules will cower in fear before such regionalisms. Using grammatical models to process language for translation simply doesn’t work that well. Language is too complex and quirky for these elegant rules to work when translating a text. There are too many edge cases. Into this gap have stepped numerous statistical approaches from the world of machine learning, in which computers ingest huge amounts of translated texts and then translate new ones based on a set of algorithms, without ever actually trying to understand or parse what the sentences mean. For example, instead of a rule saying that placing the suffix “-s” onto a word makes it plural, the machine might know that “-s” creates a plural word, say, 99.9 percent of the time, whereas 0.1 percent of the time it doesn’t, as with words like “sheep” and “deer” that are their own plurals, or irregular plurals such as “men” or “feet” or even “kine.”
Machine translation, as this computational task is more formally known, is not easy. Google Translate’s results can be imprecise, though interesting in their own way. But scientists have made great strides. What techniques are used by experts in machine translation? One early approach was to use the structured grammatical scaffolding of language I mentioned above. Linguists hard-coded the linguistic properties of language into a piece of software in order to translate from one language to another. But it’s one thing to deal with relatively straightforward sentences, and another to assume that such grammars can handle the diversity of language in the wild.
Epigenetics: How Environment Shapes Our Genes by Richard C. Francis
agricultural Revolution, cellular automata, double helix, Drosophila, epigenetics, experimental subject, longitudinal study, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, meta-analysis, phenotype, stem cell, twin studies
High levels of estradiol and low levels of natural testosterone together are responsible for what is, in the macho world of athletics, one of the most undesirable consequences of steroid abuse: shrunken testicles. Even worse, though their libido remains elevated, many longtime users experience erectile malfunctions, a truly ironic case of “the spirit is willing but the flesh is weak.” Which brings us back to Canseco’s border bust. Canseco was not only a snitch but a proudly self-proclaimed user and advocate of synthetic testosterone. He maintained his regular steroid regimen after his baseball career was over because he liked the way it made him look and feel.
The term transcription is meant to connote the transfer of information from one medium to another, as in musical transcription from, say, piano to guitar. In this case, the transcription is from DNA to RNA. During the second stage, called translation, the mRNA serves as a template for the creation of a protoprotein. The term translation is meant to connote a larger transformation of this information, like that which occurs when one language is translated into another. In protein synthesis, the translation is from the language of the base sequence of the RNA to the amino acid sequence of the protoprotein. The protoprotein is generally not functional. It must be further transformed into a functional protein through a process called posttranslational processing.
Recall from Chapter 2 that protein synthesis occurs in two stages. During the first stage, called transcription, messenger RNA (mRNA) is constructed from the DNA template. During the second stage, called translation, a protoprotein is constructed from the RNA template. Most epigenetic gene regulation occurs at the first stage, usually by inhibiting transcription. MicroRNAs, in contrast, exert their influence during the second stage, translation. Though transcription is highly regulated, it is often the case that there are too many mRNA transcripts from a particular gene for a cell’s purposes. If a cell “decides” that this is the case, it deploys microRNAs to remedy the situation.
The Last Lingua Franca: English Until the Return of Babel by Nicholas Ostler
barriers to entry, BRICs, British Empire, call centre, en.wikipedia.org, European colonialism, Internet Archive, invention of writing, Isaac Newton, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, mass immigration, Nelson Mandela, open economy, Republic of Letters, Scramble for Africa, statistical model, trade route, upwardly mobile
Without receiving preprogrammed guidance on the geometry of the various worlds of modern westernized homes, schools, and offices, a system could not easily select the correct sense for pen (‘writing instrument’, as against ‘enclosure’) in “The pen is in the box” as against “The box is in the pen.” Nor was there any general means of selecting appropriate equivalents when language was used metaphor ical ly. If it is ridiculous— because irrelevant to the context— to see references to good vodka and meat of dubious quality when encountering “The spirit is willing, but the flesh is weak” in an essay on government policy, how come in an environmental work “the rape of the countryside” does not refer to oilseed rape, a crop ubiquitous in the fields of modern England? Human reason, and even more human rhetoric, is inclined to be inscrutable. In practice, it seemed to be impossible to divorce the syntactic part of language processing from modeling the meaning of particular texts.
Although the recently acceded countries, mostly from central and eastern Europe, might have been expected to be more familiar with the use of German, or indeed Russian, in practice they seem to prefer to use English; and with more and more target languages requiring translation into them, there is a premium on reducing the number of languages from which the translation is done. The result has been increasing pressure to use English, and English only, as a common medium. Meanwhile, the other established lingua-francas of the Continent’s regions (including French and German, but also Spanish, Italian, and Russian) have enjoyed no special status as such, although it is recognized that they are the languages that— besides English— are most widely understood.* German is the native language of 18 percent of the EU population, but has been acquired as a lingua-franca by another 14 percent.
This Euro-centric situation continued for the next four years, until 2005, when Chinese, Japanese, and Korean were added. In a sense, Asia—or the Pacific Ocean— had arrived. All the new languages were still translated only to or from English. Two years later, two more languages were added (Arabic and Russian), and by November the same year (2007) yet two more (Dutch and Greek). But the doctrine of English as the only colanguage was maintained: essentially, you could translate to and from any of twelve foreign languages now—as long as your source or target was English.* It was only in 2008 that truly multilingual translation initiated. The sluice gates have been opened, in two respects. First, the number of languages began to go up exponentially (to twenty-three in late 2008, then up to fifty-one in 2009);† but second, the primacy of English as the universal colanguage was eliminated.5 Whatever the workings internally (English is likely seeing a fair amount of electronic ser vice as an interlingua within the system), the user now simply sets his or her source and target from the set of available languages and can currently choose from 2,550 directed language-pairs.
The Globotics Upheaval: Globalisation, Robotics and the Future of Work by Richard Baldwin
agricultural Revolution, Airbnb, AltaVista, Amazon Web Services, augmented reality, autonomous vehicles, basic income, business process, business process outsourcing, call centre, Capital in the Twenty-First Century by Thomas Piketty, Cass Sunstein, commoditize, computer vision, Corn Laws, correlation does not imply causation, Credit Default Swap, David Ricardo: comparative advantage, declining real wages, deindustrialization, deskilling, Donald Trump, Douglas Hofstadter, Downton Abbey, Elon Musk, Erik Brynjolfsson, facts on the ground, future of journalism, future of work, George Gilder, Google Glasses, Google Hangouts, hiring and firing, impulse control, income inequality, industrial robot, intangible asset, Internet of things, invisible hand, James Watt: steam engine, Jeff Bezos, job automation, knowledge worker, laissez-faire capitalism, low skilled workers, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, manufacturing employment, Mark Zuckerberg, mass immigration, mass incarceration, Metcalfe’s law, new economy, optical character recognition, pattern recognition, Ponzi scheme, post-industrial society, post-work, profit motive, remote working, reshoring, ride hailing / ride sharing, Robert Gordon, Robert Metcalfe, robotic process automation, Ronald Reagan, San Francisco homelessness, Second Machine Age, self-driving car, side project, Silicon Valley, Skype, Snapchat, social intelligence, sovereign wealth fund, standardized shipping container, statistical model, Stephen Hawking, Steve Jobs, supply-chain management, TaskRabbit, telepresence, telepresence robot, telerobotics, Thomas Malthus, trade liberalization, universal basic income
In June 2017, the US Army paid Raytheon four million dollars for a machine translation package that lets soldiers converse with Iraqi Arabic and Pashto speakers as well as read foreign-language documents and digital media on their smartphones and laptops. Machine translation used to be a joke. A famous example, related by Google’s director of research Peter Norvig, was what old-school machine translators did with the phrase, “the spirit is willing but the flesh is weak.” Translated into Russian and then back to English, it turned into “the vodka is good but the meat is rotten.”7 Even as recently as 2015, it was little more than a party trick, or a very rough first draft. But no longer. Now it is rivaling average human translation for popular language pairs. According to Google, which uses humans to score machine translations on a scale from zero (complete nonsense) to six (perfect), the AI-trained algorithm “Google Translate” got a grade of 3.6 in 2015—far worse than the average human translator, who gets scores around 5.1.
THE DEEP LEARNING TAKEOVER For a decade, hundreds of Google engineers made incremental progress on translation using the traditional, hands-on approach. In February 2016, Google’s AI maharishi, Jeff Dean, turned the Google Translate team on to Google’s homegrown machine-learning technique called Deep Learning. The job required huge amounts of computer muscle, but Google had that thanks to Moore’s law. The missing link was the data. That changed in 2016 when the United Nations (UN) posted online a data set with nearly 800,000 documents that had been manually translated into the six official UN languages: Arabic, English, Spanish, French, Russian, and Chinese. It is worth reflecting for a moment on how difficult it would have been to create, store, and upload that much data just a few years ago.
There is no direct equivalence of one language to another.”10 What this suggests is that the high-end translation is likely to stay in the hands of humans, but in the meantime international business will be transformed when these uninspired apprentice translators massively lower, but don’t eliminate, language barriers. Instant, free machine translation is not something that is lurking in computer laboratories. Free apps like Google Translate and iTranslate Voice are now quite good across the major language pairs. Other smart-phone apps include SayHi and WayGo. And machine translation is widely used. Google, for example, does a billion translations a day for online users. Try it out. Machine translation works on any smartphone. Just open up a foreign language website and apply Google Translate to the text. You can even use the iTranslate app to instantly translate a foreign language in real time.
Talk to Me: How Voice Computing Will Transform the Way We Live, Work, and Think by James Vlahos
Albert Einstein, AltaVista, Amazon Mechanical Turk, Amazon Web Services, augmented reality, Automated Insights, autonomous vehicles, backpropagation, Chuck Templeton: OpenTable:, cloud computing, Colossal Cave Adventure, computer age, Donald Trump, Elon Musk, information retrieval, Internet of things, Jacques de Vaucanson, Jeff Bezos, lateral thinking, Loebner Prize, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, Mark Zuckerberg, Menlo Park, natural language processing, PageRank, pattern recognition, Ponzi scheme, randomized controlled trial, Ray Kurzweil, Ronald Reagan, Rubik’s Cube, self-driving car, sentiment analysis, Silicon Valley, Skype, Snapchat, speech recognition, statistical model, Steve Jobs, Steve Wozniak, Steven Levy, Turing test, Watson beat the top human players on Jeopardy!
Enabling such capabilities would be difficult, Winograd knew, because meaning isn’t conveyed by words alone. Instead, people combine what they hear with the information already inside their heads. Computers that lack background knowledge are severely handicapped. Winograd gave the example of a computer that was programmed to translate words from Russian to English by looking up dictionary definitions. Presented with the Russian equivalent of “The spirit is willing but the flesh is weak,” the computer might generate, “The vodka is strong but the meat is rotten.” Creating a computer with human-level knowledge and reasoning power isn’t possible today, much less in the days of the primitive computers of the late 1960s.
In the 1950s, attention in the West shifted to a new enemy, Russia, and a new code of sorts, the Russian language. Intelligence officers figured that if computers could be taught to understand Russian and convert it to English, they could accomplish much more than human translators alone could do, a valuable boost during the Cold War. In 1954 Léon Dostert, a professor at Georgetown University who was working in conjunction with IBM, demonstrated a pioneering translation system. A woman who didn’t know any Russian sat at a keyboard before IBM’s first commercially available computer, the room-filling 701. She typed a Russian sentence that had been transliterated as “Mi pyeredayem mislyi posryedstvom ryechyi.”
Then one neural network encodes a phrase as a vector, and a second network decodes that vector as a translated phrase in a new language. The method, which is known as sequence-to-sequence, was so effective that in 2016 Google dumped the old version of Translate, which used statistical machine translation, and replaced it with the sequence-to-sequence one. In a period of months, the new system improved by margins that had taken the old system years to accomplish. So here comes the natural-language-generation part. Two of the Google researchers who pioneered the sequence-to-sequence technique, Oriol Vinyals and Quoc Le, had an idea. They figured that conversation, like translation, might be a problem in which you were trying to encode one sequence (what a person said) and decode it as a second sequence (what the computer should reply).
Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die by Eric Siegel
Albert Einstein, algorithmic trading, Amazon Mechanical Turk, Apple's 1984 Super Bowl advert, backtesting, Black Swan, book scanning, bounce rate, business intelligence, business process, butter production in bangladesh, call centre, Charles Lindbergh, commoditize, computer age, conceptual framework, correlation does not imply causation, crowdsourcing, dark matter, data is the new oil, en.wikipedia.org, Erik Brynjolfsson, Everything should be made as simple as possible, experimental subject, Google Glasses, happiness index / gross national happiness, job satisfaction, Johann Wolfgang von Goethe, lifelogging, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, mass immigration, Moneyball by Michael Lewis explains big data, Nate Silver, natural language processing, Netflix Prize, Network effects, Norbert Wiener, personalized medicine, placebo effect, prediction markets, Ray Kurzweil, recommendation engine, risk-adjusted returns, Ronald Coase, Search for Extraterrestrial Intelligence, self-driving car, sentiment analysis, Shai Danziger, software as a service, speech recognition, statistical model, Steven Levy, text mining, the scientific method, The Signal and the Noise by Nate Silver, The Wisdom of Crowds, Thomas Bayes, Thomas Davenport, Turing test, Watson beat the top human players on Jeopardy!, X Prize, Yogi Berra, zero-sum game
Generating human language trips up the naïve machine. I once received a voice-synthesized call from Blockbuster reminding me of my rented movie’s due date. “This is a message for Eric the Fifth Siegel,” it said. My middle initial is V. Translation between languages also faces hazards. An often-cited example is that “The spirit is willing, but the flesh is weak,” if translated into Russian and back, could end up as “The vodka is good, but the meat is rotten.” 3Watson’s avatar, its visual depiction shown on Jeopardy!, consists of 42 glowing, crisscrossing threads as an inside joke and homage that references the significance this number holds in Adams’s infamous Hitchhiker’s Guide. 4Watson was not named after this fictional detective—it was named after IBM founder Thomas J.
Natural language processing: Dursun Delen, Andrew Fast, Thomas Hill, Robert Nisbit, John Elder, and Gary Miner, Practical Text Mining and Statistical Analysis for Non-Structured Text Data Applications (Academic Press, 2012). James Allen, Natural Language Understanding, 2nd ed. (Addison-Wesley, 1994). Regarding the translation of “The spirit is willing, but the flesh is weak”: John Hutchins, “The Whisky Was Invisible or Persistent Myths of MT,” MT News International 11 (June 1995), 17–18. www.hutchinsweb.me.uk/MTNI-11–1995.pdf. Ruminations on Apple’s Siri versus Watson from WolframAlpha’s creator: Stephen Wolfram, “Jeopardy, IBM, and WolframAlpha,” Stephen Wolfram blog, January 26, 2011. http://blog.stephenwolfram.com/2011/01/jeopardy-ibm-and-wolframalpha/.
This automation is the means by which PA builds its predictive power. The hunter returns back to the tribe, proudly displaying his kill. So, too, a data scientist posts her model on the bulletin board near the company ping-pong table. The hunter hands over the kill to the cook, and the data scientist cooks up her model, translates it to a standard computer language, and e-mails it to an engineer for integration. A well-fed tribe shows the love; a psyched executive issues a bonus. The tribe munches and the scientist crunches. To Act Is to Decide Knowing is not enough; we must act. —Johann Wolfgang von Goethe Potatoes or rice?
Machine, Platform, Crowd: Harnessing Our Digital Future by Andrew McAfee, Erik Brynjolfsson
"Robert Solow", 3D printing, additive manufacturing, AI winter, Airbnb, airline deregulation, airport security, Albert Einstein, algorithmic bias, Amazon Mechanical Turk, Amazon Web Services, artificial general intelligence, augmented reality, autonomous vehicles, backpropagation, backtesting, barriers to entry, bitcoin, blockchain, British Empire, business cycle, business process, carbon footprint, Cass Sunstein, centralized clearinghouse, Chris Urmson, cloud computing, cognitive bias, commoditize, complexity theory, computer age, creative destruction, crony capitalism, crowdsourcing, cryptocurrency, Daniel Kahneman / Amos Tversky, Dean Kamen, discovery of DNA, disintermediation, disruptive innovation, distributed ledger, double helix, Elon Musk, en.wikipedia.org, Erik Brynjolfsson, Ethereum, ethereum blockchain, everywhere but in the productivity statistics, family office, fiat currency, financial innovation, George Akerlof, global supply chain, Hernando de Soto, hive mind, independent contractor, information asymmetry, Internet of things, inventory management, iterative process, Jean Tirole, Jeff Bezos, jimmy wales, John Markoff, joint-stock company, Joseph Schumpeter, Kickstarter, law of one price, longitudinal study, Lyft, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, Marc Andreessen, Mark Zuckerberg, meta-analysis, Mitch Kapor, moral hazard, multi-sided market, Myron Scholes, natural language processing, Network effects, new economy, Norbert Wiener, Oculus Rift, PageRank, pattern recognition, peer-to-peer lending, performance metric, Plutocrats, plutocrats, precision agriculture, prediction markets, pre–internet, price stability, principal–agent problem, Ray Kurzweil, Renaissance Technologies, Richard Stallman, ride hailing / ride sharing, risk tolerance, Ronald Coase, Satoshi Nakamoto, Second Machine Age, self-driving car, sharing economy, Silicon Valley, Skype, slashdot, smart contracts, Snapchat, speech recognition, statistical model, Steve Ballmer, Steve Jobs, Steven Pinker, supply-chain management, TaskRabbit, Ted Nelson, The Market for Lemons, The Nature of the Firm, the strength of weak ties, Thomas Davenport, Thomas L Friedman, too big to fail, transaction costs, transportation-network company, traveling salesman, Travis Kalanick, two-sided market, Uber and Lyft, Uber for X, uber lyft, ubercab, Watson beat the top human players on Jeopardy!, winner-take-all economy, yield management, zero day
Other challenges, however, proved much less amenable to a rule-based approach. Decades of research in speech recognition, image classification, language translation, and other domains yielded unimpressive results. The best of these systems achieved much worse than human-level performance, and the worst were memorably bad. According to a 1979 collection of anecdotes, for example, researchers gave their English-to-Russian translation utility the phrase “The spirit is willing, but the flesh is weak.” The program responded with the Russian equivalent of “The whisky is agreeable, but the meat has gone bad.” This story is probably apocryphal, but it’s not an exaggeration.
language=en. 82 “Our vehicles were driving through Mountain View”: Ibid. 83 The Japanese insurer Fukoku Mutual Life: Dave Gershgorn, “Japanese White-Collar Workers Are Already Being Replaced by Artificial Intelligence,” Quartz, January 2, 2017, https://qz.com/875491/japanese-white-collar-workers-are-already-being-replaced-by-artificial-intelligence. 83 “learn the history of past payment assessment”: Google Translate, “December 26, Heisei 28, Fukoku Life Insurance Company,” accessed January 30, 2017, https://translate.google.com/translate?depth=1#x0026;hl=en#x0026;prev=search#x0026;rurl=translate.google.com#x0026;sl=ja#x0026;sp=nmt4#x0026;u=http://www.fukoku-life.co.jp/about/news/download/20161226.pdf. 84 In October of 2016: Allison Linn, “Historic Achievement: Microsoft Researchers Reach Human Parity in Conversational Speech Recognition,” Microsoft (blog), October 18, 2016, http://blogs.microsoft.com/next/2016/10/18/historic-achievement-microsoft-researchers-reach-human-parity-conversational-speech-recognition/#sm.0001d0t49dx0veqdsh21cccecz0e3. 84 “I must confess that I never thought”: Mark Liberman, “Human Parity in Conversational Speech Recognition,” Language Log (blog), October 18, 2016, http://languagelog.ldc.upenn.edu/nll/?
siteedition=uk#axzz3QsbvnchO. 190 “BlaBlaCar drivers don’t make a profit”: Laura Wagner, “What Does French Ride-Sharing Company BlaBlaCar Have That Uber Doesn’t,” Two-Way, September 16, 2015, http://www.npr.org/sections/thetwo-way/2015/09/16/440919462/what-has-french-ride-sharing-company-blablacar-got-that-uber-doesnt. 190 the average BlaBlaCar trip is 200 miles: “BlaBlaCar: Something to Chat About,” Economist, October 22, 2015, http://www.economist.com/news/business/21676816-16-billion-french-startup-revs-up-something-chat-about. 191 operating in twenty-one countries: BlaBlaCar, accessed February 5, 2017, https://www.blablacar.com. 191 facilitating over 10 million rides every quarter: Rawn Shah, “Driving Ridesharing Success at BlaBlaCar with Online Community,” Forbes, February 21, 2016, http://www.forbes.com/sites/rawnshah/2016/02/21/driving-ridesharing-success-at-blablacar-with-online-community/#5271e05b79a6. 191 $550 million in investor funding: Yoolim Lee, “Go-Jek Raises Over $550 Million in KKR, Warburg-Led Round,” Bloomberg, last modified August 5, 2016, https://www.bloomberg.com/news/articles/2016-08-04/go-jek-said-to-raise-over-550-million-in-kkr-warburg-led-round. 191 $15: Steven Millward, “China’s Top ‘Uber for Laundry’ Startup Cleans Up with $100M Series B Funding,” Tech in Asia, August 7, 2015, https://www.techinasia.com/china-uber-for-laundry-edaixi-100-million-funding. 191 100,000 orders per day: Emma Lee, “Tencent-Backed Laundry App Edaixi Nabs $100M USD from Baidu,” TechNode, August 6, 2015, http://technode.com/2015/08/06/edaixi-series-b. 191 twenty-eight cities with a combined 110 million residents: Edaixi, accessed February 5, 2017, http://www.edaixi.com/home/about. (English version: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=http%3A%2F%2Fwww.edaixi.com%2Fhome%2Fabout.) 191 Guagua Xiche: Guagua Xiche, accessed February 5, 2017, http://www.guaguaxiche.com/web/about.html. 191 $58 million in 2015: C. Custer, “2015 Has Been Brutal to China’s O2O Car Wash Services,” Tech in Asia, November 2, 2015, https://www.techinasia.com/2015-brutal-chinas-o2o-car-wash-services. 192 Hao Chushi: Hao Chushi, accessed February 5, 2017, http://www.chushi007.com/index.html. 192 approximately $15: Jamie Fullerton, “China’s New App Brings Chefs to Cook in Your Home,” Munchies, April 8, 2015, https://munchies.vice.com/en/articles/chinas-new-app-brings-world-class-chefs-to-cook-in-your-home. 192 Ele.me, which raised over $1 billion: C.
How the Mind Works by Steven Pinker
affirmative action, agricultural Revolution, Alfred Russel Wallace, backpropagation, Buckminster Fuller, cognitive dissonance, Columbine, combinatorial explosion, complexity theory, computer age, computer vision, Daniel Kahneman / Amos Tversky, delayed gratification, disinformation, double helix, experimental subject, feminist movement, four colour theorem, Gordon Gekko, greed is good, hedonic treadmill, Henri Poincaré, income per capita, information retrieval, invention of agriculture, invention of the wheel, Johannes Kepler, John von Neumann, lake wobegon effect, lateral thinking, Machine translation of "The spirit is willing, but the flesh is weak." to Russian and back, Mikhail Gorbachev, Murray Gell-Mann, mutually assured destruction, Necker cube, out of africa, pattern recognition, phenotype, Plutocrats, plutocrats, random walk, Richard Feynman, Ronald Reagan, Rubik’s Cube, Saturday Night Live, scientific worldview, Search for Extraterrestrial Intelligence, sexual politics, social intelligence, Steven Pinker, theory of mind, Thorstein Veblen, Turing machine, urban decay, Yogi Berra
He observes that “people behave sometimes as if they had two selves, one who wants clean lungs and long life and another who adores tobacco, or one who wants a lean body and another who wants dessert, or one who yearns to improve himself by reading Adam Smith on self-command … and another who would rather watch an old movie on television. The two are in continual contest for control.” When the spirit is willing but the flesh is weak, such as in pondering a diet-busting dessert, we can feel two very different kinds of motives fighting within us, one responding to sights and smells, the other to doctors’ advice. What about when the rewards are of the same kind, like a dollar today versus two dollars tomorrow?
The enigma of war is why people volunteer for an activity that has an excellent chance of getting them killed. How could a desire to play Russian roulette have evolved? Tooby and Cosmides explain it by the fact that natural selection favors traits that increase fitness on average. Every gene contributing to a trait is embodied in many individuals in many generations, so if one individual with the gene dies childless, the success of many others with the gene can make up for it. Imagine a game of Russian roulette where if you don’t get killed you have one more offspring. A gene for joining in the game could be selected, because five-sixths of the time it would leave an extra copy in the gene pool and one-sixth of the time it would leave none.
You are about to press the button. The nation’s policy is to retaliate in kind against a nuclear attack. The policy was designed to deter attackers; if you don’t follow through, the deterrent would have been a sham. On the other hand, you are thinking, the damage has been done. Killing millions of Russians will not bring millions of dead Americans back to life. The bomb will add radioactive fallout to the atmosphere, harming your own citizens. And you will go down in history as one of the worst mass murderers of all time. Retaliation now would be sheer spite. But then, it is precisely this line of thinking that emboldened the Soviets to attack.
The Complete Novels Of George Orwell by George Orwell
Her youthful, rapacious face, with blackened eyebrows, leaned over him as he sprawled there. ‘How about my present?’ she demanded, half wheedling, half menacing. Never mind that now. To work! Come here. Not a bad mouth. Come here. Come closer. Ah! No. No use. Impossible. The will but not the way. The spirit is willing but the flesh is weak. Try again. No. The booze, it must be. See Macbeth. One last try. No, no use. Not this evening, I’m afraid. All right, Dora, don’t you worry. You’ll get your two quid all right. We aren’t paying by results. He made a clumsy gesture. ‘Here, give us that bottle. That bottle off the dressing–table.’
Men were dying because they would not abandon their true beliefs. Naturally all the glory belonged to the victim and all the shame to the Inquisitor who burned him. Later, in the twentieth century, there were the totalitarians, as they were called. There were the German Nazis and the Russian Communists. The Russians persecuted heresy more cruelly than the Inquisition had done. And they imagined that they had learned from the mistakes of the past; they knew, at any rate, that one must not make martyrs. Before they exposed their victims to public trial, they deliberately set themselves to destroy their dignity.
In the future such fragments, even if they chanced to survive, would be unintelligible and untranslatable. It was impossible to translate any passage of Oldspeak into Newspeak unless it either referred to some technical process or some very simple everyday action, or was already orthodox (goodthinkful would be the Newspeak expression) in tendency. In practice this meant that no book written before approximately 1960 could be translated as a whole. Pre-revolutionary literature could only be subjected to ideological translation–that is, alteration in sense as well as language. Take for example the well-known passage from the Declaration of Independence: We hold these truths to be self-evident, that all men are created equal, that they are endowed by their creator with certain inalienable rights, that among these are life, liberty, and the pursuit of happiness.