Hombre Máquina v: Las computadoras pueden cocinar, Escribir y pintar mejor que nosotros?

Man v Machine: Can Computers Cook, Write and Paint Better Than Us?

La inteligencia artificial ahora puede ganar un partido, reconocer su cara, incluso recurrir contra el ticket de aparcamiento. Pero, ¿puede hacer las cosas incluso los seres humanos encuentran complicado?


Desarrollado por Guardian.co.ukEste artículo titulado “V hombre máquina: pueden cocinar ordenadores, escribir y pintar mejor que nosotros?” fue escrito por Leo Benedictus, para The Guardian el sábado 4 de junio de 2016 08.00 UTC

un vídeo, para mí, cambió todo. Es imágenes del viejo juego de Atari Fugarse, aquella en la que se desliza una paleta de izquierda a derecha a lo largo de la parte inferior de la pantalla, tratando de destruir los ladrillos haciendo rebotar una pelota en ellas. Es posible que haya leído sobre el jugador del juego: un algoritmo desarrollado por DeepMind, la compañía británica de inteligencia artificial cuya AlphaGo programa también venció a uno de los mejores jugadores alguna vez, Lee Sedol, a principios de este año.

Tal vez se puede esperar una computadora para ser bueno en los juegos de ordenador? Una vez que sepan qué hacer, que sin duda lo hace más rápido y más consistente que cualquier ser humano. El jugador del desbloqueo de DeepMind nada sabía, sin embargo. No se ha programado con instrucciones sobre cómo funciona el juego; que ni siquiera se le dijo cómo utilizar los controles. Todo lo que tenía era la imagen en la pantalla y el comando para tratar de conseguir tantos puntos como sea posible.

Reloj el video. Primero, la paleta deja caer la pelota en el olvido, sabiendo hay mejor. Finalmente, simplemente curioseaba, que golpea la pelota de vuelta, destruye un ladrillo y obtiene un punto, por lo que reconoce esto y lo hace más a menudo. Después de dos horas de práctica, o alrededor 300 Juegos, se ha convertido en verdad una buena, mejor que tú o yo nunca será. Entonces, despues de 600 Juegos, las cosas se ponen fantasmagórico. El algoritmo comienza apuntando al mismo punto, una y otra vez, con el fin de excavar a través de los ladrillos en el espacio detrás. Una vez ahí, como cualquier jugador sabe Breakout, la pelota rebota alrededor de un rato, la recopilación de puntos gratis. Es una buena estrategia que el equipo se le ocurrió en su propia.

"Cuando nuestros investigadores vieron esto, que en realidad les conmocionado,"CEO de DeepMind, Demis Hassabis, dijo a una audiencia en una conferencia tecnológica en París. Usted puede ver su demostración, demasiado, y oír las risas y aplausos cuando la máquina se da cuenta de su estrategia de madriguera. El ordenador se ha convertido inteligente, un poco como nosotros.

"Inteligencia artificial" se trata sólo de los más antiguos y más publicitado de las frases de moda de todo computación. La idea fue planteada por primera vez en serio Alan Turing en Máquinas e inteligencia, la 1950 de papel en el que propuso lo que se conoce como la prueba de Turing: si una máquina podría convencer a usted a través de la conversación que era humana, que estaba haciendo tanto como cualquier ser humano podía para demostrar que estaba pensando realmente. Pero el término IA no se utiliza generalmente hasta 1955, cuando matemático estadounidense John McCarthy propuso una conferencia para expertos. Esto se llevó a cabo el año siguiente, y desde entonces el campo se ha quedado en un ciclo de dos décadas más o menos de la manía y la desesperación. (Los investigadores aún tienen un nuevo término - "invierno AI" - para describir sus hechizos de moda. Los años 1970 y 1990 fueron especialmente duras.)

Hoy en día hay una nueva manía, que se ve diferente de los demás: que cabe en el bolsillo. Un teléfono puede vencer al campeón mundial de ajedrez, reconocer canciones en la radio y fotos de sus hijos, y traducir su voz en otro idioma. El robot Nao foto con Yotam Ottolenghi puede caminar sobre dos piernas, hablar, encontrar una pelota e incluso la danza. (Es un robot, aunque, no AI: no se puede diseñar un menú.)

Escuchar acerca de los avances de la AI, usted no necesita un experto para decirle al ser excitado, o asustado. Que acaba de empezar a tener la sensación: la inteligencia está aquí. Está claro que Google tiene la sensación, demasiado, porque compró DeepMind de $ 650 millones rumoreado. En 2013, Facebook lanzó su propio proyecto, con planes para desarrollar el reconocimiento facial y lenguaje natural para el sitio. Los desarrolladores ya han comenzado a trabajar en chatbots inteligentes, el que los usuarios de Facebook podrán convocar utilizando su servicio de Messenger.

Hasta aquí, las computadoras no han sido "inteligente" en absoluto, o por un estrecho margen por lo. Han sido buenos en tareas fáciles que nos deslumbran, como las matemáticas, pero malo en los que damos por sentado, los cuales resultan ser seriamente duro. El acto de caminar es algo que los modernos robots aprenden como bebés y todavía lucha con; Arreglar tareas básicas siguen siendo sueños lejanos. "Un ejemplo es la facilidad con la que usted o yo podríamos hacer una taza de té en la cocina de otra persona,"dice El profesor Alan Winfield, un experto en robótica de la Universidad del Oeste de Inglaterra. "No es un robot en el planeta que podría hacer esto."

Para entender por qué ser humano es tan difícil, pensar en cómo se puede conseguir una computadora para reconocer a las personas a partir de fotografías. sin AI, usted tiene que saber cómo lo haces a ti mismo primero, con el fin de programar la computadora. Usted tiene que recoger y pensar en todos los posibles patrones, colores y formas de las caras, y cómo cambiar a la luz y en diferentes ángulos - y usted tiene que saber lo que es importante y lo que es sólo barro en la lente. con AI, usted no tiene que explicar: que acaba de dar una montaña de datos reales a un ordenador y dejar que aprenda. ¿Cómo se diseña el software de aprendizaje sigue siendo una cuestión esotérica, la provincia de unos científicos de la computación codiciados, pero está claro que tienen a un ganador mediante el diseño de estructuras de procesamiento de datos basado libremente en las estructuras en el cerebro. (Esto se denomina "aprendizaje profundo".) En cuanto a las montañas de datos reales, bien, eso es lo que Google, Facebook, Amazonas, Uber y todo lo demás ocurre que quienes por ahí.

En este punto, todavía no sabemos que utiliza la IA resultará mejor. Josh Newlan, un codificador de California que trabaja en Shanghai, se aburrió con la escucha de llamadas de conferencia sin fin, así construyó un cierto software para escuchar por él. Ahora, cada vez que el nombre de Newlan se menciona, su equipo al instante le envía una transcripción de la última media hora, murga 15 segundo, luego reproduce una grabación de él diciendo:, "Lo siento, No me di cuenta que mi micrófono estaba en silencio ". El año pasado, Josh Browder, un adolescente británico, construido una abogado artificial libre que hace un llamamiento en contra de multas de estacionamiento; que planea construir otra para guiar a los refugiados a través de los sistemas jurídicos extranjeros. Las posibilidades son ... Bueno, tal vez un algoritmo puede contar las posibilidades.

Así se mentes una máquina de día superar a nuestra propia? Los investigadores yo os he hablado son cautelosos, y hacer esfuerzos para enfatizar lo que sus máquinas no pueden hacer. Pero decidí poner a prueba AI: puede planear una comida, así como Ottolenghi? ¿Puede mi retrato? ¿Es la tecnología de inteligencia artificial todavía - o está empezando a ser inteligentes, de verdad?

La prueba de cocción

Bien, Yo digo que no es horrible. Los seres humanos han servido mi peor. Aunque a decir verdad el nombre que de IBM Watson Chef da este plato ("Salsa de hígado de pollo salado") es casi tan apetitosa como se merece.

Para ser justos en Chef Watson, y al propio chef-columnista del Guardian Weekend Yotam Ottolenghi, Los había fijado una tarea. Pedí un plato a base de cuatro ingredientes que parecía pertenecer a ninguna parte cerca de la otra: hígados de pollo, yogur griego, wasabi y tequila. Se podría añadir cualquier otra cosa que le gusta, pero los cuatro tenían que estar en el plato terminado, lo que me gustaría cocinar y comer. Chef Watson no vaciló, instantáneamente me da dos salsas para pasta. Ottolenghi fue más perspicaz. "Cuando llegué al desafío que pensé, 'Esto no va a funcionar,'" el me dice.

Pensé lo mismo. O al menos yo pensaba que iba a terminar de comer dos platos que lograron estar bien a pesar de sus ingredientes, en lugar de a causa de ellas. De hecho - y me lo piensa un arrastramiento, pero ¿y qué - la receta de Ottolenghi fue una revelación: hígado y cebolla y una reducción del tequila, servido con una manzana, rábano, remolacha y repollo achicoria, con un wasabi y yogur vestidor. El plato puede tener poco sentido en el papel, pero devoré una sensación de plato que cada elemento pertenecía. (Y vinagreta espesa con yogur y wasabi en lugar de mostaza: seriamente, darle una oportunidad.) Ottolenghi me dice que la receta es sólo un corto bigote de publicable.

La cosa es, ese plato él y su equipo tomó tres días para perfeccionar. Ellos fueron capaces de probar y discutir sabores, texturas, colores, temperaturas, de una manera que Watson no puede - aunque ha habido "discusiones" sobre la adición de un mecanismo de retroalimentación en el futuro, Chef ingeniero principal de Watson, Florian Pinel, me dice. "Una receta es una cosa tan compleja,"Dice Ottolenghi. "Es difícil para mí incluso a entender cómo un equipo se acercaba a ella."

Yotam Ottolenghi y los platos del chef Watson
Yotam Ottolenghi y los platos del chef Watson Fotografía: Jay Brooks por The Guardian

Watson fue construido por primera vez por IBM para ganar el programa de juego Jeopardy televisión! en 2011. En algunos aspectos, era un reto engañosa, because for a computer the tough part of a quiz is understanding the questions, not knowing the answers; for humans, it’s the other way around. But Watson won, and its technology began to be applied elsewhere, including as a chef, generating new recipes based on 10,000 real examples taken from Bon Appétit magazine.

First the software had to “ingest” these recipes, as the Watson team put it. A lot of computation went into understanding what the ingredients were, how they were prepared, how long they were cooked for, in order to be able to explain how to use them in new dishes. (The process can still go awry. Even now Chef Watson recommends an ingredient called “Mollusk”, which it helpfully explains is “the sixth full-length album by Ween”.)

A bigger problem was trying to give the machine a sense of taste. “It’s easy enough for a computer to create a novel combination,” Pinel says, “but how can it evaluate one?” Watson was taught to consider each ingredient as a combination of specific flavour compounds – of which there are thousands – and then to combine ingredients that had compounds in common. (This principle, food pairing, is well established among humans.) Finalmente, the software generates step-by-step instructions that make sense to a human cook. The emphasis is on surprises rather than practical meal planning. “Chef Watson is really there to inspire you,” Pinel explains. Each recipe comes with the reminder to “use your own creativity and judgment”.

And I need to. The first step is to “toast flat-leaf parsley”, which just isn’t a good idea. I am making, effectively, a slow-cooked spiced pork and beef ragu, including all my four ingredients, yet Watson oddly also includes cucumber and keeps telling me to “season with allspice”, which I refuse to do on principle. In the end, I have a rich sauce with a flavour rather close to the farmyard, but not uneatable. I can’t taste the wasabi or the tequila, which I’m glad about.

Yotam Ottolenghi with Nao robot
Yotam Ottolenghi with Nao robot loaned courtesy of Heber primary school, Londres. Fotografía: Jay Brooks. Styling: Lee Flude

Watson is clever and the task is tough, but I am ready to say that this is no more than a bit of fun for food nerds, until Ottolenghi stops me. “I think the idea of slow-cooking the livers with a bit of meat is great,"dice. “It intensifies the flavour. Everything will come together. If I had to start afresh with this recipe, obviously the yoghurt doesn’t fit – but I would leave the orange skin there, a few of the spices. I don’t think it’s a very bad recipe. It could work.”

Veredicto Watson hides the weirdness of the ingredients, but Ottolenghi makes them sing.

The writing test

Put little Wordsmith next to the fearsome machines of IBM and Google, and it looks as computationally advanced as a pocket calculator. Yet while Watson fumbles through its apprenticeship, Wordsmith is already at work. If you read stock market reports from the Associated Press, or Yahoo’s sports journalism, there is a good chance you’ll think they were written by a person.

Wordsmith is an artificial writer. Developed by a company in North Carolina called Automated Insights, it plucks the most interesting nuggets from a dataset and uses them to structure an article (or email, or product listing). When it comes across really big news, it uses more emotive language. It varies diction and syntax to make its work more readable. Even a clumsy robot chef can have its uses, but writing for human readers must be smooth. Hooked up to a voice-recognition device such as Echo de Amazon, Wordsmith can even respond to a spoken human question – about the performance of one’s investments, say – with a thoughtfully spoken answer, announcing what’s interesting first, and leaving out what isn’t interesting at all. If you didn’t know the trick, you’d think Hal 9000 had arrived.

The trick is this: Wordsmith does the part of writing that people don’t realise is easy. Locky Stewart from Automated Insights gives me a tutorial. You write into Wordsmith a sentence such as, “New ABC figures show that the New York Inquirer’s circulation rose 3% in April.” Then you play around. La 3% has come from your data, so you select the word “rose” and write a rule, known as a “branch”, which will change the word “rose” to the phrase “shot up” if the percentage is more than 5%. Then you branch “rose” to become “fell” if the percentage is negative. If the percentage is -5% or lower, “rose” becomes “plummeted”.

Then you feed it synonyms. So “plummeted” can also be “fell sharply by”. “The Inquirer’s circulation” can be “circulation at the Inquirer”. “Shot up” can be “soared” and so on. Then you add more sentences, perhaps about online traffic, or about which days’ print copies sold best, or about comparisons year-on-year. Then you get clever. You tell Wordsmith to put the sentences with the most newsworthy information first, defined perhaps as those that feature the greatest percentage changes. Maybe you add a branch to say that a result is “the best/worst performance among the quality titles”. Hell, you can even teach it some old Fleet Street tricks, so that if circulation plummets the piece begins “Editor Charles Kane is facing fierce criticism as”, but if circulation has “shot up” this becomes “Charles Kane has silenced critics with news that”. Insert “more” or “again” or “continues” if you get the same thing two months in a row.

“The artificial intelligence is actually the human intelligence that is building the network of logic,” Stewart says, “the same network you would use when writing a story. It could have been developed 10 o 15 hace años, in code, but to make it work at this scale has only been possible lately.” Clearly it takes longer to prepare an article on Wordsmith than to write one conventionally, but once you’ve done so, the computer can publish a fresh newspaper circulation story every month, on every newspaper, within seconds of receiving the information. It can publish millions of stories in minutes – or publish only some of them, if the data doesn’t reach a given threshold of newsworthiness. Thus it becomes an automated editor, demasiado, with adjustable tastes in thoroughness, frequency and hysteria.

For Wordsmith’s task, I suggest football: it’s a field that produces a lot of data and has a readership that wants personalised articles. Guardian football writer Jacob Steinberg volunteers to take on the computer, and I provide a table of facts from the recent Premier League: last season’s league position and this season’s position at Christmas and at the end, goals scored and conceded, top scorer’s name and total, value of summer transfers and a quote from the manager.

Working solely from this data, computer and human must each write a review of the season for a given club. Steinberg chooses Leicester City on the basis that its numbers should contain a story that anyone would see. Wordsmith doesn’t need to choose. It will do all 20.

And in fact both computer and human quickly produce quite similar work:

Leicester City footballer Jamie Vardy

Both Steinberg and Wordsmith deliver dramatic first sentences. Perhaps keen to sound authentic, Automated Insights use some clever tricks to put feeling into the latter’s article, astutely guessing that Leicester were “hoping to finish in the top 10 after a 14th place finish last season”. I look through Wordsmith’s other articles and Southampton, having finished seventh last season, have “eyes on a European spot”, while Manchester City “began the season dreaming of a league title after finishing second”.

A la inversa, Steinberg digs more meaningfully into the numbers, showing that Jamie Vardy not only scored 24 metas, but that this was a higher percentage of his team’s goals than was managed by all but two other players. Knowing how Wordsmith works, por supuesto, one could easily set it up to do the same. In fact looking through it, Steinberg’s entire article could have been created by a skilled Wordsmith programmer – with the exception of one line. “It’s a magical season,” he quotes the Leicester manager as saying, before adding, “justifiably so, given that a summer expenditure of £26.7m on transfers made them the eighth lowest spenders”. That “justifiably so” shows a writer who actually understands what he is writing.

Veredicto Steinberg is a much better writer, unless you want 20 data-heavy articles in 10 acta.

The painting test

A laptop wants me to smile. “It’s in a good mood," Simon Colton dice. He knows because he’s the scientist who programmed it. We are in the Science Museum in London, where the Painting Fool, as it is called, is giving a public demonstration. It’s important that I don’t show my teeth, Colton says, because something about the light makes them look green to the Painting Fool.

From my toothless smile the laptop creates a “conception” of what it would like to paint, based on its mood. The mood comes from a “sentiment analysis” of recent Guardian articles, as it happens (on average reading the Guardian is a downer, aparentemente, apart from the stuff about gardening). Yesterday the Fool was in such a bad mood that it sent someone away unpainted; today it is feeling “positive”.

Next the Fool attempts to paint with a simulated brush and a simulated hand (actually, an image of Colton’s hand) on the screen behind me. It learned to reflect its mood from the work of Dan Ventura, another computer scientist, at Brigham Young University in Utah, who trained a neural network to recognise the emotional attributes of images by sitting thousands of people in front of tens of thousands of paintings and asking them to tag each one with whatever adjectives came to mind. The Fool now knows that bright colours reflect a good mood, and “pencils with tight hatching” create a picture that is “cold”. When it is done, it prints out a page with a typed self-critique. “Overall, this is quite a bright portrait,” it says. “That’s OK, but my style has lowered the level of bright here. So I’m a bit annoyed about that.”

Here along with us, intrigued but too busy at her easel to watch, es Sarah Jane Moon, an artist who exhibits with the Royal Society of Portrait Painters. She doesn’t want to see my teeth, either. “We paint from life,"dice, “and you can’t hold a smile for sitting upon sitting. That’s why all the traditional portraits show quite relaxed features.”

The Painting Fool is a special machine, and even slightly famous, but I can’t deny that Moon is almost all of why I’m excited to be here. The feeling of being painted by a real person, having them look at you and think about you, is exciting and flattering. Sentiment analysis and training data, por otra parte, don’t add up to anything whose view of me I care about, and the finished portraits do not change my mind. Moon’s is a lovely, real thing, which feels straight away like one person seen by another. The Fool’s three efforts have qualities I like, but mostly they look like photographs that have gone through some kind of software filter. Colton insists the Fool is here “to learn to be better” but I look and think: so what?

Painting of Leo Benedictus by Sarah Jane Moon
Leo Benedictus as seen by Sarah Jane Moon…
Painting of Leo Benedictus by the Painting Fool computer
…and as imagined by the Painting Fool laptop. Fotografía: Murray Ballard

Then I think some more. For one thing, it turns out that art is more mechanical than I’d realised. “I try to look at Leo as an abstract set of shapes, formas, colores, tones,” Moon tells Colton, “to get away from the fact that that’s a nose. Because when you start to do that, you get caught up in what you Creo que looks like a nose.”

“What the software does is break it down into colour regions,” Colton says.

"Sí, exactamente,” Moon agrees. “I think that’s what the best painters do. It’s transcribing.” Afterwards she tells me she felt a kind of “kinship” with the software as they worked side by side.

Más importante, I realise that what matters isn’t how the machine paints; it’s how I see. Moon I understand, Yo creo que. She’s a person and I know how that feels, so I care about her picture. But what does it feel like to be the Painting Fool? Is that what its portraits are trying to tell me?

Veredicto Moon’s painting is far richer; the Fool is still learning and has centuries of practice to go.

The translation test

Google Translate was the first piece of proper science fiction to come true, y it’s already a decade old. In many ways it typifies where AI has got to. Useful, Por supuesto; impressive, without question; but still clunky as hell, despite big improvements.

If you haven’t used it, it works like this: enter text or web links in any of 103 supported languages and you get a rough translation seconds later in any of the others. The app on your phone will transcribe what you say and then speak it back, traducido (32 languages supported); it can replace the text of a foreign language sign or menu wherever you point the camera. No explanation is needed of how cool that is (and it’s free).

Globally, half a billion people use Google Translate each month, mostly those who don’t speak English (que es 80% of people) but who want to understand the internet (que es 50% Inglés). “Most of our growth, and actually most of our traffic, comes from developing or emerging markets such as Brazil, Indonesia, India, Tailandia,” says Barak Turovsky, head of product management and user experience at Google Translate. It’s surprisingly popular for dating, demasiado, he adds. “Things like ‘I love you’ and ‘You have beautiful eyes’, that’s very prevalent.”

The software has always used a form of statistical machine learning: scouring the internet for already translated text – UN declarations, EU documents – and mapping the likelihood of certain words and phrases corresponding to one another. The more data it gathers, the better it gets, but the improvement levelled off a couple of years ago. pronto, Turovsky says, they will deploy new deep learning algorithms, which will produce much more fluent translations.

Aun así, there are limits, and some seem fundamental when you talk to a human translator and realise how subtle their work is. Ros Schwartz y Anne de Freyman volunteer for this task. Both are professional French/English translators, and I need two because, in order to judge how good the translation is without being fluent in both languages, we need to translate twice – once out of English into French, once back again. Google Translate keeps no memory of the original and can do the same thing.

I choose a short passage of distinctive but not especially wild or ambiguous prose from the beginning of Herzog by Saul Bellow. Translators normally require context, so I tell Schwartz and De Freyman that it comes from a famous mid-century American novel.

Within a few days, Schwartz and De Freyman return a very smooth facsimile of the original text. Here and there some nuances have not survived, but the passage remains a pleasure to read, and the main meanings come across exactly.

Google Translate takes only a few seconds, and the result is both impressive and inadequate, weirdly good in places, in others weirdly bad – turning “he” into “it” and concocting the idea that Herzog is in love. Miraculously, it keeps “cracked” as a description of the hero. French has no word that combines the sense of “broken” and “mad” that cracked coveys in English, so De Freyman makes it “cinglé”, which comes back from Schwartz as “crazy”.

“Google Translate would look at statistical probability and say, what does ‘cracked’ mean?” Turovsky explains. “And statistically, it will try to decide whether it means ‘cracked’ or ‘crazy’ or whatever. Ese, for a machine, is a non-trivial task.” Nor is it simple for a human, even though we find it easy. You’d have to ask whether Bellow could have meant that Herzog was “cracked” as in physically fractured. Then you’d have to assume not, because human bodies don’t generally do that. So you’d wonder what he did mean and assume instead, if you were not already familiar with the usage, that he must mean “crazy”, because you understand the rest of what you’ve read. But to do all this, wouldn’t Google Translate have to be pretty much conscious, I ask? Turovsky laughs. “I don’t think I’m qualified to answer that question.”

Veredicto Some bullseyes and howlers from Google Translate, while Schwartz and De Freyman are fluent and exact.

guardian.co.uk © Guardian News & Media Ltd. 2010

Artículos relacionados