A.I. Is Harder Than You Think

The field of artificial intelligence doesn’t lack for ambition. In January, Google’s chief executive, Sundar Pichai, claimed in an interview that A.I. “is more profound than, I dunno, electricity or fire.”

Day-to-day developments, though, are more mundane. Last week, Mr. Pichai stood onstage in front of a cheering audience and proudly showed a video in which a new Google program, Google Duplex, made a phone call and scheduled a hair salon appointment. The program performed those tasks well enough that a human at the other end of the call didn’t suspect she was talking to a computer.

Assuming the demonstration is legitimate, that’s an impressive (if somewhat creepy) accomplishment. But Google Duplex is not the advance toward meaningful A.I. that many people seem to think.

If you read Google’s public statement about Google Duplex, you’ll discover that the initial scope of the project is surprisingly limited. It encompasses just three tasks: helping users “make restaurant reservations, schedule hair salon appointments, and get holiday hours.”

Schedule hair salon appointments? The dream of artificial intelligence was supposed to be grander than this — to help revolutionize medicine, say, or to produce trustworthy robot helpers for the home.

The reason Google Duplex is so narrow in scope isn’t that it represents a small but important first step toward such goals. The reason is that the field of A.I. doesn’t yet have a clue how to do any better.

As Google concedes, the trick to making Google Duplex work was to limit it to “closed domains,” or highly constrained types of data (like conversations about making hair salon appointments), “which are narrow enough to explore extensively.” Google Duplex can have a human-sounding conversation only “after being deeply trained in such domains.” Open-ended conversation on a wide range of topics is nowhere in sight.

The limitations of Google Duplex are not just a result of its being announced prematurely and with too much fanfare; they are also a vivid reminder that genuine A.I. is far beyond the field’s current capabilities, even at a company with perhaps the largest collection of A.I. researchers in the world, vast amounts of computing power and enormous quantities of data.

The crux of the problem is that the field of artificial intelligence has not come to grips with the infinite complexity of language. Just as you can make infinitely many arithmetic equations by combining a few mathematical symbols and following a small set of rules, you can make infinitely many sentences by combining a modest set of words and a modest set of rules. A genuine, human-level A.I. will need to be able to cope with all of those possible sentences, not just a small fragment of them.

The narrower the scope of a conversation, the easier it is to have. If your interlocutor is more or less following a script, it is not hard to build a computer program that, with the help of simple phrase-book-like templates, can recognize a few variations on a theme. (“What time does your establishment close?” “I would like a reservation for four people at 7 p.m.”) But mastering a Berlitz phrase book doesn’t make you a fluent speaker of a foreign language. Sooner or later the non sequiturs start flowing.

Even in a closed domain like restaurant reservations, unusual circumstances are bound to come up. (“Unfortunately, we are redecorating the restaurant that week.”) A good computer programmer can dodge many of these bullets by inducing an interlocutor to rephrase. (“I’m sorry, did you say you were closed that week?”) In short stylized conversations, that may suffice. But in open-ended conversations about complex issues, such hedges will eventually get irritating, if not outright baffling.

To be fair, Google Duplex doesn’t literally use phrase-book-like templates. It uses “machine learning” techniques to extract a range of possible phrases drawn from an enormous data set of recordings of human conversations. But the basic problem remains the same: No matter how much data you have and how many patterns you discern, your data will never match the creativity of human beings or the fluidity of the real world. The universe of possible sentences is too complex. There is no end to the variety of life — or to the ways in which we can talk about that variety.

So what should the field of artificial intelligence do instead? Once upon a time, before the fashionable rise of machine learning and “big data,” A.I. researchers tried to understand how complex knowledge could be encoded and processed in computers. This project, known as knowledge engineering, aimed not to create programs that would detect statistical patterns in huge data sets but to formalize, in a system of rules, the fundamental elements of human understanding, so that those rules could be applied in computer programs. Rather than merely imitating the results of our thinking, machines would actually share some of our core cognitive abilities.

That job proved difficult and was never finished. But “difficult and unfinished” doesn’t mean misguided. A.I. researchers need to return to that project sooner rather than later, ideally enlisting the help of cognitive psychologists who study the question of how human cognition manages to be endlessly flexible.

Today’s dominant approach to A.I. has not worked out. Yes, some remarkable applications have been built from it, including Google Translate and Google Duplex. But the limitations of these applications as a form of intelligence should be a wake-up call. If machine learning and big data can’t get us any further than a restaurant reservation, even in the hands of the world’s most capable A.I. company, it is time to reconsider that strategy.

Origin : https://www.nytimes.com/2018/05/18/opinion/artificial-intelligence-challenges.html

The Ocean Is A Strange Place After Dark

Moonlight triggers the world’s biggest orgy, strange creatures emerge from the depths, and waves glow blue. Some phenomena in the ocean can only be witnessed after dark.

1. Bioluminescence makes the sea shimmer

Dinoflagellates emit blue light when they’re disturbed, like at this bay on Vaadhoo Island in the Maldives (Credit: Naturepl.com/Doug Perrine)

You may have seen the pictures.

It’s night-time in an impossibly exotic location. Waves are breaking on the beach. The water is sparkling with electric blue lights.

The internet loves an image of a magical-looking bioluminescent bay. You may also have seen travel bloggers bemoaning the real event as not quite living up the hype.

Even if the latter is true, bioluminescence (in this case usually caused by planktonic organisms called dinoflagellates) is a pretty amazing natural phenomenon.

Dinoflagellates emit blue light when disturbed, which is why they can be seen sparkling over wave crests, around boats or when a hand or paddle runs through them.

These tiny creatures are the most common source of bioluminescence at the ocean’s surface.

So-called bioluminescent bays such as in Puerto Rico and Jamaica are among the best-known places to witness the glow. However, the ephemeral phenomenon can be found throughout the ocean where there are dense gatherings of dinoflagellates.

Sometimes dinoflagellates’ population increases rapidly causing blooms, which by day are coloured a less attractive red-brown, sometimes known as red tides. And some, but not all, of these red tides are poisonous.

These tiny creatures provide the most common source of bioluminescence at the ocean’s surface (Credit: Naturepl.com/Martin Dohrn)

Even stranger and rarer than bioluminescent bays are “milky seas”, where continually glowing water stretches for as far as the eye can see.

Milky seas have only been seen a few hundred times since 1915, mainly concentrated around north-western Indian Ocean and near Java, Indonesia.

They are not caused by dinoflagellates, but are thought to be the result of “bioluminescent bacteria that have accumulated in large numbers near the surface”, explains to Dr Matt Davis, Assistant Professor of Biology, St. Cloud State University in the US, who specialises in bioluminescence.

Reports by sailors over the centuries have described milky seas as a nocturnal whitish glow like a field of snow, but scientists have had little chance to investigate the phenomenon first-hand.

In 2005, researchers analysing archived satellite images found that milky seas could be seen from space and that one satellite had captured images of a huge area of ocean that had displayed the strange glow for three consecutive nights a decade earlier.

2. Animals glow in the dark

Bobtail squid have a symbiotic relationship with bioluminescent bacteria (Credit: Naturepl.com/Jurgen Freund)

Bioluminescence, the emission of visible light by an organism as the result of a natural chemical reaction, is common among marine life such as fishes, squid and molluscs. In the deep sea most species are bioluminescent, where it is the main source of light.

In shallower waters, most bioluminescent fish display their lights at night.

“Flashlight fishes have a specialized pouch under their eye that they can rotate to expose the light emitted from these bacteria, and they use this glow at night to hunt for food and communicate,” says Dr Matt Davis.

Flashlight fishes have a specialised pouch under their eye that use to expose bioluminescent bacteria (Credit: Matt Davis)

Ponyfish emit light from the bioluminescent bacteria housed in a pouch using transparent muscular shutters, to communicate, he explains.

Camouflage, defence and predation are among the variety of reasons fishes are thought to emit light.

For example, bobtail squid have an ingenious way of using lights. These nocturnal animals have a mutually beneficial relationship with luminescing bacteria that live in a mantel cavity on its underside. At night the squid control the intensity of this light to match the moonlight, and can reduce their silhouette to camouflage themselves from predators.

3. Moonlight triggers the planet’s biggest orgy

Mass spawning on the Great Barrier Reef is one of the extraordinary examples of synchronised behaviour on Earth (Credit: Naturepl.com/Jurgen Freund)

There is nothing more romantic than a moonlit night, especially if you are a coral on the Great Barrier Reef off Australia.

One night a year in spring, the biggest orgy on earth is triggered by lunar light.

Over 130 coral species simultaneously release their eggs and sperm into the water during a window of just 30-60 minutes.

This mass spawning event might be the most extraordinary example of synchronised behaviour in the natural world.

When the gametes – eggs and sperm cells - are released they hover for a moment, forming a ghostly replica of the reef’s shape, before dispersing into an underwater blizzard as the sperm fertilise the eggs.

Dr Oren Levy, a marine biologist and ecologist and Professor of Life Sciences at Bar-Ilan University, Israel, has studied this extraordinary event.

“This is really fascinating phenomena…we know this event is going to happen a few nights after November's full moon each year, three to five [days] post full moon,” he says.

“[It is] always amazing, in particular I am so amazed how each of the coral species year after year spawn at the same hour of the night.”

He adds: ”Once it happens it is always so exciting to see how everything is becoming so live and synchronised. It is almost [a] spiritual event and you understand the power of nature in its best.”

Moonlight triggers the phenomenon by acting as a synchroniser or “alarm” probably with other environmental signals such as sunset timings, water temperature and tides to cue the time of the gamete [egg and sperm cells] release, explains Dr Levy.

He adds that corals seem to possess photoreceptors that detect the phases of the moon, which helps with the “fine tuning” of the gamete release.

4. Sharks and seals rely on celestial light

Just when you think it's safe to go into the water... great white sharks hunt at night too (Credit: Naturepl.com/Chris & Monique Fallows)

For some seals, moonlit nights spell danger.

During winter months, the 60,000 cape fur seals on Sea Island in False Bay, South Africa run the gauntlet of being picked off by great white sharks patrolling the seas when they enter and exit the water.

One study in 2016 hypothesised seals swimming at night during a full moon are at more risk of being eaten by a shark since bright moonlight silhouetting them against the surface makes them an easy meal for predators lurking below.

However, most shark attacks on seals happen just after sunrise. Researchers behind the study, which measured shark attacks at dawn, were surprised to find seals were much less likely to be predated at this time of day if there was a full moon.

The researchers theorised that lunar illumination combined with emerging sunlight may decrease the stealth ability of the sharks and that the advantage switched from sharks to seals as night turned to day.

And seals may rely on another celestial feature to navigate - the stars.

Captive harbour seals (Phoca vitulina) are able to locate a single lodestar and steer by it, researchers have shown.

During a test using a simulated night sky, seals swam towards the brightest star and could orientate themselves when the stars were swivelled around.

In the wild, seals need to navigate the open ocean to find foraging grounds that may be separated by hundreds of kilometres.

Researcher Dr Bjorn Mauck said at the time: "Seals might learn the position of the stars relative to foraging grounds during dawn and dusk when they can see both the stars and landmarks at the coast."

5. Strange animals come to the surface every night

Humboldt squid are among the most striking creatures to surface every night (Credit: Naturepl.com/Franco Banfi)

Under the cover of darkness rarely seen creatures migrate to the ocean’s surface to feed.

The Humboldt squid, also known as the jumbo squid, is one of the most eye-catching marine animals you can see lurking in surface waters.
By day the squid lurk in the deep waters of the Eastern Pacific Ocean along the deep shelf that runs off the west coast of the Americas and every night they are one of the many ocean animals to migrate upwards to find dinner.

Vertical (or diel) migration - when ocean animals swim to the surface at dusk and disappear down again at dawn – is extremely common.

“What [Humbioldt squid are] doing largely is following their main food item, which is the so-called lantern fish,” explains Professor Paul Rodhouse, an Emeritus Fellow for the British Antarctic Survey (BAS) and former head of the organisation’s biological sciences division.

In turn, lantern fish follow vertically migrating zooplankton.

Since zooplankton are depended on by so many ocean animals, “the rest of the food chain will be following on after it,” says Prof Rodhouse.

“It is a huge movement of biomass every day,” says Prof Rodhouse. “More than a thousand metres. Some of the oceanic squid probably migrate over 1000m every day.”

He adds that almost all pelagic species (animals that live in the water column not near the bottom or shore) that can swim make the journey.

Humboldt squid are among the most striking creatures to surface every night. Their ability to change colour and flash bright red when agitated has earned them the nickname “red devils”. Although much smaller than their cousin, the 13m-giant squid, they can reach a length of about 1.5m (almost 5ft). Highly aggressive predators, they capture prey with strong tentacles and suckers and tear into it with powerful beaks, and have reportedly occasionally attacked humans.

But even ferocious Humboldts are preyed upon by bigger predators such as billfish, swordfish and sharks.

“Of course what they are all doing [by being active at night] is avoiding predation by the top predators,” says Prof Rodhouse. "The big predators that are visual predators and which stay in the surface waters and see their prey.”

“So they’re all… reducing the risk of being preyed on by going down into deep, dark waters at night.”

Origin : http://www.bbc.com/earth/story/20170818-five-amazing-things-that-happen-in-the-ocean-at-night

Hydrogen Bomb vs. Atomic Bomb: What’s the Difference?

North Korea is threatening to test a hydrogen bomb over the Pacific Ocean in response to President Donald Trump ordering new sanctions on individuals, companies and banks that conduct business with the notoriously reclusive country, according to news reports.

“I think that it could be an H-bomb test at an unprecedented level, perhaps over the Pacific,” North Korea’s Foreign Minister Ri Yong Ho told reporters this week during a gathering of the United Nations General Assembly in New York City, according to CBS News. Ri added that, “it is up to our leader.”

Hydrogen bombs, or thermonuclear bombs, are more powerful than atomic or “fission” bombs. The difference between thermonuclear bombs and fission bombs begins at the atomic level. [The 10 Greatest Explosions Ever]

Fission bombs, like those used to devastate the Japanese cities of Nagasaki and Hiroshima during World War II, work by splitting the nucleus of an atom. When the neutrons, or neutral particles, of the atom’s nucleus split, some hit the nuclei of nearby atoms, splitting them, too. The result is a very explosive chain reaction. The bombs dropped on Hiroshima and Nagasaki exploded with the yield of 15 kilotons and 20 kilotons of TNT, respectively, according to the Union of Concerned Scientists.

In contrast, the first test of a thermonuclear weapon, or hydrogen bomb, in the United States in November 1952 yielded an explosion on the order of 10,000 kilotons of TNT. Thermonuclear bombs start with the same fission reaction that powers atomic bombs — but the majority of the uranium or plutonium in atomic bombs actually goes unused. In a thermonuclear bomb, an additional step means that more of the bomb’s explosive power becomes available.

First, an igniting explosion compresses a sphere of plutonium-239, the material that will then undergo fission. Inside this pit of plutonium-239 is a chamber of hydrogen gas. The high temperatures and pressures created by the plutonium-239 fission cause the hydrogen atoms to fuse. This fusion process releases neutrons, which feed back into the plutonium-239, splitting more atoms and boosting the fission chain reaction.

Governments around the world use global monitoring systems to detect nuclear tests as part of the effort to enforce the 1996 Comprehensive Test Ban Treaty (CTBT). There are 183 signatories to this treaty, but it is not in force because key nations, including the United States, did not ratify it. Since 1996, Pakistan, India and North Korea have carried out nuclear tests. Nevertheless, the treaty put in place a system of seismic monitoringthat can differentiate a nuclear explosion from an earthquake. The CTBT International Monitoring System also includes stations that detect the infrasound — sound whose frequency is too low for human ears to detect — from explosions. Eighty radionuclide monitoring stations around the globe measure atmospheric fallout, which can prove that an explosion detected by other monitoring systems was, in fact, nuclear.

Origin : https://www.livescience.com/53280-hydrogen-bomb-vs-atomic-bomb.html?utm_content=buffere6855&utm_medium=social&utm_source=twitter

Why Do All Airplane Windows Have A Tiny Hole In Them?

Looking out the window of a plane – with many hours to burn and a god-like perspective of the world – can get you pondering some of life’s big questions: Are we alone in the universe? How did it all begin? Is there purpose to our existence? Wait, what are those little holes in airplane windows?

Thanks to the curious mind of Robbie Gonzalez from io9, the latter of those questions might be answered. After finding no sturdy answer on internet forums, he decided to track down a copy of a maintenance manual for the Boeing 737 on Wikileaks and contact Marlowe Moncur, Director of Technology for GKN Aerospace, the world leader in passenger cabin window design development.

As you might have already guessed, it’s to do with regulating pressure.

He found that most cabin windows consist of outer, middle and inner panes – all of which are made of a superstrong synthetic resin. Typically, it’s the middle pane that has the mysterious little hole.

Only the outer and middle panes are actually structural, while the inner is pretty much there as failsafe and to protect the other layers. Moncur said it’s only there to maintain cabin pressure in the extremely rare event that the outer pane becomes fractured.

Cruising at 10,600 meters (35,000 feet), the pressure is around 1.5 kilograms (3.3 pounds) per square inch. This is too low for the human body to stay conscious, so the pressure is artificially maintained at around 3.5 kilograms (8 pounds) per square inch. But of course, if you increase pressure inside, the structure has to be strong to hold the difference between the external pressure and internal pressure.

The outer pane is the thickest of these and is the primary layer that bears the pressure of the cabin. According to Gonzalez, the little hole is there to act as “as a bleed valve, allowing pressure between the air in the passenger cabin and the air between the outer and middle panes to equilibrate.” Simply put, it ensures that only the strongest outer pane is bearing the pressure, leaving the middle pane available in case of an emergency.

Phew, now that problem’s solved, go work on creating lasting nuclear fusion.

Origin : http://www.iflscience.com/physics/what-little-hole-airplane-windows/