The New York Times International - 02.08.2019

T HE NEW YORK TIMES INTERNATIONAL EDITION FRIDAY, AUGUST 2, 2019 | 7

world

As corporate giants like Ford and Gen-

eral Motors struggle to get their self-

driving cars on the road, a team of re-

searchers in China is rethinking autono-

mous transportation using a souped-up

bicycle.

This bike can roll over a bump on its

own, staying perfectly upright. When

the man walking just behind it says

“left,” it turns left, angling back in the di-

rection it came.

It also has eyes: It can follow some-

one jogging several yards ahead, turn-

ing each time the person turns. And if it

encounters an obstacle, it can swerve to

the side, keeping its balance and con-

tinuing its pursuit.

It is not the first-ever autonomous bi-

cycle (Cornell University has a project

underway) or, probably, the future of

transportation, although it could find a

niche in a future world swarming with

package-delivery vehicles, drones and

robots. Nonetheless, the Chinese re-

searchers who built the bike believe it

demonstrates the future of computer

hardware. It navigates the world with

help from what is called a neuromorphic

chip, modeled after the human brain.

In a paper published on Wednesday in

Nature, the researchers described how

such a chip could help machines re-

spond to voice commands, recognize the

surrounding world, avoid obstacles and

maintain balance. The researchers also

provided a video showing these skills at

work on a motorized bicycle.

The short video did not show the limi-

tations of the bicycle (which presum-

ably tips over occasionally), and even

the researchers who built the bike ad-

mitted in an email to The New York

Times that the skills on display could be

duplicated with existing computer hard-

ware. But in handling all these skills

with a neuromorphic processor, the

project highlighted the wider effort to

achieve new levels of artificial intelli-

gence with novel kinds of chips.

This effort involves myriad start-up

companies and academic labs, as well as

big tech companies like Google, Intel

and IBM. And as the Nature paper dem-

onstrates, the movement is gaining sig-

nificant momentum in China, a country

with little experience designing its own

computer processors, but which has in-

vested heavily in the idea of an “A.I.

chip.”

The hope is that such chips will even-

tually allow machines to navigate the

world with an autonomy not possible to-

day. Existing robots can learn to open a

door or toss a Ping-Pong ball into a plas-

tic bin, but the training takes hours or

even days of trial and error. Even then,

the skills are viable only in very particu-

lar situations. With help from neuromor-

phic chips and other new processors,

machines could learn more complex

tasks more efficiently, and be more

adaptable in executing them.

“That is where we see the big prom-

ise,” said Mike Davies, who oversees In-

tel’s efforts to build neuromorphic chips.

Over the past decade, the develop-

ment of artificial intelligence has accel-

erated thanks to what are called neural

networks: complex mathematical sys-

tems that can learn tasks by analyzing

vast amounts of data. By looking at

thousands of cat photos, for instance, a

neural network can learn to recognize a

cat.

This is the technology that recognizes

faces in the photos you post to Face-

book, identifies the commands you bark

into your smartphone and translates be-

tween languages on internet services

like Microsoft Skype. It is also hastening

the advance of autonomous robots, in-

cluding self-driving cars. But it faces

significant limitations.

A neural network doesn’t really learn

on the fly. Engineers train a neural net-

work for a particular task before send-

ing it out into the real world, and it can’t

learn without enormous numbers of ex-

amples. OpenAI, an artificial intelli-

gence lab in San Francisco, recently

built a system that could beat the

world’s best players at a complex video

game called Dota 2. But the system first

spent months playing the game against

itself, burning through millions of dol-

lars’ worth of computing power.

Researchers aim to build systems

that can learn skills in a manner similar

to the way people do. And that could re-

quire new kinds of computer hardware.

Dozens of companies and academic labs

are now developing chips specifically

for training and operating A.I. systems.

The most ambitious projects are the

neuromorphic processors, including the

Tianjic chip under development at

Tsinghua University in China. Such

chips are designed to imitate the net-

work of neurons in the brain, not unlike a

neural network but with even greater fi-

delity, at least in theory.

Neuromorphic chips typically include

hundreds of thousands of faux neurons,

and rather than just processing ones

and zeros, these neurons operate by

trading tiny bursts of electrical signals,

“firing” or “spiking” only when input

signals reach critical thresholds, as bio-

logical neurons do.

“This is about trying to bridge and

unify computer science and neurosci-

ence,” said Gordon Wilson, the chief ex-

ecutive of Rain Neuromorphics, a start-

up company that is developing a neuro-

morphic chip.

Neuromorphic chips are by no means

a recreation of the brain. In so many re-

spects, the workings of the brain remain

a mystery. But the hope for such chips is

that, by operating a bit more like the

brain, they can help A.I. systems learn

skills and execute tasks more efficiently.

Because each faux neuron fires only

on demand rather than continuously,

neuromorphic chips consume less ener-

gy than traditional processors. And be-

cause they are designed to process in-

formation in short bursts, some re-

searchers believe they could lead to sys-

tems that learn on the fly, from much

smaller amounts of data.

In the Chinese video, the bicycle is not

learning; it is merely executing soft-

ware that had been trained to handle

specific tasks, including recognizing

spoken words and avoiding obstacles.

But it is executing the software in an effi-

cient way, which is important to vehicles

that run on battery power. Researchers

believe they can eventually merge the

training process and the in-the-moment

execution, so that a bicycle could learn

as it goes, from just a few moments of

experience.

The rub is that building the right hard-

ware may require at least several more

years of research. “We are still in the

trial and error stage,” said Georgios Di-

mou, who previously worked on Intel’s

neuromorphic project.

The Chinese researchers believe that

time will bring far more than just auton-

omous bicycles. Their paper paints the

Tianjic chip as a step toward “artificial

general intelligence,” a machine that

can do anything you and your brain can

do. But that is merely the promise du

jour.

And now, a bicycle built for none

BY CADE METZ

A self-driving bike shows how a

computer might think for itself.

Misplacing your car keys is worrisome,

but it’s nothing like the free-fall panic of

losing your phone. Hollywood could

make a horror movie about somebody

just looking for an iPhone XR and going

slowly insane.

The telephone began to pervade our

lives at the end of the 19th century, and

then — as you can see in these photos

from The New York Times’s archives —

it becameour lives. Cellphones were a

significant inflection point. They made it

possible for us to be available at virtual-

ly any moment, which was so extraordi-

nary that most of us tacitly accepted

that we should be available at virtually

any moment.

History’s first call on a hand-held

wireless phone was made on April 3,

1973, by a Motorola executive named

Martin Cooper. Mr. Cooper had devel-

oped the phone himself and, having a

cheeky streak, decided to step out onto

Sixth Avenue, in Midtown Manhattan,

and call his rival at Bell Laboratories to

gloat a little. Can you hear me now?

Told recently that his call had been a

great P.R. stunt, Mr. Cooper, who turned

90 last year, said: “Remember, this was

the first public call ever made, and I only

cared about one thing: Was the phone

going to work? This thing was a hand-

made prototype — thousands of parts

carefully wired together by an engineer,

not a production guy — and there were

only two in existence.”

In one photo, Mr. Cooper’s mentor,

John F. Mitchell, shows off the brand-

new DynaTAC. “The Brick,” as it be-

came known, went on sale in 1983 for

$3,995 (more than $10,000 in today’s dol-

lars). It weighed about two pounds and

took 10 hours to charge. You could talk

for only 20 minutes before it went dead.

In 1987, however, the DynaTAC became

a symbol for the liberating power of

money when Michael Douglas walked

along a Hamptons beach in his bathrobe

in “Wall Street” and stoked Charlie

Sheen’s greed: “I’m gonna make you

rich, Bud Fox. Yeah, rich enough you can

afford a girl like Darien. This is your

wake-up call, pal. Go to work.”

What did Mr. Cooper say back in the

day when critics predicted that avail-

ability would be a curse?

“My first response was, ‘Well, we do

have an on/off switch on every phone.

You don’t haveto answer it,’” he said.

“But my more profound response is that

the cellphone is still really new in soci-

ety. I have an overweening belief in the

intelligence of people. We are going to

figure out how to communicate with

other people and live complete and

wholesome lives. Yes, I’m concerned,

but if you think about all the other dis-

tractions — somehow humanity has

managed to survive. When the TV first

came out, I spent hours watching

wrestling.”

Mr. Cooper believes the good that cell

technology can do — spark productivity,

enfranchise the poor — far outstrips any

downsides. He always buys the latest

iPhone, though he gives the phones

mixed reviews: extraordinary in terms

of tech, but needlessly complicated from

a consumer point of view.

Reminded that even his own Dy-

naTAC “Brick” had drawbacks, Mr. Coo-

per laughed. He said he always tells peo-

ple it was never a problem that the bat-

tery lasted only around 20 minutes —

the phone was so heavy that nobody

could hold it up longer than that anyway.

How phones made the world your office

PAST TENSE

Device that long pervaded

our lives became our lives

by going mobile in 1973

BY JEFF GILES

“Remember, this was the first

public call ever made, and I only

cared about one thing: Was the

phone going to work?”

THE NEW YORK TIMES WALTER CARLSON/THE NEW YORK TIMES

THE NEW YORK TIMES

BETTMANN/GETTY IMAGES CARL T. GOSSETT JR./THE NEW YORK TIMES BETTMANN/GETTY IMAGES

Clockwise from top left: Operators at

Tokyo’s international telephone exchange,

around 1950; placing a call from a vehicle

on a road on Long Island, N.Y.; Patricia

Kossman, a Doubleday editor, using a

cordless phone in 1983; Gene Johnson

using an “attaché phone” in 1969; demon-

strating a new switchboard in 1954; John

F. Mitchell showing off Motorola’s Dy-

naTAC, the first hand-held wireless

phone, in 1973; Walter S. Gifford placing

the first trans-Atlantic phone call, from

New York to London, in 1927.

DON HOGAN CHARLES/THE NEW YORK TIMES

РЕЕhelp from what is called a neuromorphichelp from what is called a neuromorphic

Л

hardware. It navigates the world with

Л

hardware. It navigates the world with

help from what is called a neuromorphichelp from what is called a neuromorphicЛ

hardware. It navigates the world withhardware. It navigates the world withИИЗ

demonstrates the future of computer

З

demonstrates the future of computer

hardware. It navigates the world withhardware. It navigates the world withЗ

demonstrates the future of computerdemonstrates the future of computerdemonstrates the future of computerdemonstrates the future of computerППОО

Д

searchers who built the bike believe it

Д

searchers who built the bike believe it

demonstrates the future of computer

Д

demonstrates the future of computer

searchers who built the bike believe itsearchers who built the bike believe itГГО

robots. Nonetheless, the Chinese re-

О

robots. Nonetheless, the Chinese re-

searchers who built the bike believe itsearchers who built the bike believe itО

robots. Nonetheless, the Chinese re-robots. Nonetheless, the Chinese re-ТТО

package-delivery vehicles, drones and

О

package-delivery vehicles, drones and

robots. Nonetheless, the Chinese re-robots. Nonetheless, the Chinese re-О

package-delivery vehicles, drones andpackage-delivery vehicles, drones andВВ

robots. Nonetheless, the Chinese re-

В

robots. Nonetheless, the Chinese re-

package-delivery vehicles, drones andpackage-delivery vehicles, drones andИИЛ

niche in a future world swarming with

Л

niche in a future world swarming with

package-delivery vehicles, drones andpackage-delivery vehicles, drones andЛ

niche in a future world swarming withniche in a future world swarming withАГ

transportation, although it could find a

Г

transportation, although it could find a

niche in a future world swarming withniche in a future world swarming withГ

transportation, although it could find atransportation, although it could find aРРУ

underway) or, probably, the future of

У

underway) or, probably, the future of

transportation, although it could find atransportation, although it could find aУП

underway) or, probably, the future of

П

underway) or, probably, the future of

transportation, although it could find atransportation, although it could find aП

underway) or, probably, the future ofunderway) or, probably, the future ofППА

cycle (Cornell University has a project

А

cycle (Cornell University has a project

underway) or, probably, the future ofunderway) or, probably, the future ofА

"What's

searchers who built the bike believe it

"What's

searchers who built the bike believe it

demonstrates the future of computer

"What's

demonstrates the future of computer

hardware. It navigates the world withhardware. It navigates the world with"What's

News"

package-delivery vehicles, drones and

News"

package-delivery vehicles, drones and

robots. Nonetheless, the Chinese re-

News"

robots. Nonetheless, the Chinese re-

searchers who built the bike believe it

News"

searchers who built the bike believe it

VK.COM/WSNWS

package-delivery vehicles, drones and

VK.COM/WSNWS

package-delivery vehicles, drones and

robots. Nonetheless, the Chinese re-

VK.COM/WSNWS

robots. Nonetheless, the Chinese re-

searchers who built the bike believe it

VK.COM/WSNWS

searchers who built the bike believe it

demonstrates the future of computer

VK.COM/WSNWS

demonstrates the future of computer

hardware. It navigates the world with

VK.COM/WSNWS

hardware. It navigates the world with

help from what is called a neuromorphichelp from what is called a neuromorphicVK.COM/WSNWS