Page 138 - Fortune-November 01, 2018

+ manuf ac turing

A NEW EDGE FOR …AND FOR THE
PRO INVESTORS… AMATEURS TOO

IN THE WORLD OF FINANCE, “ROBO-ADVISER” SER-
there’s been such an explo- vices, offered by startups
sion of data collected over including Betterment and
the past decade that even Robinhood and traditional
those twentysomething discount brokerages like
analysts working around Charles Schwab, are
the clock don’t stand a already using A.I. to serve
chance of being able to the investing masses.
process it all. But machines Their low-fee investment
might. Bloomberg, FactSet tools rely on algorithms to
Research Systems, and determine how your assets
Thomson Reuters have all should be split among
developed an array of data- stocks, bonds, and other
science tools and tech- assets, based on your needs
niques—including machine and your stomach for risk.
learning, deep learning, and Their A.I. can automatically
natural language pro- rebalance your portfolio; it
cessing (NLP)—to quickly can also nudge a (non-
unearth valuable insights robotic) adviser to call you
for thousands of financial when the algorithms pre-
professionals. dict you need help with tax
Bloomberg was a pioneer strategy or estate planning.
of sentiment analysis (an The next frontier: A.I.
example of NLP), which it smart enough to help savers DESIGNING
began developing around make good decisions about
a decade ago, in which long-term, buy-and-hold
machine-learning tech- investments. Bank of MORE EFFICIENTLY
niques are used to flag a America Merrill Lynch and
news story or tweet as being Morgan Stanley are among
relevant to a stock and as- the bigger players in an SURE, COMPUTER ALGORITHMS ARE TAKING
sign a sentiment score. A.I. emerging discipline known
is also spreading to wealth (awkwardly) as quanta- over tech and science and medicine…but
management. Investment mental analysis. They aim the creatives are still safe, right? Not
groups have more than to combine the quantitative exactly. A new program from software
quadrupled their number of processing for which basic developer Autodesk called Dreamcatcher
“alternative data” analysts A.I. is best suited (basically, (rendering above) can use A.I. techniques
over the past five years, as the capacity to spot pat- to assist human designers as they go
asset managers scramble terns in gargantuan loads about their creative tasks. Already in use
to unlock the potential of of data) with additional by companies including Airbus, Under
trading signals contained in algorithms trained in the Armour, and Stanley Black & Decker, the
website scrapes, language sophisticated analysis as-
analysis, credit card pur- sociated with supersmart software is an example of the burgeon-
chases, and satellite data. humans—assessing, say, ing field of generative design. A designer
Firms reported to be using the growth potential of an inputs requirements, limitations, and
A.I. for investment research industry or the strategic other qualities into the program—even the
RENDERING: COUR T ESY OF AU T ODESK T. Rowe Price. BlackRock, learning could eventu- of options. As the human designer win-
total cost of materials. The software then
acumen of a company’s
include BlackRock, Fidelity,
produces hundreds or even thousands
management. Machine
Invesco, Schroders, and
the world’s largest asset
ally enable a quantamental
nows the choices, the software susses out
system to learn from its
manager, has been a fore-
preferences and helps iterate even better
mistakes. The ultimate
runner in adopting A.I. and is
options. Airplane manufacturer Airbus
goal: Warren Buffett–like
setting up a BlackRock Lab
used the software to redesign an interior
for Artificial Intelligence.
stock-picking wisdom at
—Scott DeCarlo
low prices—and perhaps a
design that was 66 pounds, or 45%, lighter
name catchier than “quan-
than the previous setup. —A.P.
tamental.” —Matt Heimer partition in the A320 and came up with a
101
FEEDBACK LETTERS@FORTUNE.COM FO R T U N E. CO M // N O V. 1 . 1 8