[A] The next logical step is to automate the physical process of cutting
the diamond, and Dr Holden seems to have found the ideal partner. He is talking
to Calibrated Diamonds. a company based in Johannesburg, South Africa. about
combining his optimisation techniques with an advanced laser-cutting system.
Traditionally, diamonds are cut and polished using other diamonds. But m recent
years, lasers have been introduced to make rough cuts and m carry, out
"bruting", the bevelling process used to give diamonds their
characteristic sharp-edged shapes. John Bond. the founder of Calibrated
Diamonds, says his laser-cutting method can make much more precise cuts and can
even polish diamonds, though he is reluctant to explain how it works. He
believes that combining his laser-cutting with Dr Holden’s software, and
automating the whole process, could both reduce waste dramatically and cut the
turnaround time from months to days. "Currently, people are losing up to 70% of
the diamond." he says. [B] For over 600 years lapidarists, or stone-cutters,
have been using essentially the same techniques to cut diamonds and determine
their value, says Dr Holden. There is a great need for automation, he says—a
sentiment echoed within the industry. The Gemological Institute of America, for
example, is devising software to enable retailers and consumers to compare
diamonds of different cuts by grading the cut automatically. [C] In actual
use, the system is fed models of gemstones, which are produced by scanning the.
stones using a desktop X-ray tomography machine. When a stone is identified as a
borderline case between two grades, the system uses an optimisation technique,
called a genetic algorithm, to explore the different ways in which the stone
could be cut to maximise its value. Sometimes bigger is not necessarily better:
removing Imperfections, known as inclusions, may reduce the size and caratage of
a stone, but could also elevate it to a more valuable grade. Tests of iGem
showed that it could increase the value of a rough stone by as much as
23%. [D] Even a small improvement can yield a significant increase in value,
says Dr Holden. who specialises in applying technology to improve
decision-making in business. Together With his colleague Matee Serearuno. he has
developed an optimisation system called iGem. Besides automatically working out
the grade of a rough diamond, it also suggests how best to cat it in order to
maximise the value of the resulting stones. [E] Mr Bond hopes that this
combination of technologies will help his home continent to benefit more from
its natural wealth. More than 60% of the world’s rough diamonds come from
Africa, he says, but a lack of expertise and relatively high labour costs means
that the stones are usually shipped overseas, to countries such as India, for
assessment and cutting. He hopes to have his first factory up and running later
this year. [F] The system uses a set of roles, distilled from the judgments
of four diamond experts, m determine the clarity, and hence the grade, of each
stone. Each expert was asked to classify 503 different "virtual
stones"--computer models of stones containing different Wives of flaws. The
experts’ verdicts were then boiled down into a set of rules, so that when a new
gemstone is presented to the system, it can determine how the experts would
probably have graded it. Data from more experts could have been used, but four
proved to be enough to produce a robust and accurate system, says Dr
Holden. [G] While a diamond may be for ever, its value is far from set in
stone. It depends on the four Cs: carat, cut, colour and clarity. But while the
first three can be measured objectively, assessing a diamond’s clarity involves
a certain amount of subjectivity and can leave experts disagreeing about the
grade—and hence the value—of a stone. Now researchers at Cambridge University’s
Institute for Manufacturing claim m have devised a way to make the grading of
diamonds and other precious stones more consistent. Tony Holden, the project’s
leader, says automating this process could do more than just load m more
accurate valuations. It could also make rough stones more valuable, by reducing
the amount of waste during cutting. Order: