rays,â said Martin.
âIt will read them,â said Michaels, âbut there could still be bugs in the program. What you have to do now is run the program against as many skull X rays as you can find that you have read in the past. If there are problems, I think they will be in the area of false negatives. Meaning the program will say the X ray is normal when pathology is really present.â
âThatâs the same problem with radiologists,â said Philips.
âWell, I think weâll be able to eliminate that in the program,â said Michaels. âItâs going to be up to you. Now to work this thing, first turn it on. I think even a doctor of medicine will be able to do that.â
âWithout doubt,â said Philips, âbut weâll need a Ph.D. to plug it in.â
âVery good,â Michaels laughed. âYour humor is improving. Once the unit is plugged in and turned on, you insert the cassette program into the central unit. The output printer will then inform you when to insert the X-ray film into the laser scanner.â
âWhat about the orientation of the film?â asked Philips.
âDoesnât matter, except the emulsion side has to be down.â
âOkay,â said Philips, rubbing his hands together and eyeing the unit like a proud parent. âI still canât believe it.â
âI canât, either,â said Michaels. âWho would have guessed four years ago that we could have made this kind of progress? I can still remember the day you arrived unannounced in the Department of Computer Science, plaintively asking if anyone was interested in pattern recognition.â
âIt was just pure luck that I bumped into you,â returned Philips. âAt the time I thought you were one of the undergraduates. I didnât even know what the Division of Artificial Intelligence was.â
âLuck plays a role in every scientific breakthrough,â agreed Michaels. âBut after the luck, thereâs lots of hard work, like whatâs facing you. Remember the more skull films you run with the program, the betterit will be, not only to debug the program, but also because the program is heuristic.â
âLetâs not pull any big words on me,â said Philips. âWhat do you mean âheuristicâ?â
âSo you donât like some of your own medicine,â laughed Michaels. âI never thought Iâd hear a doctor complain about incomprehensible words. A heuristic program is one that is capable of learning.â
âYou mean this thing will get smarter?â
âYou got it,â said Michaels, moving toward the door. âBut itâs up to you now. And, remember, the same format will be applicable to other areas of radiology. So in your spare time, as if youâre going to have any, start the flow sheet for reading cerebral angiograms. Iâll talk to you later.â
Closing the door behind Michaels, Philips went over to the worktable and eyed the X-ray-reading apparatus. He was eager to begin to work with it immediately, but he knew the burden of his daily routine proscribed it. As if in confirmation, Helen walked in with a pile of correspondence, telephone messages, and the cheerful news that the X-ray machine in one of the cerebral angiography rooms wasnât functioning properly. Reluctantly Philips turned his back on the new machine.
4
âLisa Marino?â asked a voice, causing Lisa to open her eyes. Leaning over her was a nurse named Carol Bigelow, whose dark brown eyes were the only portion of her face visible. A flower-print hat contained her hair. Her nose and mouth were covered by a surgical mask.
Lisa felt her arm lifted and rotated so the nurse could read her identification bracelet. The arm was replaced and patted. âAre you ready for us to fix you up, Lisa Marino?â asked Carol, releasing the brake mechanism on the gurney with her foot, and pulling the