Tools


Siemens Worldwide

Pictures of the Future

Contact

Contact

sts.components.contact.mr.placeholder Sebastian Webel
Mr. Sebastian Webel

Editor-in-Chief

Tel: +49 (89) 636-32221

Fax: +49 89 636-35292

Werner-von-Siemens-Straße 1
80333 Munich

Pictures of the Future
The Magazine for Research and Innovation
 

Artificial Intelligence

Invisible Prophet

As she waits on a treatment table to have a damaged cardiac valve repaired, the inventor of a unique artificial intelligence technlogy knows that competitors may try to kill her at any moment. Little do the criminals know, however, that once the software is triggered, its agents will scrupulously track every conceivable lead, gradually zeroing in on their target, while documenting every shred of evidence.

2035. The inventor of software that learns to detect potential threats to businesses and their owners realizes that she has become the target of an insidious plot by competitors. What the perpetrators do not realize, however, is that the software, though still in its pilot stage, has learned to predict their actions.

What do you expect them to do, Denise?” said Dr. Higgs, squinting down at me in consternation after I had explained my predicament. “I don’t know for sure,” I said. “But whatever it is, it’s likely to happen any minute now.” I was trying to relax on a treatment table in the hospital’s interventional cardiology section while a multi-modal scanner produced a 3D image set of my heart’s damaged mitral valve. On the other hand, I was also brimming with anticipation, fascinated by the idea that a recently-activated pilot version of software from my company, Prophet Analytics, was probably about to provide definitive evidence against the perpetrators of what appeared to be a fiendish plan to get rid of me. Let me see if I can explain the events that led up to my being here…

I founded Prophet Analytics only about a year ago. The idea was to create a company that would use a group of unique learning algorithms I had patented to discover threats to businesses and people and help users predict — and avoid — the outcomes of those threats. In its pilot version, the software, which was trained to identify anomalous events in my extended environment, was designed to run as a very high-security app on my communicator. The software dispatches agents through networks and the agents monitor data produced by any and all relevant sensors. Once confronted with an anomaly, Prophet formulates hypotheses and compares them with what actually happens. In the process, it learns from experience and refines its accuracy. Although the software’s development had been top secret, someone had gotten wind of its potential for putting them out of business and had seemingly decided to take it, and its inventor, out of the picture.

A Prophet agent in an AV sensor in our lobby caught a tone of voice or a facial expression that raised a red flag.

But they were late. The product was already operating on a trial — and secret — basis. And when a visiting scientist from a potential competitor casually brought up the subject of a summer picnic for our top managers and asked a close colleague if I had any allergies, a Prophet agent in an AV sensor in our lobby caught a tone of voice or a facial expression that raised a red flag. Immediately thereafter, Prophet agents discovered that someone had attempted to — and possibly managed — to gain access to the files of my childhood doctor. Knowing, as it does, that I once experienced an acute reaction to a bee sting, Prophet’s hypothesis generation engine led it to search for connections to apitoxin, and particularly melittin, the principal active component in bee venom. And sure enough a professor who had recently met with Dr. Shanti — the visiting scientist — owned a small business in nano devices used to deliver industrially-produced melittin to cancer cells.

High-Tech Trap

Then, about a month ago, while playing in a benefit women’s volleyball match, a highspeed ball managed to zip between my fingers and slammed against my chest just below the neck. The impact was so intense that it knocked me down. An on-the-spot knowledge-based ultrasound test indicated a possible rupture of one of the cord-like tendons — chordae tendineae — in my heart that hold the flaps of the mitral valve in place. Such ruptures can cause the valve to regurgitate blood, which causes weakness. In the heat of the moment, the exam was automatically sent to a nearby hospital with only minimal data security.

The next morning, while thinking over Prophet’s — until then — ambiguous results, it occurred to me that my accident and its associated security breach might work as a trap to flush out any would-be bad guys. What’s more, if indeed someone was working against me or the company, the more casually I managed my information, the less likely it would be that they would realize I was onto them. As a result, information about my ultrasound scan went back and forth on a low security level between my office and the hospital’s cardiology department — including an appointment for a follow-up test and a probable interventional procedure to clip the two leaflets of the valve together, thus restoring normal function.

Prediction Engine

By afternoon, Prophet was turning up all sorts of information. Snippets of conversation and pictures picked up by image sensors in vehicle, traffic, street, parking, private and satellite infrastructures indicated that the highly automated Prosthetic Device Production Center (PDPC) in my hospital was the focus of considerable attention.

Of particular interest was what appeared to be a chance encounter at a vehicle “KwickC” charging station between Dr. Shanti and a former Prophet Analytics supervisor, Dr. Clark Hallick during which both had entered the restroom area at the same time. In fact, I happened to remember Clark very well, as we had dated for a while way back in our pre-IPO days. But all the good looks in the world couldn’t blind me to Clark’s overbearing and selfcentered attitude. We had parted bitterly and he had left the company shortly thereafter. Now it turned out that he was leading a robotic learning optimization program at the hospital — and that air filtration sensors in the charge station lavatory had picked up molecules of something very unusual for such a venue: melittin.

Having acquired a huge amount of data regarding the potential perpetrators’ personalities, histories, and realtime activities, Prophet’s prediction engine began to zero in on probable scenarios. And the one that received the highest level of probability — with a 93 percent chance of realization — was exactly what began to unfold.

Late that evening, Hallick’s department apparently began preparations for a test to see if the hospital’s robotic vision systems and associated sensors could be hacked. According to official records, the test would be conducted two days later during the small hours of the night — in other words only shortly before my admission and treatment time.

Now, as I lay on the treatment table and Dr. Higgs examined the representation of my mitral valve on a swivel arm-based display, the outlines of an optimized prosthetic clip automatically took shape on the screen. He clicked a few virtual buttons on the screen’s interface, dispatching the dataset to the PDPC for production. Moments later a bright blue light on a nearby control panel began blinking, indicating that a custom-made prosthetic clip had arrived via pneumatic line.

“We’ll have you squared away in no time, Denise,” said Higgs reassuringly. “I’ll just give you a sedative…” He started to reach for a small, glowing dial attached to the catheter line that had been connected to my groin.

But before he could touch the dial I caught his hand. “Dr. Higgs, let me take a look at the prosthesis first,” I said. “Just open the pneumatic drawer and let me examine it with my smartphone.” The drawer opened and I pointed the phone at the tiny object.

The latest phones are designed to recognize objects — regardless of how unusual they may be — to catalogue them complete with price and location information. Specialized applications allow users to virtually “open” devices by interrogating their internet datasets to see, hear, analyze, or price-compare their inner parts, or to remotely control them. But equipped with Prophet technology, such a phone will look for anything that, based on previous knowledge, fits a predicted outcome.

A moment later, a warning appeared on the phone indicating that the PDPC had experienced a data outage prior to production of the prosthesis. And worse: the phone’s built-in laser diode spectrometer had found traces of melittin embedded in the prosthesis, indicating a potential time-release delivery mechanism.

“Good God!” Exclaimed Higgs. “With your medical history that would have…” “Exactly,” I said. “And nothing would have happened until well after my release from the hospital.”

Suspect in a Red Circle

I pointed my phone at a long, horizontal wall monitor normally used for displaying cellular, physiological and anatomical inter-relationships pertinent to complex procedures. Instantly, recognizing the emergency nature of the data, the monitor became an extension of the phone’s display. As Prophet ran through what it had learned, it reconstructed the crime it had uncovered through images and reports that snapped into focus on the wall gong back through time. First came an extreme close-up of my prosthesis and associated laser spectrometry analysis, then there was an image of a model M6 sanitation robot with a report showing that traces of melittin and Clark’s DNA had been detected on gloves it had picked up and automatically analyzed. Third was an image of Clark entering the PDPC at 4:35 this morning. And finally, as I now pressed my index finger against a Prophet icon that said “Transmit Report to Security,” there was a bird’s eye view of the hospital’s lobby with a red circle around one person — marked “Hallick C” — who was apparently heading for an exit and a blue circle around a uniformed person near the entrance that read “Security Notified.”

“Well,” I said to Higgs, “I guess that just about wraps it up.” “Guess you’re right, Denise,” he answered. “And now I’d recommend ordering a fresh prosthesis and getting on with the procedure.”

Arthur F. Pease