Spam filters and network firewalls

improved significantly when they began to rely on statistical models, called Bayesian networks that are built by machine-learning algorithms. The user gives the algorithm many examples of desirable messages and also some counterexamples of undesired traffic. “The software identifies all the variables that influence

the property that you are interested in [for example, not spam], then searches over all feasible relationships among those variables to find the model that is most predictive,” Horvitz explains.

Bayesian networks can be eerily accurate. “They use probabilities, so they are wise in the sense that they know that they can’t know everything,” Horvitz elaborates. “That allows them to capture subtle behaviors that would require thousands of strict rules.” In January he

plans to present the results of a field trial of a model trained on 559 past appointments taken from a manager’s datebook. When challenged with 100 calendar entries it had never seen, the model correctly predicted whether the manager

would attend the meeting 92 percent of the time. And in four out of every five cases, the model matched the manager’s own estimate of the cost of interruption during the meeting.

That sounds impressive, but some experts in the field remain skeptical. Users may have a very low tolerance for a system that erroneously suppresses one out of every 10 important calls. “The more ‘attentive’ things become, the more unpredictable they are,” warns Ben Shneiderman of the University of Maryland. “We have a history in this community of creating ‘smart’ devices that people don’t use because they can’t understand how they operate.”

Indeed, Vertegaal reflects, “artificial intelligence couldn’t deliver the personal secretary, because it was too complicated.” Nevertheless, he adds, “I’m pretty sure we can deliver a receptionist.”

That would be welcome, but will considerate computing really reduce interruptions and boost productivity? At least for certain specialized tasks, the answer is: unquestionably.

Consider Lockheed Martin’s HAIL-SS (Human Alerting and Interruption Logistics-Surface Ship) system. In much the way that Bestcom interposes itself between the phone system and an office worker, HAIL-SS keeps an eye on the sailors operating an Aegis naval weapons system and mediates the many alerts

that Aegis produces. In combat simulations, HAIL-SS cut the number of interruptions by 50 to 80 percent, allowing sailors to handle critical alerts up to twice as quickly. The software lowered the perceived difficulty and stressfulness

of the job by one quarter. The U.S. Navy now plans to deploy HAIL-SS throughout the fleet.

No comparable studies have yet been done in the office environment, however. Even with Bestcom diverting callers to voice mail and squelching e-mail alerts, Horvitz was interrupted 14 times in the course of our five-hour interview. Two fire alarms, a FedEx deliveryman and

numerous colleagues poking their head into the office were merely examples of a large class of disruptions that will never disappear, because they benefit the interrupter.

Vertegaal is optimistic nonetheless. “By opening up these new sources of information about how available someone is, people will naturally adapt and use them to apply existing social rules of etiquette,” he predicts. “So just by virtue of letting people know when you’re busy,

you’ll get fewer interruptions.”

W. Wayt Gibbs is senior writer.

Part1(up to ‘Horvits himself’)

Task1. Find words or expressions meaning the following.

1. a raised platform at the end of a hall , for speakers or important people;

2. similar to;

3. a small mistake in words or behavior;

4. to laugh quietly;

5. to interrupt a chain of reflections;

6. to become attuned to, to get used to;

7. steady and persistent;

8. not aware of;

9. very foolish, absurd;

10. an overwhelming ,concentrated outpouring;

11. to bear calmly and patiently, polite and respectful;

12. a stupid and clumsy person;

13. to draw a conclusion;

Task2. Find sentences in support or against the following.

1.When people are disturbed their performance falls dramatically.

2.A considerate computer may prove to be the Big Brother.

3.With a minor change of software we can make our gadgets more considerate.

4.The IBM research proved that most people would not like to be disturbed during their work.

5.The research showed that the system accuracy depended on the subjects.

Task3. Answer the following questions.

1.When his notebook interrupted him during a conference Normann was a)annoyed; b)angry; c)confused; d)wished he wasn‘t there; e)was puzzled.

2. How did the audience react?

3. What would you do in the situation?

4. What do numerous interruptions result in?

5. Explain the sentence ‘It seems to add up to a feeling of frustration.‘

6. What should be done to prevent numerous gadgets from behaving like idiots?

7. .Explalin the sentence ‘They videotaped the subjects and had them rate their interruptability.‘

8 .Can we say that Horvits came to the same conclusion in his research as Fogarty‘s team did?

Part2(from ‘Horvits himself’ up to the end)

Task1Find words or expressions meaning the following.

1. sth that worries or upsets;

2 likely;

3. strangely, frighteningly;

4. using;

5. to want to do sth;.

6. avoiding work;

7. to get rid of as useless;

8. commonplace;

9. to send another way;

10. to discover or guess as if by magic;

11. to silence

Task2Find words or expressions meaning the following.

1.If. such a system is installed , some employees may be accused of being idle.

2. People would feel apprehensive if they had to reveal their personal information to some software they have no authority over.

3.Horvits and his team are oblivious to possible dangers of such software.

4.Horvits is annoyed by the fact that a smart computer cannot recognize his presence while a dumb public convenience can.

5.Computers which can react to your look seem like a miracle.

6. To become practical such systems have to be based on AI.

7. A good example of artificial intelligence model is Bayesian networks.

8. The smarter the systems are the harder it is to say how they will behave.

9.There is no doubt that considerate systems will improve our performance in every sphere of office work. As there will be fewer disruptions.

Date: 2015-01-29; view: 848