Should Computers Be Able To Make Mistakes?

Do you think that advanced computers should be able to make mistakes?  It’s an interesting question in advanced computer science, one proposed a while back.  One of the problems with modern computers, theorized Michael Rabin, is that we demand perfection from them.  However, isn’t that the point of a computer?  Let’s take a closer look at what he meant:

For computer scientist Michael Rabin, the moment of truth came about a decade ago when he realized that perhaps one reason computers cannot do all we might want is because we demand perfection. Perhaps, he reasoned, if we allowed computers to make mistakes–as humans do–we might be able to get them to do more.

The idea was disarming, but, as Rabin showed, it works. He and many other investigators have devised extremely powerful probabilistic algorithms–ways of solving problems that almost always work but have a small but definite chance of being wrong (Science, 4 June 1976, p. 989). The next step for Rabin, who has a joint position at Harvard University and at Hebrew University in Jerusalem, was to go on to the workings of the computer itself. When computers were simple autonomous units, it was relatively easy to design ways to control which part of the computer was doing what. But with the advent of computer networks and of parallel processing, the problem of keeping order in the computer becomes increasingly complex. Some practical situations, in fact, are so complicated taht they simply have no workable, deterministic solution.

Rabin says that the theme of his work is to “achieve order through disorder,” in much the same way as occurs in statistical mechanics. There, he notes, a large system, such as the molecules of a gas, behaves in an orderly fashion as a result of randomized behavior of the individual constituents. As examples of the power of this approach, he tells of probabilistic solutions to several well- known problems in computer science that had frustrated investigators who were looking for simple deterministic solutions.

The fanciful problem of the dining philosophers invented by Edwin Dijkstra of Einthoven in Holland is one of these. Says Richard C. Holt of the University of Toronto, who put a picture of the dining philosphers on the cover of his book on computer science, the problem has an intrinsic fascination of its own. “It is easy to state and you can imagine solving it. But there is no easy solution.”

The story is that a group of philosophers is sitting around a table, talking and thinking. Between every pair of philosophers is a fork and in the center of the table is a plate of spaghetti. Each philosopher needs two forks to eat spaghetti. From time to time, the philosophers become hungry and want to eat some spaghetti. The system works well if no two philosophers sitting next to each other want to eat at the same time. But suppose, Rabin says, that they all become hungry at once. Each philosopher turns to his right and picks up a fork. Then each turns to his left. All the left forks, however, are now taken. “There is a deadlock. The philosophers never get to eat,” Rabin remarks. The difficulty is caused, he says, because the individual philosophers operate sequentially but the group operates concurrently. “If philosopher a were the only one who wanted to eat and then philosopher b were the only one who wanted to eat, the system would work,” he explains.

How can a protocol be devised so that deadlock is impossible? One way, Rabin notes, is to make one of the philosophers king. He tells each of the others when to eat. But if all of the philosophers are equal, Rabin says, “there are some schedules in which the philosophers will starve.””

Kolata, Gina. “Order out of chaos in computers; computer scientists are controlling decision-making with probabilistic methods, which work in cases where determinism does not.” Science 223 (1984): 917+.

Do you think that advanced – say, quantum – computing would be better off with a orderly but somewhat random way of doing things, such as in the molecules of a gas as noted in the excerpt above.

Bill Gordon, blogger at We Hate Malware, is somewhat vocal on the issues – “I think that quantum computing is really the next step in breaking the next barrier of technology.  With quantum states, however comes a sort of unpredictable, organized chaos way of looking at the computer and its systems.”

It’s interesting to think about, but quantum computing would definitely introduce the random factor that was talked about in that article back in 1984.  I’m sure they weren’t even thinking about quantum computers back then.



Did CERN Open A “Doorway”?

Did CERN open a doorway?  There are so many conspiracy theories about this that it’s very interesting to delve into some of them.  What do you think?


The Problem With Diminishing Returns In Astronomy

Quantum_ComputerDo you think that there are diminishing returns in the field of Astronomy right now?  Do you agree there are more and more people that are turning away from participating in this exciting field because it’s just too difficult?

Here’s an interesting article from not too long ago that looks at the different ways we can view this problem:

Perhaps the most striking discovery, however, is the nondiscovery of life elsewhere in the solar system. In all of the many worlds in our solar system, we have found not a shred of evidence for extraterrestrial life, even though much of the stuff of life (organic molecules, chemicals, and water) are there. It is too early to draw conclusions. We haven’t done enough looking. But we wonder whether our earthly existence is unique, and what special responsibility is implied.

The exploration of the solar system–the lessons, the data, the information, the knowledge gained–has been an enterprise not just for a few scientists or for one nation, but for our whole planet and for all humankind. It is natural to consider it as an earthly activity.

Its conduct, as well as its motivation, has become increasingly international. The Soviets carried out landings on Venus, while the Americans made two successful ones on Mars. The United States has landed people on the Moon and returned samples; the Soviet Union landed automatic vehicles there and returned samples robotically. The U.S. explored the outer planets; Japan, Europe, and the U.S.S.R. explored Halley’s comet. Now the world has changed, and current missions are less and less national and more and more international. The next Mars landings are an international mission: Mars 94/96 will be led by Russians but with heavy involvement of Europeans. The U.S. Cassini mission to Saturn will have as its centerpiece an atmospheric probe to Titan, Huyghens, built by the Europeans. And advanced thinking about sending humans to Mars almost always has the missions being conducted internationally. The costs, the resources, the technical expertise, and the benefits are too great for one nation alone.

Space benefits are inherently global. We have mentioned space science and exploration, but we must also cite meteorology, communications, remote sensing and land-use, search-and-rescue, and navigation satellites. Their implementation and benefit are global–but their application can reach to the region, nation, community, and even individual level. The technology exists today to track individual land vehicles from space or to pinpoint the best location for a village to build a wall.

The global aspect of space has one other increasingly important application. Space, in some ways uniquely, provides an educational motivation for high achievement. The creative application of science and technology and the inspiration of space exploration have been associated with education since the beginning of the space age. This is not true just in the leading spacefaring nations. In the Planetary Society’s H. Dudley Wright International Student Contest, “Together to Mars,” thousands of entries came from around the world, and twenty winners hailed from such countries as Argentina, Malaysia, Hungary, Israel, Columbia, and Portugal. The society also has a joint program with the United Nations and the European Space Agency for workshops in developing nations whose educators are enthusiastic about space science and planetary exploration in the curriculum in these nations.

Dealing with “brain drain” is a major challenge in many countries–young people are turning away from creative enterprises of high achievement because of lack of opportunity. Some bright young people interested in science and technology go to major industrialized countries to seek opportunity, but others are “turned off” by the lack of opportunity and follow less productive paths to simply make quick money. In the past, weapons technology was an area of opportunity in the U.S. and U.S.S.R. for creative engineering. Happily, the end of the Cold War is now reducing the demand for new weapons. But the need for creative outlets for the human imagination is still there–and the opportunity for it to be exercised peacefully is now greater. Space work is not the only opportunity, but it is one of the best and certainly one of the most motivating.

Our present challenge is to work together globally on these space projects to produce the best results and employ the most talent. Three areas that might provide a focus, and which certainly have a need, provide examples of where new international efforts might be mounted. In the Planetary Society’s presentation to the Committee on Peaceful Uses of Outer Space at the United Nations, we suggested three specific examples of programs that could be emphasized:”

Friedman, Louis. “Spaceflight and global unification: the benefits of space exploration.” National Forum 72.3 (1992): 38+.

Supposedly the improvement in computing should be helping us break through barriers in technology much faster, however we think that things have stalled.  The way that we will get through this speed bump is through the advent of quantum computing.