Home
Sometimes I think intelligence is nothing but pattern recognition. Someday in the near future a computer scientist will write code that rapidly compares and stores complex patterns. To populate the computer model with data, the programmer will let his software read the entire Internet, or as much as it can, and look for patterns. After a few days of chewing through content on the Internet, the software will appear to understand everything about humans, from our language to our history. The reality of this apparent "understanding" would be nothing more than pattern recognition.

Consider a simple example: A computer could learn by pattern recognition that humans raised in specific parts of the world usually say "bless you" to a person within earshot who has just sneezed. Then overlay the pattern that people usually whisper in a library and your artificial intelligence knows it should whisper "bless you" when someone in a library sneezes. In a more complicated scenario there might be hundreds of patterns intersecting, but a computer could easily contrast and compare them.

I could be wrong, but I suspect that artificial intelligence will grow out of one page of clever pattern recognition code combined with exposure to every pattern revealed on the Internet. It would take about a week to turn the pattern recognition code into what appears to us a sentient being with super intelligence. Futurists call that day the singularity.

As with most of my posts, some of you will tell me about all of the fiction books that say the same thing but said it first. I haven't read any of those books. Nor do I know anything about actual AI research, so perhaps it's obvious that pattern recognition is the key, and the real problem is that it's a hard nut to crack.

I came to my hypothesis that intelligence is just pattern recognition because people who are not terribly bright have trouble understanding analogies. And analogies are just patterns.

Logic isn't a big part of human intelligence. Put three humans in a room with a problem and each will have a different idea of the logical solution. Humans are rationalizers, not logical beings. Computers don't need logic to act human because humans don't have enough logic that anyone would notice some was missing.

Experience is little more than having more patterns to draw from. New situations are never identical to old ones, but they might follow a general pattern. For example, when I was in my twenties and someone said they would call me back I assumed it was true. Now I look at the entire situation, and all the patterns involved, and half the time I correctly tell myself I'll never hear from that person.

My two questions for the day:
  1. Is intelligence much more than pattern recognition?
  2. If intelligence is mostly pattern recognition, how likely is it that someone will write code that scours the Internet for patterns and uses that as a base to create super intelligence?
 

 
Rank Up Rank Down Votes:  +25
  • Print
  • Share
  • Share:

Comments

Sort By:
+4 Rank Up Rank Down
Feb 28, 2014
You mean computers will someday learn about humans by studying our browser history and the disrespectful way we treat each other on anonymous internet forums?

That can't be good.
 
 
+14 Rank Up Rank Down
Feb 28, 2014
Scott, this experiment has already been run with IBM's "Watson" system. Machine learning from applying pattern matching to the Internet does make an AI good at "Jeopardy" and medical diagnosis. Nothing that looks like consciousness has emerged. (Darn!) So, something beyond recursive pattern recognition in the recipe for consciousness is still missing. If, indeed, consciousness IS an emergent property, then it is unlikely that we will predict it before it appears. I just hope that, after it appears, we will be able to fully understand what made it appear. We humans are much better at working backward. As you have correctly noted, we follow behind observation and rationalize it.
 
 
+2 Rank Up Rank Down
Feb 28, 2014
1. Yes. (Once you recognize a pattern, what comes next? Is the pattern relevant? What's the context? Etc.)

2. Martin Armstrong (Pattern Recognition. Claims of AI, but not superintelligence)
 
 
+2 Rank Up Rank Down
Feb 28, 2014
Knowledge is pattern recognition. Intelligence is the ability to manipulate your thinking to adapt to new surroundings. Being "smart" is something I haven't really figured out despite being knowledgeable, intelligent, and meeting several "smart" people. Machines can be incredibly knowledgeable, can feign intelligence better than many people can manage on a good day, and may one day be able to act "smart" better than I can.

Still, should it try to replace me, it won't be able to protect itself when I walk up to it with a large sledge hammer in my hands. And shoot it.
 
 
+2 Rank Up Rank Down
Feb 28, 2014
Scott, what is the "pattern" used to solve a problem you've never seen before?

For example: if I give you a scale and 7 identical balls plus 1 ball that weighs slightly less. What is the minimal number of times you would need to weigh those 8 balls in order to determine with 100% certainty which ball weighed slightly less?

Let's try another one: how do you determine the area under a sphere, without first knowing calculus? that is, how do you use pattern recognition to come up with something new?

additional reading, for those interested:
http://lemire.me/blog/archives/2013/09/16/bricolage/

http://efgh.com/math/impossible.htm

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

[for those who care, the answer to the first question is 2.]
 
 
Feb 28, 2014
1. No; intelligence IS pattern recognition.
2. Writing the code is not the hard part. AI is more of a hardware problem, not a software problem. Computer processors still need to get much faster to simulate intelligence as we recognize it. This is explored in detail in several of Ray Kurzweil's books, especially "The Age of Spiritual Machines."
 
 
Feb 28, 2014
You should read "The Emperor's New Mind" by Roger Penrose and pay special attention to the part about mathematical proof of Penrose Tilings and why that proof cannot be done by a machine.

There is a fascinating history of Penrose Tilings. I'll be brief. Chemists noticed in the 1960's that there are some patterns of crystals that defied any explanation. Penrose had worked out the math by then of Penrose tilings but the Nobel Prize committee held off on giving the prize to the chemists who discovered this in nature until 2011 because pattern recognition told them it was crank science.

http://en.wikipedia.org/wiki/Quasicrystal

Background on Penrose Tilings:

http://en.wikipedia.org/wiki/Penrose_tiling

Sometimes, humans are SMARTER than PATTERN RECOGNITION.

The point of this is there is rational cognitive ability which biological minds possess - and no amount of algorithmic combination can produce the same insight in a machine. Penrose discusses this very elegantly in his books "The Emperor's New Mind" and "Shadows of the Mind".

If you are wondering who Roger Penrose is - he is the foremost mathematical physicist alive today.

Pattern recognition is an important quality of intelligence but so is the ability to discern valid patterns in the junk noise of data.

If you were to give a machine access to the entire internet anything intelligible or intelligent would be drowned out by the sheer volume of noise data.

You can do a simple test of this. Google does this sort of "pattern recognition" and you can view your Google profile and see who Google thinks you are. I read mine - and it's hilarious. They got it all wrong with me. Not very smart, imo - and a good example of how the noise of too much data and blindly applying machine algorithms to that data produces hilarity.

This link explains how to see what Google's pattern recognition thinks of you.

Commence laughter.

http://google.about.com/od/adsense/qt/Does-Google-Think-You-Are-A-Man.htm
 
 
Feb 28, 2014
Actually, intelligence is just a construct. Psychologists devised many definitions of intelligence througout years. The most cynical one said that intelligence is the trait that lets us solve intelligence tests.
I see your point, but "Is intelligence much more than pattern recognition?" is a wrong question. If this definition suits you, then it's fine.
The real question is, if your hypothetical master-pattern-recognizing robot existed, how good would he be in competition with humans on a given field. Would he be a good talker? Military or marketing strategist? Political leader? Artist? Lover? Intelligence may or may not be success requirement on this fields, depending how you define it.
 
 
Feb 28, 2014
I think that intelligence is pattern recognition, but I don't think the process of pattern recognition is nearly as simple as you make it out to be.

Our brains are meat computers that have developed pattern recognition abilities over millions of years, and are specialised in recognising the complex patterns of human interactions.

The synapses and mental algorithms involved in this pattern recognition are so complex that even our own brains - the current champions of pattern recognition - are lost in it. We do nearly all of it unconsciously.

I suppose it is possible that some genius programmer might be able to distill all of the complex interactions that make up the world as we experience it into a few lines of code, but I consider it unlikely. I think that a synthetic brain that was able to recognise the complex patterns of human interaction as well or better than a human would essentially be a synthetic human brain.

I think it more likely that the singularity will happen gradually, as we augment our meat computers with synthetic synapses one by one until every component of our brains are synthetic, but still fundamentally human in function and form.
 
 
+3 Rank Up Rank Down
Feb 28, 2014
I think AI is on the verge of explosive advancement. Natural language processing and pattern recognition (aka data mining) will yield amazing insights in all areas of our world. I'll call this "extracting knowledge from data" for lack of a better description. You can call it intelligence if you like.

However...
I believe there is more to "consciousness" than reductive logic. I was a design engineer in the super computer field in a previous life and I don't see any path where our current technology will scale to the point where a hardware/software system will say "Hey, wait a minute ...I've got a great idea!" I believe there is an "emergent" property to human cognition that involves processes beyond perception, memory and logic. I'm not saying that conscious, intelligent machines are impossible, just that it will require a different technology than we currently have.
 
 
Feb 28, 2014
It seems to me that IBM's Watson is basically this already - they built a pattern-recognition algorithm and fed it all of Wikipedia (plus a few other knowledge databases).
 
 
+13 Rank Up Rank Down
Feb 28, 2014
It's scary to imagine what a computer would "learn" about humans based on the internet. There aren't a lot of "bless you's" recorded.

Just think of what other cultures learn from American movies and multiply how bad it would be.
 
 
Feb 28, 2014
There are a couple of major problems with teaching computers to recognize patterns. The first is transferring the data into a format that the computer can run pattern recognition on. This sounds trivial, but is actually incredibly complex. Case in point, Siri. It's errors are very well documented and using that data will quickly lead to a garbage in-garbage out recognition. Parsing visual data is even harder, just look at how much trouble the UK spy agency has had filtering out !$%* from their yahoo cam data.

The second problem once you have all that data is teaching computers what data to ignore. In Scott's example of the library, humans intuitively eliminate non-important variables (temperature, light, phase of the moon, hair style, shirt color, dust motes, beams of sunlight) etc. to just focus on the two main factors, volume and the presence of the library. The possible permutations quickly explode the complexity of the calculation. While you can overcome this with an increase in processing power, we are many years away from being able to brute force these types of correlations.
 
 
Feb 28, 2014
[Speech is just pattern too. -- Scott]

Right. I agree.

In fact, I'd say human speech is a highly complex pattern. Not all information can be extracted just by "reading the words". There is stuff like context, emphasis, cultural differences, etc. And then there is humor. Once a computer can really understand jokes, I don't think there will be anything that can really separate a human mind from an artificial one.
 
 
Feb 28, 2014
Scott, if appears that you are saying everything is made up of patters, so let's flip this question on it's head. Is there any component of intelligence that does not involve patterns?

As long as the underlying natural sciences to our universe do not change, then it seems to me that everything will have a pattern. There are complexities ad infinitum, but computer will be better than us the closer we get to infinitum.

Also, this seems related on the point that all science can be boiled down to basic patterns: http://xkcd.com/435/

So, unless the beings running the computer simulation we are in start changing the rules, the singularity is inevitable.
 
 
Feb 28, 2014
[Simulating human thought is easy once you have pattern recognition solved. All you have to do is introduce random errors. It's the stupidity that defines us as human, not the thinking we do right. -- Scott]

Apparently you didnt get my point so let me try again: give a computer pattern recognition and well find out the same thing we found out when we gave them logic, memory and number crunching. Namely, that theres more to human thought than we thought and were not there yet.
 
 
Feb 28, 2014
According to Ray Kurzweil, director of engineering at Google, the key to real AI is speech recognition. In this article - http://www.theguardian.com/technology/2014/feb/22/robots-google-ray-kurzweil-terminator-singularity-artificial-intelligence?CMP=twt_gu - Language, he believes, is the key to everything.

"And my project is ultimately to base search on really understanding what the language means. When you write an article you're not creating an interesting collection of words. You have something to say and Google is devoted to intelligently organising and processing the world's information. The message in your article is information, and the computers are not picking up on that. So we would like to actually have the computers read. We want them to read everything on the web and every page of every book, then be able to engage an intelligent dialogue with the user to be able to answer their questions."

[Speech is just pattern too. -- Scott]
 
 
Feb 28, 2014
Once upon a time, in the sixties and seventies, it was beleived that computers were on the verge of replacing human intelligence because they kept getting better and better at logic, number crunching and memory. Today they're tops at number crunching and memory and any deficiencies they have in logic are due to programming not hardware, yet they havent replaced human thought. Why? Because when computers became what they are today we discovered there was more to human thought than logic, number crunching and memory. My pattern recognition tells me your hypothetical computer will be similar; it will be a very impressive advance but it still wont simulate thought.

[Simulating human thought is easy once you have pattern recognition solved. All you have to do is introduce random errors. It's the stupidity that defines us as human, not the thinking we do right. -- Scott]
 
 
 
Get the new Dilbert app!
Old Dilbert Blog