After the first day I would say that there is a definite advantage to the machine with the button pushing. On many questions where any Jeopardy! player would know the answer without hesitation Watson wins the race to push the button. So they are demonstrating a mechanical advantage. That is not much of technological breakthrough.
There have also been some interesting flubs on the part of the programmers.
First, the logistical problem. Watson gets information by text and electronic signal on when to answer, and apparently hasn’t been designed to interpret speech. In one case it gave a wrong answer after that same wrong answer had been given. It simply didn’t know that that had happened. It did know that it could try to answer – probably part of the whole signal button activation message – but it didn’t know that it was giving the same wrong answer. That one’s not a big deal. They focused on language interpretation which makes this a problem, we’ll see how serious, but it falls outside of the stated goals. It’s also impractical to surmount because it would require someone texting the response given by other contestants to Watson, and that wasn’t part of the plan.
The other issues are more programming problems. On more than one occasion Watson gave responses that were obviously wrong because it did not evaluate the entire answer. The response Watson gave seemed to be based on analysis of the first part of the answer that ignored the end. That’s either because of a prioritization based on length or punctuation, or it’s a deliberate shortcut. Either way it causes problems.
The two that I can think of demonstrated that Watson ignored or did not recognize crucial words in the answer. One was a question about trains where the response was ‘finis’ when the correct response was ‘terminal’. Watson evaluated the language reference in the answer but missed the association with trains. I don’t know if that, and the other case that escapes me at the moment, indicate a slip in textual analysis or a bug in the instructions.
I think I noticed one strategy in that Watson selects categories that have not been chosen yet when it has the chance. It seems to need a sample answer to evaluate the category. That’s not unlike human contestants, but I don’t remember ever seeing a human deliberately bounce around the board to do it, though.
So, there have been some interesting things, but I think Watson’s lead can be attributed to speed on the button rather than an advantage in ‘skill’. The graphic showing the possible responses and their weighting that Watson has picked is very interesting.
The fact that Watson knows Sauron but not Voldemort I will attribute to the age and preferred reading of the programmers – though I could be wrong. Maybe it was just the wording. After all, I know both of those.
Tuesday, February 15, 2011
Monday, February 14, 2011
Man vs Machine
Well, man vs a machine and all of the people who designed the processes behind it.
If you haven’t heard, an IBM computer is on Jeopardy!. I heard a few details on the radio that confirm my suspicions. There will be cheating involved.
At first the folks at IBM didn’t want to use the signal button, they wanted to just electronically transmit the answer. When the Jeopardy! people insisted that they use the button, IBM complained that they were imposing human limitations on their machine. Well, yeah, since this is supposed to be a competition against humans and that’s part of the rules.
Next comes the part I was thinking about – when the Watson computer hits the button to answer. Your average human contestant has to wait until the answer is read and then hit the button, so they anticipate when that will be. If they ring in early there is a .25 second penalty. Watson knows when the light goes on indicating that it can hit the button and never makes a mistake on the timing and it can respond faster than a human. That sure is fair and just like a human.
One of the comebacks is that the humans have the advantage because they understand the nuance of the questions and it takes Watson 2 – 3 seconds to figure it out. In other words, less time than it takes to read the answer. And I thought the whole point was to demonstrate that the computer could understand nuance. So they cheat.
Also, Watson gets the answer as electronic text, no reading involved the way the humans do it – or listening if they are blind as has happened. It’s getting more and more human all the time here.
Now let’s get to the Daily Double wagering. Watson calculates exactly how many questions are left, how much money can still be won and how much it needs – and what it’s opponents scores are and determines it’s wager based on what it can afford to lose and what it needs to win based on criteria that few if any humans can track at any time let alone on a TV stage in the middle of a Jeopardy! game. Not much of an advantage there. Well, that one is not that incredible and advantage, but it is one.
And then there’s the memory that the machine has. Do you think it can store more answers than the average human can?
Look, I’m intrigued by the logic behind this. Interpreting the answers and then searching for the right answers based on English language is an interesting challenge. They say it has shown no insight into how the human brain works, but I can see it leading to language research. It is being mentioned as a way to do fast research since it can search, say, all medical journals and texts and then return an answer that a doctor can evaluate and use or discard.
I think this is just another example where, when humans can quantify how a task is performed then figure out how to mechanize it, they can build a machine that is faster and stronger than a human. We’ve had steam hammers, electronic calculators, computers, computers that play chess (Deep Blue) and now this. That’s fine. We’ll see a test of how well people can make a machine to perform a task that has only been done before by humans. Just don’t call it a fair contest when the machine is allowed to have advantages unrelated to the human component that is being touted – namely language interpretation and solution selection. That’s like saying a car racing a human is fair because all you’re testing is how to navigate the race course mechanically.
Yeah, I’ll watch it. Like I said, it’s an intriguing problem they’ve apparently solved and I want to see the demonstration. I just want them to admit that there is no fair contest being waged here. What we have is a very long, and hopefully interesting, commercial for IBM.
If you haven’t heard, an IBM computer is on Jeopardy!. I heard a few details on the radio that confirm my suspicions. There will be cheating involved.
At first the folks at IBM didn’t want to use the signal button, they wanted to just electronically transmit the answer. When the Jeopardy! people insisted that they use the button, IBM complained that they were imposing human limitations on their machine. Well, yeah, since this is supposed to be a competition against humans and that’s part of the rules.
Next comes the part I was thinking about – when the Watson computer hits the button to answer. Your average human contestant has to wait until the answer is read and then hit the button, so they anticipate when that will be. If they ring in early there is a .25 second penalty. Watson knows when the light goes on indicating that it can hit the button and never makes a mistake on the timing and it can respond faster than a human. That sure is fair and just like a human.
One of the comebacks is that the humans have the advantage because they understand the nuance of the questions and it takes Watson 2 – 3 seconds to figure it out. In other words, less time than it takes to read the answer. And I thought the whole point was to demonstrate that the computer could understand nuance. So they cheat.
Also, Watson gets the answer as electronic text, no reading involved the way the humans do it – or listening if they are blind as has happened. It’s getting more and more human all the time here.
Now let’s get to the Daily Double wagering. Watson calculates exactly how many questions are left, how much money can still be won and how much it needs – and what it’s opponents scores are and determines it’s wager based on what it can afford to lose and what it needs to win based on criteria that few if any humans can track at any time let alone on a TV stage in the middle of a Jeopardy! game. Not much of an advantage there. Well, that one is not that incredible and advantage, but it is one.
And then there’s the memory that the machine has. Do you think it can store more answers than the average human can?
Look, I’m intrigued by the logic behind this. Interpreting the answers and then searching for the right answers based on English language is an interesting challenge. They say it has shown no insight into how the human brain works, but I can see it leading to language research. It is being mentioned as a way to do fast research since it can search, say, all medical journals and texts and then return an answer that a doctor can evaluate and use or discard.
I think this is just another example where, when humans can quantify how a task is performed then figure out how to mechanize it, they can build a machine that is faster and stronger than a human. We’ve had steam hammers, electronic calculators, computers, computers that play chess (Deep Blue) and now this. That’s fine. We’ll see a test of how well people can make a machine to perform a task that has only been done before by humans. Just don’t call it a fair contest when the machine is allowed to have advantages unrelated to the human component that is being touted – namely language interpretation and solution selection. That’s like saying a car racing a human is fair because all you’re testing is how to navigate the race course mechanically.
Yeah, I’ll watch it. Like I said, it’s an intriguing problem they’ve apparently solved and I want to see the demonstration. I just want them to admit that there is no fair contest being waged here. What we have is a very long, and hopefully interesting, commercial for IBM.
Monday, February 7, 2011
It’s probably not as easy as they think
As the Super Bowl was about to start yesterday I wondered if people bothered to rehearse the national anthem or if they all figured they knew the song and had it covered. The Star Spangled Banner has a tough range. And while you may know the lyrics and can sing it in a crowd, I imagine it’s different when you’re standing in a stadium filled with more than 100,000 people, on TV with millions of viewers, at the largest single sporting event there is.
Christina Aguilera wasn’t the first to flub it, probably not the worst and I’m sure she won’t be the last. This will be on the news for a few days and on YouTube forever, but she’ll laugh it off eventually. Maybe it will be a lesson to others who step up there to sing.
Christina Aguilera wasn’t the first to flub it, probably not the worst and I’m sure she won’t be the last. This will be on the news for a few days and on YouTube forever, but she’ll laugh it off eventually. Maybe it will be a lesson to others who step up there to sing.
Icons
I understand the appeal of icons. I also understand the appeal of Ronald Reagan as a conservative icon. What I don’t understand is how Reagan actually is a conservative icon today.
In the 80s Reagan represented what the Republican party stood for. He didn’t actually do those things, but he was the public face of party policy. He was the President so he was the image of the party’s position on the issues. The public image was in large part opposite of what was happening. But the President gets the benefit of the doubt especially when the messaging is so good.
But today the facts are obvious and well documented and disseminated. Yet the icon remains.
I suppose it’s the need for a symbol that is driving the message. Very little that Reagan actually did lives up to modern standards for Republicans. So the focus is kept narrow to allow the image to survive. Without that there is nothing to rally around. That gives strength to the message. It gives strength to the Republicans. So the icon must be supported.
It is a weakness for Democrats that they don’t have that sort of image to use to focus attention and action. It makes the message weak, whatever the strength of the message. The image outweighs the facts. A strong image supports the message even if it is built on lies and half-truths. A weak image destroys the message however important the message is.
Marketing strategy is very important when you are trying to sell something.
In the 80s Reagan represented what the Republican party stood for. He didn’t actually do those things, but he was the public face of party policy. He was the President so he was the image of the party’s position on the issues. The public image was in large part opposite of what was happening. But the President gets the benefit of the doubt especially when the messaging is so good.
But today the facts are obvious and well documented and disseminated. Yet the icon remains.
I suppose it’s the need for a symbol that is driving the message. Very little that Reagan actually did lives up to modern standards for Republicans. So the focus is kept narrow to allow the image to survive. Without that there is nothing to rally around. That gives strength to the message. It gives strength to the Republicans. So the icon must be supported.
It is a weakness for Democrats that they don’t have that sort of image to use to focus attention and action. It makes the message weak, whatever the strength of the message. The image outweighs the facts. A strong image supports the message even if it is built on lies and half-truths. A weak image destroys the message however important the message is.
Marketing strategy is very important when you are trying to sell something.
A Centenary
Ronald Reagan was born 100 years ago yesterday. He is idolized by quite a few people in the country, and I’ve never understood why. I get that the people he helped would like him, as well as those who supported his policies. What I don’t get is why so many people love him for being something he wasn’t.
Contemporary conservatives who espouse small government and low taxes and fighting terrorism and having bigger and better bombs seem to love the guy. But for a conservative icon Ronald Reagan did some strange things. Yes, he attacked unions, used divisive campaign tactics, opposed communism, demonized government and he cut taxes – once.
He also raised taxes 11 times, 7 of his 8 years in office.
He increased the size of the federal government.
He increased government spending.
He more than doubled the deficit, increasing it more than all previous Presidents combined.
He supported amnesty for undocumented immigrants.
He sold weapons to Iran.
He helped fund South American rebels.
He got Marines killed in Lebanon and then cut and ran.
He saved Social Security – in part by raising taxes.
He signed the first START treaty.
He wanted a world free of nuclear weapons.
I’m not saying that all of those things are bad, but I’m really not seeing the conservative hero in there.
ETA: This isn't meant as an attack on Reagan or anyone else. I just don't understand the idolization when the facts don't match the desired message.
Contemporary conservatives who espouse small government and low taxes and fighting terrorism and having bigger and better bombs seem to love the guy. But for a conservative icon Ronald Reagan did some strange things. Yes, he attacked unions, used divisive campaign tactics, opposed communism, demonized government and he cut taxes – once.
He also raised taxes 11 times, 7 of his 8 years in office.
He increased the size of the federal government.
He increased government spending.
He more than doubled the deficit, increasing it more than all previous Presidents combined.
He supported amnesty for undocumented immigrants.
He sold weapons to Iran.
He helped fund South American rebels.
He got Marines killed in Lebanon and then cut and ran.
He saved Social Security – in part by raising taxes.
He signed the first START treaty.
He wanted a world free of nuclear weapons.
I’m not saying that all of those things are bad, but I’m really not seeing the conservative hero in there.
ETA: This isn't meant as an attack on Reagan or anyone else. I just don't understand the idolization when the facts don't match the desired message.
Wednesday, January 26, 2011
State of the Union - Follow up
I got most of it right.
There was talk about sacrifice and compromise, though not always in those words. Big time talk about the tax code and corporate taxes, whoopee. Investing in the country, but that was probably always a gimme since it’s such a popular topic. Same for the strong future.
I was flabbergasted that the long-term concept I expected turned out to be a 5-year freeze on spending levels. What a monumentally bad idea. Someone explain to me how we invest without spending.
I’d say the not raising taxes was not explicit, so I’ll call that a push.
The talk on health care was as I expected, so, I got it right but I’m not impressed. The same for talking about the amazing American worker – he mentioned it, but with no specifics or real enthusiasm.
I’d say it was mostly a move to what people in DC like to call the center.
Not enough about protecting Social Security from the predators, but no talk of changes yet. I didn’t really hear much specific about the economy – did I miss it?
So, I claim a high level of accuracy, if not being very thorough in my predictions. It was a standard State of the Union address – vague, bland and apparently designed to leave little room for criticism. Pretty much what I expected (did I say that already?).
I can’t wait until next year when we can all be called on to look to the future and make our country stronger without any mention of how nothing got done based on this speech.
I didn’t pay attention to the so-called rebuttals.
Oh, there was one surprise. I didn't know that Biden was that much taller than Boehner. That's about it.
There was talk about sacrifice and compromise, though not always in those words. Big time talk about the tax code and corporate taxes, whoopee. Investing in the country, but that was probably always a gimme since it’s such a popular topic. Same for the strong future.
I was flabbergasted that the long-term concept I expected turned out to be a 5-year freeze on spending levels. What a monumentally bad idea. Someone explain to me how we invest without spending.
I’d say the not raising taxes was not explicit, so I’ll call that a push.
The talk on health care was as I expected, so, I got it right but I’m not impressed. The same for talking about the amazing American worker – he mentioned it, but with no specifics or real enthusiasm.
I’d say it was mostly a move to what people in DC like to call the center.
Not enough about protecting Social Security from the predators, but no talk of changes yet. I didn’t really hear much specific about the economy – did I miss it?
So, I claim a high level of accuracy, if not being very thorough in my predictions. It was a standard State of the Union address – vague, bland and apparently designed to leave little room for criticism. Pretty much what I expected (did I say that already?).
I can’t wait until next year when we can all be called on to look to the future and make our country stronger without any mention of how nothing got done based on this speech.
I didn’t pay attention to the so-called rebuttals.
Oh, there was one surprise. I didn't know that Biden was that much taller than Boehner. That's about it.
Monday, January 24, 2011
State of the Union Predictions
At the moment I would expect to see the following things mentioned in some form in the SOTU tomorrow:
Shared sacrifice
Everyone has to be involved
Compromise
Bipartisanship
Changes to Social Security
Raise the retirement age
No benefit cuts
Tax code reform, emphasis on corporate and cuts
Invest in the country
Look to a strong future
The economy is strong but not yet recovered
Implicit move to the so-called center
A plan for something that won’t be completed until after 2012
And some form of ‘everything on the table’ where the ‘everything’ excludes raising the high end marginal tax rate
List accomplishments
Something about health care reform
Lip-service to workers
Pabulum, promises and no real plan to fix things.
Mostly I expect it to be vague, with maybe one specific big item to use as a focus – which would also work as a distraction from the major problems that aren’t being addressed.
Shared sacrifice
Everyone has to be involved
Compromise
Bipartisanship
Changes to Social Security
Raise the retirement age
No benefit cuts
Tax code reform, emphasis on corporate and cuts
Invest in the country
Look to a strong future
The economy is strong but not yet recovered
Implicit move to the so-called center
A plan for something that won’t be completed until after 2012
And some form of ‘everything on the table’ where the ‘everything’ excludes raising the high end marginal tax rate
List accomplishments
Something about health care reform
Lip-service to workers
Pabulum, promises and no real plan to fix things.
Mostly I expect it to be vague, with maybe one specific big item to use as a focus – which would also work as a distraction from the major problems that aren’t being addressed.
Subscribe to:
Posts (Atom)