I’ve had a cold, woke up, visited Twitter, and read a thread that informed me that Anthony Watts fears Googles new SEO algorithm which involves a new “trust” or “truth” rankings of some sort. After taking another sip of cold syrup, I indulged my curiosity about the new “trust” ratings, clicked, read the paper describing how Google is going to come up with “trust indices”: I suspect Anthony has nothing to fear!
Yes: It seems Google is going to add some sort of “trust” index to determine whether sites contain “true” facts or “false” facts.
My impression is that Anthony’s fear stemmed from thinking biased humans were going to be involved in creating the “trust” indices and that potentially biased “trust” index would be used to change page ranks. As Anthony currently gets a lot of Google traffic, he might fear losing some.
But I also thought: using honest to goodness humans in real time that won’t scale for numerous topics. Beyond that it wouldn’t stay up to date. (What are the current FDA dietary recommendations. Did they ditch the food pyramid? Anyone know?)
Anyway, one of the tweets contained a link to the theoretical paper describing the algorithm for determining whether pages are “truthful”. The paper appears to be “Knowledge Based Trust: Estimating the Trustworthiness of Web Sources” by Dong et al. The researchers evidently had access to “a database of 2.8B facts extracted from the web” and “estimate[d] the trustworthiness of 119M webpages”.
The method seems to involve extracting “triples” of information like c(Obama, nationality, USA). If your ‘triple’ is right, you get brownie points. If your triple is wrong (e.g. you said his nationality is Kenya) you get dinged. If you page contains no “facts” that can be tested against the other facts in the database, that’s neutral. But the algorithm does a bit more, like figuring out the topic of your site, detecting triples, deciding if those triples are ‘relevant’ to your web site, and then deciding if the triples indicate the presence of a “true fact” and whether the fact is “trivial”. We don’t know the exact definition of “trivial”; the example given is “if most sampled triples from a Hindi movie website state that the language of the movie is Hindi, we consider it as trivial”
So, possibly, the following “triples” might gain a blog on climate change brownie points:
Or possibly not. We can’t be sure. But I strongly suspect that despite its detractors negative opinion, automatic extractors combing WattsUpWithThat will find quite a number of clearly true triples. These include triples that are created when Anthony or his co-bloggers describe what the IPCC AR4, AR5, TAR and what have you predicted correctly and then proceed to explain why they think the predictions will not pan out.
But we can be fairly certain that as soon as Google starts using the ‘truth’ algorithm, the people who make a living in “Search Engine Optimization” will start finding ways to detect which sorts of “triples” help pages rank well. They’ll study how to inject relevant “triples” that don’t appear unnatural.
My tentative advice to Anthony: If you are hoping to amp up your “trust” level, try to wedge in roughly 3 triples per page, making sure some are already in the “fact” data base, but including new ones. Like, for example “Australian government has decided to not renew any of the management board of the Australian CSIRO.” might eventually be trust boosting triple. Provided it’s true; ( I haven’t checked.)
I have a bad cold. Talk about almost whatever you like. Try to avoid carrying over food fights from other blogs though. I’m going to go take some more cough syrup! Meanwhile, speculating on possible SEO strategies based on this “trust” index are welcome!
“Provided it’s true; ( I haven’t checked.)”
It’s actually totally untrue. They aren’t giving the chairman another term. Chair is a fixed term position usually occupied by some prominent businessman. I’m not aware of any that have been renewed in the past. I think wanting another term is fairly unusual.
Other board member terms (usually 3 yr) will expire in due course. McKeon thinks the Gov’t won’t renew them. Gov’t denies such a policy.
But I’m not sure how a computer would check.
Nick Stokes
My impression– not for the linked paper but other articles– is Google’s determination of truth actually depends on what pages on the internet say! So if 100% of pages that comment on the discovery of America say Columbus discovered American in 1492, then that fact is “true”. If 80% say it’s true, it’s likely true. If 80% say something different (like 1066 or something) then the claim he discovered it in 1066 is “true”.
It all has to be automated!
There was another article that discussed interviewing medical experts on things like vaccinations, but my impression was that was a separate system.
Given how everything has to scale and they can’t have too much human intervention, I’m really not sure how this will pan out. I don’t know if there is some threshold of number of people saying something before a fact is deemed “true” and so on.
Also: my impression of how deeming a triple “true” will occur may be entirely mistaken. 🙂
While I can understand Anthony’s concern (remember the doubts about Google search rankings of WUWT), I suspect there is enought difference of opinion on the web that nobody is going to define ‘the truth’ very easily. OTOH, if Google decides that not ‘being evil’ includes promoting a specific view on global to enforce green ideology, then all bets are off.
“if Google decides that not ‘being evil’ includes promoting a specific view on global to enforce green ideology, then all bets are off.”
How else would one avoid being evil?
Stop eating meat!! No wait, stop eating vegetables! No, wait, we are supposed to burn vegetables, pet meat and squeeze fruit. No, no, that’s not it either. I’m just not sure anymore.
Jeff,
You miss the point: there are far too many humans in existence to ensure ‘a sustainable future’. The moral prohibitions on virtually all foods are an attempt to make you starve to death in a sea of plenty, thus lessening the human burden on Gaia and her many noble (non-human) species. Let’s face it, the world would be a far better place if you, me, and most of the rest of humanity did not exist. Ask Greenpeace activists or the british Royal Family if you have any doubt. 😉
Feel better Lucia.
If Google groups like sites , then give a trustworthy rating to them , sites not grouped will be labeled as untrustworthy however valid their triplets.
Seems like something Google could do using that site formula put up by the pretty blonde at JC a couple of months ago, with shades of Lewindowsky grey.
JeffId,
Yes. If Google decided what is “true” based on some values they hold, all bets are off. But my impression is Google really, really like things to be determined by “algorithms”. Among other things, the algorithms scale. So, my guess is that’s not what’s going to happen even if some people at google would like the outcome of the algorithm to have specific results (like possibly dinging Anthony’s site.) The thing is: Google culture really values having an ‘algorithm’. (And that’s what’s made them money in the end.)
Beyond that– if they do start injecting the judgement of individual people at Google, their search engine will become less ‘useful’ to those doing search because people won’t find what they are trying to find.
As a silly example: Suppose Satan himself had a personal web site. I want to find that site— for whatever reason.
So I google “Satan”. His site doesn’t show. Then I google “Satan’s web site”. It doesn’t show. Then I google “Satan’s personal web site” and so on. But now suppose google has — for whatever reason– deemed Satan’s web site “not truthful” and so doesn’t return that in Google search. To find the site, I will have to use a different search engine– or email some friends or find an email list serve or something. (Hey! That’s what happened when Yahoo had their stupid human curated search! If you wanted to find coned knitting yarn, you had to find an email list. The human’s at Yahoo didn’t like “amateur non-polished sites” and wouldn’t list any of the small businesses that had coned knitting yarn. Then Google came along, and suddenly we could find coned knitting yarn!)
But anyway, Google knows perfectly well that there is a limit to how much of their own values they can inject into search before their search engine become “not useful to people who want to search”. At that point people would switch to other search engines — some do exist!
“I suspect Anthony has nothing to fear! ”
Famous last words.
I hope the good folks at Google have a copy of Samuel Arbesman’s ‘The Half Life of Facts.’ When I went to school we had nine planets and 105 elements in the Periodic Table… dose were de fax. (Dat’s me imitating de effex of a bad code… get better quickly!)
Tom,
Heh! Yeah. Some “true facts” are unstable. I’d say official dietary advice may contain the largest number of unstable facts. Is dietary cholesterol bad? Which is worse for you, high or low salt consumption?
BTW: those aren’t really questions at all. They are examples of questions whose ‘correct’ answers seem to have changed. I think the only stable dietary advice has been: vegetables and produce are good for you. But even with that– the food pyramid era spent an awful lot of time making it seem you should live on dry bread, rice and so on. Yes, the pyramid said to eat veggies and fruits– but I think everything I read emphasized Grains! Grains! Grains!
Definitely wanted you to get calories with low fat.
The statement ‘cholesterol is not a nutrient of concern’ was always true. It was the dietary advice that was wrong for twenty years or so. High levels of salt intake raise blood pressure for a small fraction of the population, ~10% IIRC, but don’t matter for the rest.
The statement that there were 105 elements in the periodic table in 19xx is still a fact. A statement in 19xx that there would never more than 105 elements was not a fact even then.
Speaking of grains, one can determine from skeletal remains when agriculture began to be practiced as opposed to hunter/gatherer. General health always deteriorated when agriculture was first introduced. Replacing meat with grain is tricky when one doesn’t know anything about amino acids and human metabolism, not to mention the micro-nutrients.
DeWitt,
Are you sure on the cholesterol advice?
“http://www.cnn.com/2015/02/19/health/dietary-guidelines/”
Quotes
It may be the case that the previous advice to not overcomsume cholesterol was misguided and that possibly the DGAC should not have been concerned. But it would appear the DGAC was concerned about over-consumption of cholesterol.
So, it was a nutrient of “concern”– in the sense that people are concerned, and in this case, plenty of well respected people in the nutritional field were “concerned” about overconsumptoin and advised against it.
On the hunter/gather thing and eating grains.
I think part of the problem was replacing non-grain veggies with grains. We may put the word “hunter” at the beginning of “hunter-gatherer” but gathering tended to supply a large amount of calories.
Mind you, I like grains. I even think grains are fine for you. But in many places, agriculture swung a balance in the direction of making grains an over-abundant portion of calories because they tend to be easy to grow, they can be stored fairly easily, they are filling, people do need calories to live and so on. It is very easy for an settled group of people with access to a grain that does well (e.g. corn) and a few other food sources that do well to plants a relatively restricted variety of foods and exist on those.
When populations switch to agriculture, lost of people manage to make it to childbearing years, they have lots of kids, their kids survive. But yes, general health actually degrades– as in their teeth rot and their bones are soft and so on. Except they don’t actually die from rotten teeth, weak bones and so on. In the process of not dieing, and reproducing, they displaced the hunter/gatherers– even if the hunter/gatherers are in some sense “more healthy”.
In contrast, by its nature, gathering results in calories coming from a wider variety of foods (fruits, nuts, roots, seeds and so on.) These carry along with them more minerals and vitamins– and even generally fiber. That does seem to be good for general health. On the other hand, the lifestyle is rather marginal!
“I have a bad cold. Talk about almost whatever you like. Try to avoid carrying over food fights from other blogs though. I’m going to go take some more cough syrup! Meanwhile, speculating on possible SEO strategies based on this “trust†index are welcome!”
That must be a very bad cough or a very good cough syrup.
I have not read in detail what has motivated Google to think about using or actually using a truth algorithm, but I would suspect it might be from some users of the internet bad experiences getting very bad advice or very wrong self diagnostics. Unfortunately if the user of the internet does not have some very healthy skepticism about what is found there and does not do extensive research on all sides of an issue that user could proceed down a path that can spin all the truth from a proposition.
If Google becomes an arbitrator of the truth on the internet, in the end for the careful skeptic it would be only one additional source of information or guide to information. If Google’s algorithm went the way of the consensus on AGW then it would of course have all the baggage that any consensus has – including being necessarily vague and based on potentially uninformed and/or biased opinions.
The similarities between cholesterol and ‘climate change’ exaggerations are interesting.
Both theories rose to prominence over objections.
.
Both were promoted by advocates who rose through ‘authoritative’ organizations ( American Heart Association, IPCC, NASA-GISS ).
.
Both tended to ignore or supress contradicting evidence.
.
Both co-opted the cause for economic gain ( Con-Agra, climate funding ).
.
Both were used by governments with an agenda who didn’t have the luxury of waiting for evidence ( McGovern ).
.
Both vids are worth watching ( when you have time ):
.
Clips from ‘Fat Head’:
.
https://www.youtube.com/watch?v=xbFQc2kxm9c
.
Big Fat Fiasco:
.
https://www.youtube.com/watch?v=exi7O1li_wA
See ‘Big Fat Fiasco’ for the science take.
.
https://www.youtube.com/watch?v=exi7O1li_wA
“””On the hunter/gather thing and eating grains.
I think part of the problem was replacing non-grain veggies with grains. We may put the word “hunter†at the beginning of “hunter-gatherer†but gathering tended to supply a large amount of calories.”””
.
Consider what was being gathered. Most things were much lower density ( in calories, but more importantly in carbohydrates ) so the glucose load of eating a bunch of leaves or greens or even roots was a lot lower than high fructose corn syrup and white bread.
.
The Tom Naughton videos at one point discusses the max load of glucose in the blood being about a teaspoon or two. Glucose is both necessary for brain activity but also toxic above that small amount, so when modern diet loads up glucose, the body MUST produce insulin to reduce the toxic load, and fat cells are REQUIRED to store the glucose. The damaging effects of insulin do seem to account for heart disease and obsesity, diabetus, and coronary disease do seem all linked.
.
Grains enabled civilization because they were so easily stored and transported, enabling specialization in villages instead of hunting and gathering. But grains are very dense in carbohydrates and modern processing ( Corn Syrup ) are even denser.
Kenneth
I wouldn’t be surprised to learn the main motivation is to improve search in the sense of helping people find what they want to find and they think this would help.
You might not notice it, but if one googles on highly commercial topics, pages that are pure advertising trickle up. Try “laser face lift”. Click a few links– most are pure advertising. Also, entertainment questions often end up dominated by gossip pages. It may very well be that google knows people want to find pages that aren’t just sales outlets. ( Do laser face lift site:www.mayoclinic.org instead. Maybe that’s what people want to find. But Mayoclinic doesn’t spend on SEO the way a company selling laser facelift equipment does. Also, I only picked Mayoclinic because I know they have a site. There may be other sites that would be better.)
Also, the paper mentions sites like “yahoo answers”. My nephew sent me a photo of his diffyQ homework. He was frustrated he couldn’t get an answer anywhere near close to what was right on yahoo answers. Guess why not? Because the answer highlighted as “best” on yahoo answers was WRONG. Why was it highlighted as “best”? Possibly, it was highlighted by student who was unable to tell it was wrong. (I mean, really, do people who know how to do this things visit yahoo answer to check whether the answers are right?)
I’m not sure if this algorithm is going to help– especially on math at yahoo answers. But it might help for people to know that sites like “yahoo answers” are not really checked.
Climate Weenie,
Fructose may be worse than glucose. Fructose, like ethanol, can only be metabolized in the liver. The problem with glucose could be more not enough activity rather than too much glucose.
lucia,
That some people were concerned about over consumption of cholesterol was a fact. But their being concerned did not make the concern valid. Serum cholesterol level is mainly a proxy for serum lipoprotein level. The idea that one could maintain the same fat intake but lower cholesterol intake to reduce serum lipoprotein concentration and thus arterial plaque formation was never valid. Your cells make about as much cholesterol as you absorb from your diet. The converse, that if you reduced fat intake but maintained the same cholesterol intake, it would not reduce serum lipoprotein is also not true. That should have been blindingly obvious. That it took this long to be recognized doesn’t do much for the credibility of the FDA’s dietary advice in general.
Interesting concept that what we eat determines whether we are healthy or not, as in hunter good [sorry hunter], gatherers bad.
“general health actually degrades– as in their teeth rot and their bones are soft and so on. ”
Looked at in a different light we were probably gatherers for many more millenia than hunters and our teeth problem, diabetes, and cardiovascular issues are inbuilt rather than diet caused.
We are programmed to develop diseases and die, fortunately our life span is 80 years . A
. A horse or dog will only live 20 years, one vegetarian one a carnivore. There are fat dogs and thin horses and they all have their specific dietary diseases.
Saying a hunter is healthier than a gatherer [in general] is equating to blaming the diet [like cholesterol] when really the diet does not matter in normal circumstances.
There are only so many heart beats to a heart before it stops and also only so many spoonfuls of sugar to developing diabetes, but only to those people who are born to have heats stop or develop diabetes [like Gattica].
Well, our varied teeth indicate that we’re omnivores, all right.
But it’s not hunter -or- gatherer. It’s what we gather. In evolutionary past, maybe we gathered up a lot more weeds like dandelions and a lot fewer hershey bars. A lot of the fossil evidence of past diet is sketchty, but it looks like wheat, rice, corn, sugar and citrus fruit are really recent ( 2 to 10 thousand years old ) inventions. A long time for history, but a brief time for evolution.
Then why do we sweat? The answer is that humans evolved as cursorial or endurance hunters. We ran down our prey on the plains of Africa. The ability to sweat allows humans to run farther than any other animal except some canidae and hyenas. Wolves don’t sweat but their muzzles are extremely efficient at rejecting excess body heat.
Ya – so I’m going for bar-b-que and salad, but no sweet tea or cobbler.
Given that the average lifespan of hunter/gatherers was about 21, it is questionable whether dietary imbalances played a significant role in mortality/morbidity.
Lucia is correct to note that the bulk of calories were provided by the gathering part of the menu, with meat serving more as a prestige offering. Most animal protein came from creepy crawling things rather than the results of the great hunts, which may have been more an excuse to get out of the house than anything else, if modern studies of hunter gatherers are relevant.
Tom,
I’m betting that the average life span statistic is heavily biased by high infant mortality for both sexes and high mortality for women from bearing children. I would be curious to see if there were statistics on life expectancy for males that reached puberty and survived the manhood initiation rites, but not curious enough to actually look.
lucia (Comment #135713)
You are probably correct about the motivations of Google for developing a truth algorithm. I doubt the cases where advertisements get in the way of a search will be high on Google’s priority list since advertisements are the Google business plan.
I suspect anything that that Google might do to “help/protect” the internet user/searcher might end up like a lot of government regulations in that it gives those being protected a false sense of security with those being regulated finding ways around the regulation.
Kenneth,
Some advertisers pay Google. But other web sites are “ads” for services etc. with little else. Those advertisers are not paying google. They place information– often of dubious accuracy– on their sites to make google think it’s an information site. But it’s really just a site hawking face lifts, diet aids, pharmaceuticals etc.
Google knows that when these things end up “high” in the search ranks those doing search don’t find what they want– which is often more objective information about plastic surgery, diet, drug side effects etc. So Google probably does want an automated way to detect that a site is about… well… new unproven slenderizing pills or some such.
” But now suppose google has — for whatever reason– deemed Satan’s web site “not truthful†and so doesn’t return that in Google search.”
You’ll just have to scroll down the 66th page, Lucia.
So this is sorta like a Cronbach’s alpha for the interwebs? Echo chamber, anyone?
What if most sites on the web say the Sun orbits the Earth?
Because of Google’s crazy and arrogant revision of its compose feature on gmail about 18 months ago, I have no trust in Google. I have used gmail for many business purposes and have contracts that were done solely by email. One day (semi out of the blue), Google decided to reduce my usable screen by about 60% and prevent me from editing the subject line in the standard gmail when I replied to emails. It became totally unusable, and I hated it.
Thousands of people griped to Google (about 98% of the commenters hated it.) Google arrogantly stated “tough luck” and lied about their reason for not giving an option to keep the old standard gmail, saying that it would be too expensive. See https://productforums.google.com/forum/#!topic/gmail/EFUIeCXRTug%5B651-675%5D Luckily, someone came up with a Firefox addon called “old compose”, and it works. I was going to switch to Outlook and then noticed that it quite often crashed.
In the meantime, I switched to Bing as my search engine and try avoid all new Google products. So, it may very well be that Google’s proposed move is harmless, or maybe even useful. However, when dealing with Google, I always have my guard up.
JD
What do you do if you can’t stand google, hate microsoft, won’t have anything to do with yahoo, despise mozilla, can’t use linux, etc. Nothing left to do but become a hunter-gatherer.
Then again, maybe google can sift through and find the actual truth ( ie the world hasn’t warmed for x years). An algorithm that really found the truth – that would be worth waiting for!
Google can try this. There’s no doubt in my mind that people will leave for other places. Sometimes even big companies self destruct. McDonalds is floundering after years of coddling to health warriors that don’t eat hamburgers or want happy meals. They even fired Ronald. Google will be the trusted place for information that nobody wants.
Goodle could name the new truth algorithm, “Diogenes”
http://en.wikipedia.org/wiki/Diogenes_of_Sinope
….for many reasons…..
I see. So he just lists a bunch of facts and then doesn’t get dinged for “Climate sensitivity is zero,” “Sea levels aren’t rising” and “I’ll be publishing my results soon.” Interesting. 🙂
Boris,
I know you think the other things would matter more– but how would an automated algorithm by google realize your opinion on which matter more is the correct one? Real question btw. I can’t see how it would. I especially can’t see how it would figure out “I’ll be publishing soon” is a “fact” that has anything to do with “climate change” nor that it is “wrong”.
“Yes, the pyramid said to eat veggies and fruits”
Eleven years ago my sister told me of an incident when she drove our father to the VA hospital for a checkup. His doctor told him he should eat more fruit and vegetables. His response, leaving the doctor stunned, was,
“Look, I’m 84 years old!! Just how much longer do you expect me to live if I start eating that crap?”.
He’s still alive at 95, and hasn’t changed his diet. I suspect that heridity and natural variablity play a much larger factor than diet.
http://www.nationalreview.com/corner/416289/memories-pizza-just-memory-google-least-ian-tuttle