Are we really looking carefully enough?

Discuss your experiences with and ideas about Stardust@home here.

Moderators: Stardust@home Team, DustMods

h.d.thoreau
Posts: 10
Joined: Sun Aug 06, 2006 7:54 pm

Are we really looking carefully enough?

Post by h.d.thoreau »

I fear that too many people are only doing this for their rank. After all, is it not basic human nature to be greedy. It is not necessarily our fault, but rather a survival mechanism. After all, should you believe that life is governed by the overall rule that the fittest survive, in other worlds evolution, than it would make scientific sense that those who wanted to get that piece of meat from the other monkeys in their tribe would be more fit to reproduce. But we must remember that the reason we are doing this is
to learn about our universe, not to be of the top 100 'stardust viewers', as answering calibration movies correctly will not help to correctly identify interstellar dust. After all, is that not why we are here, to help scientists, not to impede them by flashing through the movies as fast as possible, possibly skipping over tracks. So please, take your time while looking at the movies. Don't even look at the ranking that often, because
remember, answering calibration movies correctly won't help to unlock the secrets of our solar system, as well as, possibly, the origin of life. Here are some examples of promising tracks that have been skipped over.

I ask you to please look at these very carefully, as many of these possible tracks are very small. Also, we can't expect every track to look exactly like those in the calibration movies. The calibration movies are the BEST possible examples that they could find in order to make sure that everyone really knows what to look for, but remember that we are looking for the unexpected, which might mean smaller tracks that what we are used to from the calibration movies. Remember, there is a big difference between dust shot out of an accelerator on earth, and dust collected in the harsh, cold hypersonic flight in 0g, in which this dust was collected. The dust may be embedded much farther in than in the calibration movies. This is likely because of the sheer speed of the dust particles, which is measured in 10s of thousands of miles per hour.

http://stardustathome.ssl.berkeley.edu/ ... =1912977V1
Image
81 views, marked 14 times

http://stardustathome.ssl.berkeley.edu/ ... =6898743V1
Image
86 views, marked only 23 times

this one is absolutely dead on:
http://stardustathome.ssl.berkeley.edu/ ... =5334186V1
Image
98 views with only 30 marks

http://stardustathome.ssl.berkeley.edu/ ... =3097428V1
Image
94 views, 28 flags

It actually looks like this one might have 3 tracks in it:
http://stardustathome.ssl.berkeley.edu/ ... d=612028V1
Image
97 views and only 10 flags

http://stardustathome.ssl.berkeley.edu/ ... =9164498V1
Image
97 views and only 21 flags

So, while I may be looking too closely at these pictures, for I actually care about the results of this mission (which cost the taxpayers a couple million by the way), I say that all who are serious about finding these grains of dust, which should be everybody participating, should slow down, and give each picture a couple more scans. In the end, those few seconds extra that you give may be the difference between making this mission a success or a waste of taxpayer money.
And finally, a word of advice, make sure that you set your monitor's resolution to 1024x768 at 32 bit, it makes the image appear bigger on your screen, so it is easier to see small details.
plexuz
Posts: 11
Joined: Thu Aug 10, 2006 7:51 am
Location: GTA, On, Canada

Post by plexuz »

Well, I totally with you here. I have very similar findings with very low agreement rate. However, to me they seem very realistic. I wonder if what we are finding are just artefacts?

Here are mine:
http://stardustathome.ssl.berkeley.edu/ ... =9337800V1
Image

http://stardustathome.ssl.berkeley.edu/ ... =4805525V1
Image

http://stardustathome.ssl.berkeley.edu/ ... =1079720V1
Image
icebike

Re: Are we really looking carefully enough?

Post by icebike »

h.d.thoreau wrote:
So, while I may be looking too closely at these pictures, for I actually care about the results of this mission (which cost the taxpayers a couple million by the way), I say that all who are serious about finding these grains of dust, which should be everybody participating, should slow down, and give each picture a couple more scans.
Some of us here have been begging for some guidance on this topic since the day 1. These are the so called "inclusions" which are seemingly dismissed by someone who's name I am forbidden to utter on the Internet.

I find a lot of them. Some deep enough that they never do resolve, others clearly come to a point if you focus deep enough, and still others that resolve to something quite a bit bigger than a single spec.

I keep marking them. Others keep ignoring them. I'm beginning to think its a huge waste of time, and maybe its best to ignore them.

It would be great to know if we should bother. Are they too small to recover? (Some are almost down on the micron level). Are they just bubbles or manufacturing defects? Are such defects common?

Are they embedded particles which entered from the other side of the tile, (is that possible)? Did the entry wound heal over the many years these tiles were in space? Did the return to atmospheric pressure close the channels again?

Did the existence of these specs escape the attention of mission planners when the areogel was selected for this project, or when the Training was being developed?

There are lots of questions. But two things are clear:

1) if these are recoverable space dust the project is already a roaring success, because there is a boatload of them, but to get them all will
take careful scanning.

2) if these are either not recoverable, or not space dust, we can proceed a hell of a lot faster


Maybe it take both Rabbits and Tortoises. Rabbits to run through the VMs like it was a video game, finding obvious hits (if any) which are then sent to 100s of others to verify. Tortoises to plod along and do a more complete job.
the moon
Posts: 177
Joined: Thu Aug 03, 2006 12:34 am

Post by the moon »

Relax guys, you're movies will get examined by the team. Out of the 23 movies I've marked, only 1 of them has more then a 25% agreement rate. So 21 out of 97 and scores like that will be pretty high on their list. It's not like the team is going to say "oh well, none of the movies have a 90% agreement rate so I'm gonna go home and take a nap."
icebike

Post by icebike »

the moon wrote:Relax guys, you're movies will get examined by the team. Out of the 23 movies I've marked, only 1 of them has more then a 25% agreement rate. So 21 out of 97 and scores like that will be pretty high on their list. It's not like the team is going to say "oh well, none of the movies have a 90% agreement rate so I'm gonna go home and take a nap."
You miss the point entirely.

Its not about whether THEY look at them, its about whether WE should continue to look for these tiny faint events.

Its not the amount of agreement, its the total disagreement about what we are to look for, and spend our time on.

I suspect 3/4 of all participants, and some of the dust mods have already decided to ignore these. Based on nothing but an offhand comment by she-who-can-not-be-named with no real investigation of the issue.

How many hundreds of volunteer hours could be saved or re-focused by simply examining some of these and telling us one way or another? While they are waiting for some arbitrary criteria to be met before digging out the high-rez images and looking at these faint events, the most careful volunteers are spending hundreds or thousands of hours marking them. They risk the disaffection of the best volunteers. What happens when they EVENTUALLY get around to looking at these and tell h.d.thoreau that he just wasted several weeks being careful?
the moon
Posts: 177
Joined: Thu Aug 03, 2006 12:34 am

Post by the moon »

Does it really take you that much longer to look for these little specks, as opposed to quickly scanning for the large tracks like in the examples? I don't think it would save me any time at all. You still have to look carefully.

Also I think you're missing the point of making the project available to the public. Since even the researchers aren't sure exactly what to look for, they made it sort of open ended in the hopes that 10,000 people searching in 10,000 different ways would find the most tracks. Because, until they can actually find some real tracks, and some false positives to use as examples there is no one best way to search for the things.
icebike

Post by icebike »

the moon wrote:Does it really take you that much longer to look for these little specks,
Yes. Dramatically longer. I could scan tor the fat-Albert types found in the CMs in under 2 seconds per movie. The little guys you have wiggle that mouse quite a bit to spot.

the moon wrote: Also I think you're missing the point of making the project available to the public. Since even the researchers aren't sure exactly what to look for,
It was my impression that when they looked at the prospect of 20 years of work to find one spec they had to find another way. At least that's what the web site said. Did I miss the part about even the professionals having no clue?

the moon wrote: Because, until they can actually find some real tracks, and some false positives to use as examples there is no one best way to search for the things.
They had tracks from experiments, and also russian information from Mir. They had a pretty good idea of what to look for.

I'm sure everyone is mystified by these tiny objects. So we all speculate... (pun intended)
benj1989
Posts: 18
Joined: Wed Aug 09, 2006 4:32 am
Location: Opwijk -Belgium

Post by benj1989 »

I think they are a default from the aerogel,
because I think they really know what they are looking for and what we should looking for, is it nasa or not? :D
but till we don't get confirmation about those little points I will click them :roll:
h.d.thoreau
Posts: 10
Joined: Sun Aug 06, 2006 7:54 pm

Post by h.d.thoreau »

If memory serves me right, the gel was exposed to the dust trail for several days, before returning to Earth. If so, there should be ALOT of dust specs, albeit small. Remember, the comet (or asteroid, I'm not sure) that the dust came from was supposedly formed millions, perhaps billions of years ago. Any dust which is trapped in aerogel would not neccesarily be huge. It has undergone alot of stress in its lifetime, and I atleast imagine the particles to be quite small from numerous interstellar impacts. Either way, I will continue to mark the small trails while many simply fly through looking for huge dust tracks like some found on earth under controlled conditions. I'd rather not miss even the smallest of possible dust tracks. Imagine if those were all dust particles, but not enough peolpe marked them, so they were wasted. What if one of those held the secret to the start of the solar system, or the universe. I don't want that to be wasted. Please, even if a track is small, and seemingly unimportant, it could end up being that new particle which revolutionizes science as we know it.
petterip
Posts: 16
Joined: Sun Aug 06, 2006 1:12 pm
Location: Finland

Post by petterip »

h.d.thoreau wrote:If memory serves me right, the gel was exposed to the dust trail for several days, before returning to Earth.
According to the article Interstellar Dust: The Hunt for the Building Blocks of the Universe by Amir Alexander in The Planetary Report vol. 26 number 4, the gel was exposed up to seven months. The article very thoroughly addressed the nature of the interstellar dust and the fact that the stream is so thin that not even the scientist expect there to be many traces in the Aerogel.

Additional details are found in the official Stardust mission homepage, especially in the article "Science-in-depth".

Some good points have been made in this topic. I agree with the guys who are taking this as an interesting opportunity to participate and - within the given resources - do the best we can helping the project to succeed.

In the movies which we are scanning, or "dusting" :wink: , not all grains of dust are even within the resolution. According to A. Alexander, "...these grains, all ranging from less than 1/100th of a micron to 10 microns in size". So the grains come in various sizes. And thus I would agree with icebike for there are both Rabbits and Tortoises, and everything in between.

Happy dusting, everyone :D

Edit: corrected a typo :oops:
Last edited by petterip on Mon Aug 14, 2006 12:09 pm, edited 1 time in total.
While not stardusting, I'd rather be driving with my Alfa.
elainekeefe
Posts: 190
Joined: Fri Aug 04, 2006 4:38 am
Location: Massachusetts, USA

Are we really looking carefully enough?

Post by elainekeefe »

Interestingly, h.d.thoreau, I have the third movie you've pictured, #5334186V1, twice on my events list. Obviously, I agreed with you there! (Not once, but twice!)
Betelgeuze
Posts: 19
Joined: Sun Aug 06, 2006 1:24 am
Location: Belgium

Post by Betelgeuze »

I want to add:
http://stardustathome.ssl.berkeley.edu/ ... =2090529V1
Image

27 viewings and Im the only one who has seen it so far! :? :x
Please tell me you guys see it!
PocketLint
Posts: 20
Joined: Mon Aug 07, 2006 1:02 pm

What?

Post by PocketLint »

Of all the posted movies in this thread, NONE of them even come close to what we were trained for, well except for one... I'll let you decide which one I think it is after you read my observations.

And I have to say, this is just MY observations. I post these comments only because you imply that somehow we all as a group are so involved in ranking that we are missing the possible tracks.
I have to differ, I think some people are so juiced to be the first to find something that they are clicking on anything that is below surface without actually giving thought to the process and the posted examples give testament to that thought.

Heck I could have written a pattern matching program that could do that. Heck about a million other people here could have done that. The operators of this project do not just want passable possible hits. They want realistic hits. In my opinion.

I'm willing to explore a few of the examples given and explain why I wold have rejected them as possible hits. Again, remember these are just my interpretations of the training movies and calibration movies.

Take this one for example..
http://stardustathome.ssl.berkeley.edu/ ... =5334186V1
It is commented as DEAD ON..

If you look closely, you will see there are TWO other pings coming into focus at the same rate as the one highlighted. (lower right corner for example).
The odds of TWO to THREE out of the estimated 45 total tracks being in the same movie makes this a no go for me.
Plus.. It is coming into focus with no shadows. A true track THAT large, as I have been trained to see, will have a shadow as you track the focus. A true shadow is not just the object comming into focus. The shadow is the track walls that will come into focus as the current ring goes out. The end result will seem to be a continuous ring in focus for more than one frame of the movie with a shadow still in view.


Lets take this one..
http://stardustathome.ssl.berkeley.edu/ ... =1912977V1

At first glance, I thought that might be a track but once I spent a few seconds looking at it I would have to say again, no track. I would however had clicked on bad focus for this one.
It at first seems to have the shadow, but if you look closely you will see that is an artifact of compression. With the gel having folds in it, and the other blips on the higher levels of the gel the folds coming into focus first, I would have to say that the blip of interest here is really on the surface of the true gel lowest level. Notice how a few smaller blips are also starting to come into focus at the same time in the same patch of space. Also, in the last two frames of the movie, the upper half of the ring is going out of focus. For this one, I would want a re-scan of it with a better surface focus range set.

http://stardustathome.ssl.berkeley.edu/ ... =3097428V1
Notice how ALL the blips in this movie all have trails bleeding to the bottom left. The final blip that come into focus has that same trait. I think in this movie the gel has a gradient and that is all we are seeing here.


These are my observations based on the training and calibration movies and extrapolated rules.

All that I'm saying is don't be so quick to group people in such a derogatory way. While I've been in the top 10 and still have viewed over 10,000 movies and have a ratio of 99.98 and 99.15 that in no way implies that I have not given study to EACH and EVERY movie that I viewed.


BTW: Which movie, of the ones posted in this thread do you now think I consider a possible hit if you apply my logic to them?
icebike

Re: What?

Post by icebike »

PocketLint wrote:
And I have to say, this is just MY observations. I post these comments only because you imply that somehow we all as a group are so involved in ranking that we are missing the possible tracks.
Because you didn't use the Quote button we have no idea to whom you are replying, and no idea who this "you" is.
PocketLint wrote:
I have to differ, I think some people are so juiced to be the first to find something that they are clicking on anything that is below surface without actually giving thought to the process and the posted examples give testament to that thought.
True, people are over anxious. However, in the initial training no examples of embedded particles were given. Its like who ever created the training was unaware of their existence, and had not looked at very many movies prior to preparing the training.

The training did show tracks that started sub-surface, and continued downward. Few if any of these tracks showed any "entrance wounds."

This led many to believe that subsurface particles, with or without entrance holes, were suspect.

Secondly: The initial training also did not give very good examples of sideways entry, (angled entry) where the tracks run sideways, and the particle is visible. They have since been improving the training, and its worth re-taking the training occasionally.

Finally: some particles have been flagged in users event list as having been accepted for level 1 review. At least some of these are nothing at all like the training movies.
http://stardustathome.ssl.berkeley.edu/ ... =6193890V1
http://stardustathome.ssl.berkeley.edu/ ... =5262416V1

Point being: It would seem that the initial training was not particularly accurate, as no one has found anything that looks remotely like the huge tracks shown there, and first review is accepting things that also bear no resemblance to the training.

So rather than no thought being given to what is being clicked, I think its just the opposite. Volunteers are thinking ahead of the scientists (correctly or incorrectly), and reasoning for themselves that subsurface particles need to be investigated, until some clear word comes from on high to the contrary.

Further, the first review team seems to agree with them in a few cases.
PocketLint
Posts: 20
Joined: Mon Aug 07, 2006 1:02 pm

agree

Post by PocketLint »

You are echoing my point exactly I think in the end.

I was taking offence to the original posters taint that somehow, people that view a large number of calibration movies, IE: are up there in the rankings, are some how deficient in their ability to detect possible tracts.

All that I was trying to say was, "do not be so narrow minded". I gave examples as to why I would not have considered the given examples as possible tracks.

My personal rules for excluding possible tracks are probably wrong. But if you go by the numbers, then I just might have picked up on the thing the training movies where trying to impart. Ok, I know they really don't have an exact idea on what a real track will look like but they have an educated guess. they created the training movies from that guess. they explained what a track is what we might see as a tell-tail sign and I extrapolated from that.
I'm not in any way saying that I'm correct.

What I'm saying is... Don't tell me that I'm somehow in it for the rankings. That I'm not doing my job or I'm doing a disservice to the project just because I can look at a movie and within a few seconds tell if it has what I would consider a possible track or not.

Just because it takes someone 30 minutes to decide that there might be a track and I can look at it, run the probabilities and make a judgement in 3 seconds in no way means that I 'skipped over' them as the original poster comments.

I saw them, I evaluated them, I rejected them OR validated them as I saw fit.

My point in the end is.. My observation are no more valid NOR invalid as anyone person's observations. It IS the rankings of all of participants and how they interpreted the movies that will decide on the order that the possible hits gets evaluated and that is all there is to it.




BTW: without a quote, 'YOU' would infer the to the author of the post I would think.
Post Reply