Good work, Daniel. A laboriously-gathered overview of current practice, and discussion of how to determine whether uses of illegally-obtained data are justified.
That's a different field entirely. Just because some American hospitals have trouble organising medical ethics reviews, does not mean that European or even American CompSci researchers will run into the same problems. Indeed, Daniel is (I think) positive perhaps based on good interactions with the ethics committee here (my wife is certainly happy with the ethics reviews she's had).
Even in that SSC discussion linked, many SSC commenters agree that their own ethics system was far easier, even in medicine.
You seem to be extrapolating from an extremely small sample, and one that mostly makes fun of things that would exist (eventually) in any such process. It's not really a useful rebuttal, and you probably should not draw strong conclusions from it.
To anyone who is interested in this topic, I published a paper a few years back on the topic of 'accidentally illicit' datasets. It's not my best work but someone might find it interesting.
Amusing anecdote: I wrote this paper (2) after a reviewer insisted they would reject my other paper (1) unless I tested the paper's algorithm on 10% of all images on the Internet.
Talk about a hostile review!
(They also demanded I remove the performance comparison which showed the new technique to be some 1000x faster than existing techniques and more reliable... Hmm).
The reviewer then dragged out the review/response process for so long that I had time to write/review/publish the ethics paper above, in between one round of reviewer comments (!)
I then took the freshly published ethics paper to the editor for (1), and asked them to disqualify the hostile reviewer for making unethical demands and refusing to withdraw them even when this was pointed out.
The editor agreed. The reviewer was then replaced by someone else, who replicated the entire work of (1) completely from scratch using only the description in the paper, confirmed the result using their own datasets they gathered privately, and who approved publication.
'Reviewer 1', they're always either the hero or the villain. It was an interesting feeling to see the very worst type of reviewer being replaced by the very best.
Anyway, that's the strange story behind this paper :-)
1) The reviewer may have had similar work and wanted to hold up the publishing of OP’s paper while finishing their own work. Or there’s some personal animus. Or OP’s paper may have threatened to surplant the reviewers work. There’s a whole plethora of reason for a particularly hostile review. It’s a big enough problem that my partner who’s going through grad school now had the option of requesting specific people NOT be on the reviews for her (first publishing in gradschool as a first author!) paper because of scooping, animus towards her PI, etc. that may have resulted in an overly hostile review.
1) My first publication in the topic, iirc. The only reason I can see them doing this when all the other reviews were good was eg. They have grant applications underway and feel undermined, or they are petty and don't want their own technique to look bad etc. (the extreme roadblocking and 'I won't let you compare performance or accuracy against other techniques!!' is a little telling imho...)
There are a lot of people out there who think a field of research and all the money in it belongs to them.
And there's some angry people who have effectively no power or influence over their own everyday life that act like little tyrants when they get to review papers.
(fwiw, I coauthored (1) without any grant support, just for fun, and don't think I ever applied for a grant in that area? I can't remember).
Who knows? Trying to figure out hostile reviewers is a guessing game and not very helpful for your own happiness
2a) travelling on holiday, posting from my phone.
And
2b) this has personal identifying info and I can't remember everything I've ever posted on my other account right now, (or even the u/p for my desktop account, haven't posted in months)
P.s. Glad you enjoyed it, I've been looking for an excuse to mention that story for years :-)
There is a very fine line between authorized data, technically public but implicitly unauthorized data, and illegally obtained, unauthorized data. Here’s an example of each in the financial sector, from my personal experience:
1. Financial account aggregators and “budget apps” like Min monetize their business, in part, by selling huge amounts of data to the financial sector. Sometimes companies like Second Measure take raw data from companies like Yodlee and clean it, then resell it. Nowadays there is an entire industry of alterative market research that has had all sorts of participants, from Foursquare (locations) to Spark (email enhancement). This is technically authorized, because it’s in the TOS. The users effectively contribute their own data.
2. I developed an extremely accurate, reasonably generalizable method of forecasting vehicle production at several companies that relies on implementing a VIN searching algorithm in conjunction with legally required NHTSA recall lookup portals hosted by each manufacturer. This data is what you’d call unauthorized, because no entity explicitly endorses your use of it. For example, several colleagues and I knew well ahead of time that Tesla would miss on production of the Model 3s because they were utterly unrepresented in our data. But this data is public, so it’s fine to use from a legal and compliance standpoint. It was lucrative data specifically because it had a high signal for revenue, yet was hitherto unused and unidentified.
3. I once found, in the course of looking for legally usable data, an actual security vulnerability disclosing all users of a publicly traded QSR’s online delivery service, along with their phone numbers, email addresses and last four digits of credit cards. This is both unauthorized and illegal, because the data is contaminated with personally identifiable information and it clearly requires a vulnerability (not just scraping) to acquire.
I’ve seen overzealous data vendors accidentally slip from #2 into #3, which is really bad for all concerned. It’s not a great look for the vendor, who will likely be fired, and it represents a breach for the company who owns the data and its users. Any firm that has purchased the data will likely be contamined and be forced into a trading lockdown of that security for a period of time by compliance.
My real concern is that illicit data like this is used in machine learning research. Machine learning is already pretty frustrating - it’s common for me to find research from a conference that I’m simply unable to replicate because the training or experiment data is not available (this is annoyingly the case with A/B experiment optimization research put out by giant companies in particular). I worry that this trend of accepting machine learning research without any requirement for total data transparency will incentivize researchers to conduct their experiments using illicit data that doesn’t need to be sourced.
I believe dsacco's posted about doing similar work before. Apparently there are teams of people at hedge funds that comb the web to find these kinds of datasets - non-obvious signals for financial metrics. Very eye-opening stuff. I always find his/her anecdotes interesting.
We gave the Japanese amnesty for https://en.wikipedia.org/wiki/Unit_731 in exchange for their biological warfare data. We had the option to do the right thing (burn it, put it on display), but we validated it.
> We had the option to do the right thing (burn it, put it on display), but we validated it.
I'm not sure complete amnesty was the best thing to do, but I don't think destroying the data would be the best either. If we had burned the data that was already gathered, it would have been a complete waste of human life. Those people experimented on were murdered, and burning the data would have made those murders pointless. By using that data, we could prevent human death in the future.
If I remember correctly, there was also a similar set of unethical decisions in Nazi Germany that generated a body of knowledge of how humans survive under "situations", like lack of oxygen, rapid-(de)pressurization, decompression, % body burned and live. Just a whole lot of horrible things.
The problem was, this was done scientifically. Controls, and all that. The problem was they were intentionally treating their subjects (Jewish) like lab rats.
But when WWII ended, we realized some of their data could be used for the space race. And the US used it.
The ethical issues here are all encompassing. Do you get rid of it and let them die in vain? Do you use it, and benefit on people executed over genocide conditions? There's no good way to answer this, as the converse has a valid argument as well.
I look back and say, well, I was born in the early 80's. It's about 60 years before my time (would have to be born in the 20's to serve in the 40's).
Those same perverse incentives also power the perpetrators. We absolutely need to destroy all ill gotten gains without question. All deaths are pointless, even our own when lived out naturally. But deaths that power a vile knowledge, that knowledge cannot, must not live. Because when you allow it, even for an instant, you make it ok. And it will happen again, and the perpetrator will feel like they are doing good. And the receiver will feel like they are respecting the life given, but in fact it is a perfectly crafted crime. The gains must destroyed, every time, without question.
phew that was one the darkest, bleakest wikipedia entries I've ever read. Does anyone know what it was about the Japanese army in WWII that made them so particularly vicious?
20 comments
[ 4.6 ms ] story [ 46.4 ms ] threadhttp://slatestarcodex.com/2017/08/29/my-irb-nightmare/
HN discussion of the above
https://news.ycombinator.com/item?id=15127271
Even in that SSC discussion linked, many SSC commenters agree that their own ethics system was far easier, even in medicine.
To anyone who is interested in this topic, I published a paper a few years back on the topic of 'accidentally illicit' datasets. It's not my best work but someone might find it interesting.
http://firstmonday.org/ojs/index.php/fm/article/view/2739/24...
Amusing anecdote: I wrote this paper (2) after a reviewer insisted they would reject my other paper (1) unless I tested the paper's algorithm on 10% of all images on the Internet.
Talk about a hostile review!
(They also demanded I remove the performance comparison which showed the new technique to be some 1000x faster than existing techniques and more reliable... Hmm).
The reviewer then dragged out the review/response process for so long that I had time to write/review/publish the ethics paper above, in between one round of reviewer comments (!)
I then took the freshly published ethics paper to the editor for (1), and asked them to disqualify the hostile reviewer for making unethical demands and refusing to withdraw them even when this was pointed out.
The editor agreed. The reviewer was then replaced by someone else, who replicated the entire work of (1) completely from scratch using only the description in the paper, confirmed the result using their own datasets they gathered privately, and who approved publication.
'Reviewer 1', they're always either the hero or the villain. It was an interesting feeling to see the very worst type of reviewer being replaced by the very best.
Anyway, that's the strange story behind this paper :-)
1) What would be the possible motivation for such hostility from the first reviewer?
2) Why did you create a temporary throwaway account but then promote a paper with your real name and information?
There are a lot of people out there who think a field of research and all the money in it belongs to them.
And there's some angry people who have effectively no power or influence over their own everyday life that act like little tyrants when they get to review papers.
(fwiw, I coauthored (1) without any grant support, just for fun, and don't think I ever applied for a grant in that area? I can't remember).
Who knows? Trying to figure out hostile reviewers is a guessing game and not very helpful for your own happiness
2a) travelling on holiday, posting from my phone.
And
2b) this has personal identifying info and I can't remember everything I've ever posted on my other account right now, (or even the u/p for my desktop account, haven't posted in months)
P.s. Glad you enjoyed it, I've been looking for an excuse to mention that story for years :-)
1. Financial account aggregators and “budget apps” like Min monetize their business, in part, by selling huge amounts of data to the financial sector. Sometimes companies like Second Measure take raw data from companies like Yodlee and clean it, then resell it. Nowadays there is an entire industry of alterative market research that has had all sorts of participants, from Foursquare (locations) to Spark (email enhancement). This is technically authorized, because it’s in the TOS. The users effectively contribute their own data.
2. I developed an extremely accurate, reasonably generalizable method of forecasting vehicle production at several companies that relies on implementing a VIN searching algorithm in conjunction with legally required NHTSA recall lookup portals hosted by each manufacturer. This data is what you’d call unauthorized, because no entity explicitly endorses your use of it. For example, several colleagues and I knew well ahead of time that Tesla would miss on production of the Model 3s because they were utterly unrepresented in our data. But this data is public, so it’s fine to use from a legal and compliance standpoint. It was lucrative data specifically because it had a high signal for revenue, yet was hitherto unused and unidentified.
3. I once found, in the course of looking for legally usable data, an actual security vulnerability disclosing all users of a publicly traded QSR’s online delivery service, along with their phone numbers, email addresses and last four digits of credit cards. This is both unauthorized and illegal, because the data is contaminated with personally identifiable information and it clearly requires a vulnerability (not just scraping) to acquire.
I’ve seen overzealous data vendors accidentally slip from #2 into #3, which is really bad for all concerned. It’s not a great look for the vendor, who will likely be fired, and it represents a breach for the company who owns the data and its users. Any firm that has purchased the data will likely be contamined and be forced into a trading lockdown of that security for a period of time by compliance.
My real concern is that illicit data like this is used in machine learning research. Machine learning is already pretty frustrating - it’s common for me to find research from a conference that I’m simply unable to replicate because the training or experiment data is not available (this is annoyingly the case with A/B experiment optimization research put out by giant companies in particular). I worry that this trend of accepting machine learning research without any requirement for total data transparency will incentivize researchers to conduct their experiments using illicit data that doesn’t need to be sourced.
I'm not sure complete amnesty was the best thing to do, but I don't think destroying the data would be the best either. If we had burned the data that was already gathered, it would have been a complete waste of human life. Those people experimented on were murdered, and burning the data would have made those murders pointless. By using that data, we could prevent human death in the future.
The problem was, this was done scientifically. Controls, and all that. The problem was they were intentionally treating their subjects (Jewish) like lab rats.
But when WWII ended, we realized some of their data could be used for the space race. And the US used it.
The ethical issues here are all encompassing. Do you get rid of it and let them die in vain? Do you use it, and benefit on people executed over genocide conditions? There's no good way to answer this, as the converse has a valid argument as well.
I look back and say, well, I was born in the early 80's. It's about 60 years before my time (would have to be born in the 20's to serve in the 40's).