{ "id": 35153, "title": "An idealistic suggestion to promote international collaboration", "url": "/competitions/passenger-screening-algorithm-challenge/discussion/35153", "author_name": "Heads or Tails", "post_date": "2017-06-22T23:28:52.684000", "votes": 0, "comment_count": 5, "views": 0, "content": "

I'm far from being competitive in this challenge regardless of citizenship and I have joined Kaggle only recently. Having read most of the comments in this discussion I feel like instead of being angry we could turn the (understandable) frustration into something positive for the Kaggle community and for the international collaboration between (aspiring) data scientists of all nations.

\n\n

Elsewhere, it has been suggested that the US winners could donate the prize money to make a statement that we succeed or fail as a world-wide community, not as individuals or single countries (the latter part mostly my interpretation). Surely, if we would be able to arrange something like this then we might as well aim for something bigger?

\n\n

The prize money is substantial, so why not use it to organise an international Kaggle Community Conference? Sounds naive and crazy? Hear me out, and hopefully it will sound just naive afterwards ;-)

\n\n

These are the main ideas:

\n\n\n\n

Would it be difficult? Yes. There are potentially lots of visas, travel time restrictions, personal preferences, and much more to be taken care of. Would it be impossible? I don't think so. A couple of 100 people aren't that difficult to gather. Would it be fun? Hell, yes :-)

\n\n

I'm aware that it still sounds a bit naive. After all, the most important requirement would be that the winner(s) agree, voluntary and without any pressure, to donate the money for this cause. But if we as a community could pull off such a thing, and turn frustration into a worthwhile challenge, then maybe things aren't so bad after all in terms of fostering collaboration between a truly international group of data science enthusiasts.

\n\n

Who's with me?

", "messages": [ { "id": 195205, "postDate": "2017-06-22T23:28:52.683Z", "content": "

I'm far from being competitive in this challenge regardless of citizenship and I have joined Kaggle only recently. Having read most of the comments in this discussion I feel like instead of being angry we could turn the (understandable) frustration into something positive for the Kaggle community and for the international collaboration between (aspiring) data scientists of all nations.

\n\n

Elsewhere, it has been suggested that the US winners could donate the prize money to make a statement that we succeed or fail as a world-wide community, not as individuals or single countries (the latter part mostly my interpretation). Surely, if we would be able to arrange something like this then we might as well aim for something bigger?

\n\n

The prize money is substantial, so why not use it to organise an international Kaggle Community Conference? Sounds naive and crazy? Hear me out, and hopefully it will sound just naive afterwards ;-)

\n\n

These are the main ideas:

\n\n\n\n

Would it be difficult? Yes. There are potentially lots of visas, travel time restrictions, personal preferences, and much more to be taken care of. Would it be impossible? I don't think so. A couple of 100 people aren't that difficult to gather. Would it be fun? Hell, yes :-)

\n\n

I'm aware that it still sounds a bit naive. After all, the most important requirement would be that the winner(s) agree, voluntary and without any pressure, to donate the money for this cause. But if we as a community could pull off such a thing, and turn frustration into a worthwhile challenge, then maybe things aren't so bad after all in terms of fostering collaboration between a truly international group of data science enthusiasts.

\n\n

Who's with me?

", "rawMarkdown": "I'm far from being competitive in this challenge regardless of citizenship and I have joined Kaggle only recently. Having read most of the comments in this discussion I feel like instead of being angry we could turn the (understandable) frustration into something positive for the Kaggle community and for the international collaboration between (aspiring) data scientists of all nations.\n\nElsewhere, it has been suggested that the US winners could donate the prize money to make a statement that we succeed or fail as a world-wide community, not as individuals or single countries (the latter part mostly my interpretation). Surely, if we would be able to arrange something like this then we might as well aim for something bigger?\n\nThe prize money is substantial, so why not use it to organise an international *Kaggle Community Conference*? Sounds naive and crazy? Hear me out, and hopefully it will sound just naive afterwards ;-)\n\nThese are the main ideas:\n\n- the prize money should be sufficient to pay for a decent conference centre in a well-connected but not too expensive city for a few 100 people for a couple of days\n\n- the top 10-ish get invited with all expenses paid\n\n- the top 50-ish get free flights\n\n- the top 100-ish get free meals\n\n- the conference would need to be planned well in advance, reducing the costs. A location in Europe might be best for overall visa requirements and flight times\n\n- I noticed that there exist already local kaggle meetups in several locations, like Paris, Berlin, or London. The existing organisational infrastructure could be used to make the planning significantly easier.\n\n- the Kaggle team takes care of the organisation, call it a gesture of good will :-), with optional volunteering help of local Kagglers\n\n- a decently sized and very international group of top Kagglers meet for a couple of days to exchange ideas and discuss future collaborations, projects, and careers\n\nWould it be difficult? Yes. There are potentially lots of visas, travel time restrictions, personal preferences, and much more to be taken care of. Would it be impossible? I don't think so. A couple of 100 people aren't that difficult to gather. Would it be fun? Hell, yes :-)\n\nI'm aware that it still sounds a bit naive. After all, the most important requirement would be that the winner(s) agree, voluntary and without any pressure, to donate the money for this cause. But if we as a community could pull off such a thing, and turn frustration into a worthwhile challenge, then maybe things aren't so bad after all in terms of fostering collaboration between a truly international group of data science enthusiasts.\n\nWho's with me?" }, { "id": 196375, "postDate": "2017-06-27T07:38:04.513Z", "content": "

Hai trovato il posto giusto per questa fiera ? Milano sarebbe ideale. Mi piace molto l'idea ;)

", "rawMarkdown": "Hai trovato il posto giusto per questa fiera ? Milano sarebbe ideale. Mi piace molto l'idea ;)" }, { "id": 195294, "postDate": "2017-06-23T07:04:37.507Z", "rawMarkdown": "", "votes": 5, "isDeleted": true, "replies": [ { "id": 195371, "postDate": "2017-06-23T12:23:21.767Z", "content": "
\n

now your changes are 1 in 100 :) for $500 K or 8 in 100 for $100 K,\n beats lottery scratchers. Maybe, just maybe its worth a go and small\n investment in a graphics card or Google Cloud service.

\n
\n\n

That's a naive way to see things. The places are not randomly distributed and you can fully expect a hard fight for top 8, a fight involving heavy use of computational resources. We know that the winning models are going to be CNNs, possibly 3D, and possibly ensembled. The development/training costs of such models can be comparable to the potential prize money.

", "rawMarkdown": "> now your changes are 1 in 100 :) for $500 K or 8 in 100 for $100 K,\n> beats lottery scratchers. Maybe, just maybe its worth a go and small\n> investment in a graphics card or Google Cloud service.\n\nThat's a naive way to see things. The places are not randomly distributed and you can fully expect a hard fight for top 8, a fight involving heavy use of computational resources. We know that the winning models are going to be CNNs, possibly 3D, and possibly ensembled. The development/training costs of such models can be comparable to the potential prize money." }, { "id": 195406, "postDate": "2017-06-23T14:14:11.547Z", "content": "

If past competitions are any example, then it is clear that price of developing top-10 solution is nowhere near 100k$ (if we don't count all participants' time and potential cost of all knowledge generated and shared across forum and kernels). It is more likely in range of 2-3k$ which will buy you an excellent DevBox with couple 1080ti. Most precious resource in this case isn't computing power but time to deal with all those engineering issues and brain power to generate clever ideas.

", "rawMarkdown": "If past competitions are any example, then it is clear that price of developing top-10 solution is nowhere near 100k$ (if we don't count all participants' time and potential cost of all knowledge generated and shared across forum and kernels). It is more likely in range of 2-3k$ which will buy you an excellent DevBox with couple 1080ti. Most precious resource in this case isn't computing power but time to deal with all those engineering issues and brain power to generate clever ideas." }, { "id": 195428, "postDate": "2017-06-23T15:07:36.633Z", "content": "

I appreciate the comments and the differing opinions. However, I think that they somewhat miss to address the big-picture issue here: how a global community that values its international spirit in the pursuit of knowledge can address a frustrating problem such as pretty arbitrary eligibility rules in a high-profile competition. Kaggle is defined by its user base, and instead of getting angry about these eligibility rules (or ignoring the competition altogether) the user base could decide to turn the situation into a positive outcome for everyone, regardless of nationality.

\n\n

Of course, anyone who will get close to the prize money will have thoroughly earned it through hard and persistent work. No question about that. Keeping the money makes perfect sense. And if those that are eligible don't mind the arbitrary advantage they have been given then they may well run with it. I don't have an issue with that either. I just think that if there is enough of a will, then we as a community could do much better than that. I was trying to sketch a constructive solution to a situation that many seem to regard as a problem. That's all.

", "rawMarkdown": "I appreciate the comments and the differing opinions. However, I think that they somewhat miss to address the big-picture issue here: how a global community that values its international spirit in the pursuit of knowledge can address a frustrating problem such as pretty arbitrary eligibility rules in a high-profile competition. Kaggle is defined by its user base, and instead of getting angry about these eligibility rules (or ignoring the competition altogether) the user base could decide to turn the situation into a positive outcome for everyone, regardless of nationality.\n\nOf course, anyone who will get close to the prize money will have thoroughly earned it through hard and persistent work. No question about that. Keeping the money makes perfect sense. And if those that are eligible don't mind the arbitrary advantage they have been given then they may well run with it. I don't have an issue with that either. I just think that if there is enough of a will, then we as a community could do much better than that. I was trying to sketch a constructive solution to a situation that many seem to regard as a problem. That's all." } ] } ], "comments": [ { "id": 196375, "author_name": "mezoganet", "author_url": "", "post_date": "2017-06-27T07:38:04.513000", "content": "

Hai trovato il posto giusto per questa fiera ? Milano sarebbe ideale. Mi piace molto l'idea ;)

", "votes": 0, "replies": [] }, { "id": 195294, "author_name": "", "author_url": "", "post_date": "2017-06-23T07:04:37.507000", "content": "", "votes": 5, "replies": [ { "id": 195371, "author_name": "redstr", "author_url": "", "post_date": "2017-06-23T12:23:21.767000", "content": "
\n

now your changes are 1 in 100 :) for $500 K or 8 in 100 for $100 K,\n beats lottery scratchers. Maybe, just maybe its worth a go and small\n investment in a graphics card or Google Cloud service.

\n
\n\n

That's a naive way to see things. The places are not randomly distributed and you can fully expect a hard fight for top 8, a fight involving heavy use of computational resources. We know that the winning models are going to be CNNs, possibly 3D, and possibly ensembled. The development/training costs of such models can be comparable to the potential prize money.

", "votes": 0, "replies": [] }, { "id": 195406, "author_name": "Cut Onion", "author_url": "", "post_date": "2017-06-23T14:14:11.547000", "content": "

If past competitions are any example, then it is clear that price of developing top-10 solution is nowhere near 100k$ (if we don't count all participants' time and potential cost of all knowledge generated and shared across forum and kernels). It is more likely in range of 2-3k$ which will buy you an excellent DevBox with couple 1080ti. Most precious resource in this case isn't computing power but time to deal with all those engineering issues and brain power to generate clever ideas.

", "votes": 0, "replies": [] }, { "id": 195428, "author_name": "Heads or Tails", "author_url": "", "post_date": "2017-06-23T15:07:36.633000", "content": "

I appreciate the comments and the differing opinions. However, I think that they somewhat miss to address the big-picture issue here: how a global community that values its international spirit in the pursuit of knowledge can address a frustrating problem such as pretty arbitrary eligibility rules in a high-profile competition. Kaggle is defined by its user base, and instead of getting angry about these eligibility rules (or ignoring the competition altogether) the user base could decide to turn the situation into a positive outcome for everyone, regardless of nationality.

\n\n

Of course, anyone who will get close to the prize money will have thoroughly earned it through hard and persistent work. No question about that. Keeping the money makes perfect sense. And if those that are eligible don't mind the arbitrary advantage they have been given then they may well run with it. I don't have an issue with that either. I just think that if there is enough of a will, then we as a community could do much better than that. I was trying to sketch a constructive solution to a situation that many seem to regard as a problem. That's all.

", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "195205": "I'm far from being competitive in this challenge regardless of citizenship and I have joined Kaggle only recently. Having read most of the comments in this discussion I feel like instead of being angry we could turn the (understandable) frustration into something positive for the Kaggle community and for the international collaboration between (aspiring) data scientists of all nations.\n\nElsewhere, it has been suggested that the US winners could donate the prize money to make a statement that we succeed or fail as a world-wide community, not as individuals or single countries (the latter part mostly my interpretation). Surely, if we would be able to arrange something like this then we might as well aim for something bigger?\n\nThe prize money is substantial, so why not use it to organise an international *Kaggle Community Conference*? Sounds naive and crazy? Hear me out, and hopefully it will sound just naive afterwards ;-)\n\nThese are the main ideas:\n\n- the prize money should be sufficient to pay for a decent conference centre in a well-connected but not too expensive city for a few 100 people for a couple of days\n\n- the top 10-ish get invited with all expenses paid\n\n- the top 50-ish get free flights\n\n- the top 100-ish get free meals\n\n- the conference would need to be planned well in advance, reducing the costs. A location in Europe might be best for overall visa requirements and flight times\n\n- I noticed that there exist already local kaggle meetups in several locations, like Paris, Berlin, or London. The existing organisational infrastructure could be used to make the planning significantly easier.\n\n- the Kaggle team takes care of the organisation, call it a gesture of good will :-), with optional volunteering help of local Kagglers\n\n- a decently sized and very international group of top Kagglers meet for a couple of days to exchange ideas and discuss future collaborations, projects, and careers\n\nWould it be difficult? Yes. There are potentially lots of visas, travel time restrictions, personal preferences, and much more to be taken care of. Would it be impossible? I don't think so. A couple of 100 people aren't that difficult to gather. Would it be fun? Hell, yes :-)\n\nI'm aware that it still sounds a bit naive. After all, the most important requirement would be that the winner(s) agree, voluntary and without any pressure, to donate the money for this cause. But if we as a community could pull off such a thing, and turn frustration into a worthwhile challenge, then maybe things aren't so bad after all in terms of fostering collaboration between a truly international group of data science enthusiasts.\n\nWho's with me?", "196375": "Hai trovato il posto giusto per questa fiera ? Milano sarebbe ideale. Mi piace molto l'idea ;)", "195294": "" } }