{ "id": 35487, "title": "idle_spec for the win?", "url": "/competitions/passenger-screening-algorithm-challenge/discussion/35487", "author_name": "", "post_date": "2017-06-29T13:18:29.836509600Z", "votes": 16, "comment_count": 14, "views": 0, "content": "

Knowing all the circumstances, I can already place my bets idle_speculation will get ez #1 and ez money in this competition :) good luck for others

", "messages": [ { "id": "197370", "postDate": "06/29/2017 13:18:29", "content": "

Knowing all the circumstances, I can already place my bets idle_speculation will get ez #1 and ez money in this competition :) good luck for others

", "rawMarkdown": "Knowing all the circumstances, I can already place my bets idle_speculation will get ez #1 and ez money in this competition :) good luck for others", "votes": null }, { "id": "197508", "postDate": "06/29/2017 18:07:13", "content": "

Remember that there is a second stage, with a \"data-blind\"; i.e. the current leaderboard could be very deceptive.

", "rawMarkdown": "Remember that there is a second stage, with a \"data-blind\"; i.e. the current leaderboard could be very deceptive.", "votes": null }, { "id": "197830", "postDate": "06/30/2017 09:43:12", "content": "

You seem to antagonize people here ;)

\n\n

But this seems likely, idle_speculation is the only US one among top 10 Kagglers.

\n\n

Except that giba is now located in the US too ;) May not be a permanent resident yet though.

", "rawMarkdown": "You seem to antagonize people here ;)\n\nBut this seems likely, idle_speculation is the only US one among top 10 Kagglers.\n\nExcept that giba is now located in the US too ;) May not be a permanent resident yet though.", "votes": null }, { "id": "197900", "postDate": "06/30/2017 13:13:17", "content": "

good luck with that!!

", "rawMarkdown": "good luck with that!!", "votes": null }, { "id": "198019", "postDate": "06/30/2017 18:44:32", "content": "

I have an idea: Kaggle should set up another competition where only non-US residents are allowed to compete, and the objective is to predict how well will the US residents do in this competition.

", "rawMarkdown": "I have an idea: Kaggle should set up another competition where only non-US residents are allowed to compete, and the objective is to predict how well will the US residents do in this competition.", "votes": null }, { "id": "198056", "postDate": "06/30/2017 21:25:47", "content": "

Several companies and top CS universities have formed teams, specifically for this competition consisting of only US citizens/residents. I personally know of a lab at MIT that formed a team specifically for this contest. Anything could happen!

", "rawMarkdown": "Several companies and top CS universities have formed teams, specifically for this competition consisting of only US citizens/residents. I personally know of a lab at MIT that formed a team specifically for this contest. Anything could happen!", "votes": null }, { "id": "198063", "postDate": "06/30/2017 22:03:15", "content": "

Until academic research teams decide to enter in the game ... then anything may happen.

\n\n

I red some blog posts where winners are academics who focus only on one competition (Data Science Bowl 2017 for the most recent or one of the first applications of deep NN here )

\n\n

I am sure some researchers from Willow at Ecole Normale Supérieure in Paris, would be eager to participate in this competition if the rules were not so unfavorable to non-US citizens

", "rawMarkdown": "Until academic research teams decide to enter in the game ... then anything may happen.\n\nI red some blog posts where winners are academics who focus only on one competition (Data Science Bowl 2017 for the most recent or one of the [first][1] applications of [deep NN][2] here )\n\nI am sure some researchers from [Willow ][3] at Ecole Normale Supérieure in Paris, would be eager to participate in this competition if the rules were not so unfavorable to non-US citizens\n\n\n [1]: http://blog.kaggle.com/2012/11/01/deep-learning-how-i-did-it-merck-1st-place-interview/\n [2]: http://blog.kaggle.com/2012/10/31/merck-competition-results-deep-nn-and-gpus-come-out-to-play/\n [3]: http://www.di.ens.fr/willow/", "votes": null }, { "id": "258333", "postDate": "12/16/2017 00:10:05", "content": "

Too bad you don't get a prize for predicting this.

", "rawMarkdown": "Too bad you don't get a prize for predicting this.", "votes": null }, { "id": "258345", "postDate": "12/16/2017 00:34:37", "content": "

Which one is the MIT team?

", "rawMarkdown": "Which one is the MIT team?", "votes": null }, { "id": "258346", "postDate": "12/16/2017 00:35:12", "content": "

Congrats to idle_spec!

", "rawMarkdown": "Congrats to idle_spec!", "votes": null }, { "id": "258501", "postDate": "12/16/2017 09:09:12", "content": "

This is fun to watch how my prediction was confirmed. I think it was easy prediction anyway given idle_specs track record :)

", "rawMarkdown": "This is fun to watch how my prediction was confirmed. I think it was easy prediction anyway given idle_specs track record :)", "votes": null }, { "id": "258563", "postDate": "12/16/2017 12:46:24", "content": "

congrats to idle_spec! and other winners, reading this months ago gave me a little extra motivation :), I must say I took some personal pleasure when he had not made a final submission with only a few hrs to go. I thought he had over slept or something :), but your bet turned out in your favor. your boy good!!!

", "rawMarkdown": "congrats to idle_spec! and other winners, reading this months ago gave me a little extra motivation :), I must say I took some personal pleasure when he had not made a final submission with only a few hrs to go. I thought he had over slept or something :), but your bet turned out in your favor. your boy good!!!", "votes": null }, { "id": "258651", "postDate": "12/16/2017 17:51:44", "content": "

@raddar Can you tell more about the circumstances that made you believe idle_spec will win?

", "rawMarkdown": "raddar Can you tell more about the circumstances that made you believe idle_spec will win?", "votes": null }, { "id": "258728", "postDate": "12/16/2017 22:58:15", "content": "
\n

raddar wrote

\n \n
\n

I think it was easy prediction anyway given idle_specs track record :)

\n
\n
\n\n

I would not have guessed that, to be honest. Kagglers tend to cluster as deep learning / CNN / computer vision people on the one hand, and XGBoost / LightGBM / random forest people on the other hand. There is some overlap in skills and people, of course, but I think this pattern exists. If I'm not mistaken, this should be idle_speculation's first big success in a computer vision type of domain on Kaggle. Congrats, BTW! I hope we see a detailed write-up.

", "rawMarkdown": "> **raddar wrote**\n> \n> > I think it was easy prediction anyway given idle_specs track record :)\n\nI would not have guessed that, to be honest. Kagglers tend to cluster as deep learning / CNN / computer vision people on the one hand, and XGBoost / LightGBM / random forest people on the other hand. There is some overlap in skills and people, of course, but I think this pattern exists. If I'm not mistaken, this should be idle_speculation's first big success in a computer vision type of domain on Kaggle. Congrats, BTW! I hope we see a detailed write-up.", "votes": null }, { "id": "258903", "postDate": "12/17/2017 09:01:17", "content": "

@Bastiaan, at that time he was #1 (and had simillar score to private LB one). Plus he has access to huge server resources. If you looked up at Expedia (https://www.kaggle.com/c/expedia-hotel-recommendations) he said he had access to at least 700 gb RAM if i remember correctly.

", "rawMarkdown": "Bastiaan, at that time he was #1 (and had simillar score to private LB one). Plus he has access to huge server resources. If you looked up at Expedia (https://www.kaggle.com/c/expedia-hotel-recommendations) he said he had access to at least 700 gb RAM if i remember correctly.", "votes": null } ], "comments": [ { "id": 197508, "author_name": "jangerhofer", "author_url": "", "post_date": "06/29/2017 18:07:13", "content": "

Remember that there is a second stage, with a \"data-blind\"; i.e. the current leaderboard could be very deceptive.

", "votes": null, "replies": [] }, { "id": 197830, "author_name": "cpmpml", "author_url": "", "post_date": "06/30/2017 09:43:12", "content": "

You seem to antagonize people here ;)

\n\n

But this seems likely, idle_speculation is the only US one among top 10 Kagglers.

\n\n

Except that giba is now located in the US too ;) May not be a permanent resident yet though.

", "votes": null, "replies": [] }, { "id": 197900, "author_name": "godaibo", "author_url": "", "post_date": "06/30/2017 13:13:17", "content": "

good luck with that!!

", "votes": null, "replies": [] }, { "id": 198019, "author_name": "tunguz", "author_url": "", "post_date": "06/30/2017 18:44:32", "content": "

I have an idea: Kaggle should set up another competition where only non-US residents are allowed to compete, and the objective is to predict how well will the US residents do in this competition.

", "votes": null, "replies": [] }, { "id": 198056, "author_name": "kernele", "author_url": "", "post_date": "06/30/2017 21:25:47", "content": "

Several companies and top CS universities have formed teams, specifically for this competition consisting of only US citizens/residents. I personally know of a lab at MIT that formed a team specifically for this contest. Anything could happen!

", "votes": null, "replies": [ { "id": 258345, "author_name": "u39kun", "author_url": "", "post_date": "12/16/2017 00:34:37", "content": "

Which one is the MIT team?

", "votes": null, "replies": [] } ] }, { "id": 198063, "author_name": "serigne", "author_url": "", "post_date": "06/30/2017 22:03:15", "content": "

Until academic research teams decide to enter in the game ... then anything may happen.

\n\n

I red some blog posts where winners are academics who focus only on one competition (Data Science Bowl 2017 for the most recent or one of the first applications of deep NN here )

\n\n

I am sure some researchers from Willow at Ecole Normale Supérieure in Paris, would be eager to participate in this competition if the rules were not so unfavorable to non-US citizens

", "votes": null, "replies": [] }, { "id": 258333, "author_name": "brandenkmurray", "author_url": "", "post_date": "12/16/2017 00:10:05", "content": "

Too bad you don't get a prize for predicting this.

", "votes": null, "replies": [ { "id": 258346, "author_name": "u39kun", "author_url": "", "post_date": "12/16/2017 00:35:12", "content": "

Congrats to idle_spec!

", "votes": null, "replies": [] } ] }, { "id": 258501, "author_name": "raddar", "author_url": "", "post_date": "12/16/2017 09:09:12", "content": "

This is fun to watch how my prediction was confirmed. I think it was easy prediction anyway given idle_specs track record :)

", "votes": null, "replies": [ { "id": 258563, "author_name": "godaibo", "author_url": "", "post_date": "12/16/2017 12:46:24", "content": "

congrats to idle_spec! and other winners, reading this months ago gave me a little extra motivation :), I must say I took some personal pleasure when he had not made a final submission with only a few hrs to go. I thought he had over slept or something :), but your bet turned out in your favor. your boy good!!!

", "votes": null, "replies": [] }, { "id": 258651, "author_name": "bastiaanbergman", "author_url": "", "post_date": "12/16/2017 17:51:44", "content": "

@raddar Can you tell more about the circumstances that made you believe idle_spec will win?

", "votes": null, "replies": [] }, { "id": 258728, "author_name": "olegtrott", "author_url": "", "post_date": "12/16/2017 22:58:15", "content": "
\n

raddar wrote

\n \n
\n

I think it was easy prediction anyway given idle_specs track record :)

\n
\n
\n\n

I would not have guessed that, to be honest. Kagglers tend to cluster as deep learning / CNN / computer vision people on the one hand, and XGBoost / LightGBM / random forest people on the other hand. There is some overlap in skills and people, of course, but I think this pattern exists. If I'm not mistaken, this should be idle_speculation's first big success in a computer vision type of domain on Kaggle. Congrats, BTW! I hope we see a detailed write-up.

", "votes": null, "replies": [] }, { "id": 258903, "author_name": "raddar", "author_url": "", "post_date": "12/17/2017 09:01:17", "content": "

@Bastiaan, at that time he was #1 (and had simillar score to private LB one). Plus he has access to huge server resources. If you looked up at Expedia (https://www.kaggle.com/c/expedia-hotel-recommendations) he said he had access to at least 700 gb RAM if i remember correctly.

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "197370": "Knowing all the circumstances, I can already place my bets idle_speculation will get ez #1 and ez money in this competition :) good luck for others", "197508": "Remember that there is a second stage, with a \"data-blind\"; i.e. the current leaderboard could be very deceptive.", "197830": "You seem to antagonize people here ;)\n\nBut this seems likely, idle_speculation is the only US one among top 10 Kagglers.\n\nExcept that giba is now located in the US too ;) May not be a permanent resident yet though.", "197900": "good luck with that!!", "198019": "I have an idea: Kaggle should set up another competition where only non-US residents are allowed to compete, and the objective is to predict how well will the US residents do in this competition.", "198056": "Several companies and top CS universities have formed teams, specifically for this competition consisting of only US citizens/residents. I personally know of a lab at MIT that formed a team specifically for this contest. Anything could happen!", "198063": "Until academic research teams decide to enter in the game ... then anything may happen.\n\nI red some blog posts where winners are academics who focus only on one competition (Data Science Bowl 2017 for the most recent or one of the [first][1] applications of [deep NN][2] here )\n\nI am sure some researchers from [Willow ][3] at Ecole Normale Supérieure in Paris, would be eager to participate in this competition if the rules were not so unfavorable to non-US citizens\n\n\n [1]: http://blog.kaggle.com/2012/11/01/deep-learning-how-i-did-it-merck-1st-place-interview/\n [2]: http://blog.kaggle.com/2012/10/31/merck-competition-results-deep-nn-and-gpus-come-out-to-play/\n [3]: http://www.di.ens.fr/willow/", "258333": "Too bad you don't get a prize for predicting this.", "258345": "Which one is the MIT team?", "258346": "Congrats to idle_spec!", "258501": "This is fun to watch how my prediction was confirmed. I think it was easy prediction anyway given idle_specs track record :)", "258563": "congrats to idle_spec! and other winners, reading this months ago gave me a little extra motivation :), I must say I took some personal pleasure when he had not made a final submission with only a few hrs to go. I thought he had over slept or something :), but your bet turned out in your favor. your boy good!!!", "258651": "raddar Can you tell more about the circumstances that made you believe idle_spec will win?", "258728": "> **raddar wrote**\n> \n> > I think it was easy prediction anyway given idle_specs track record :)\n\nI would not have guessed that, to be honest. Kagglers tend to cluster as deep learning / CNN / computer vision people on the one hand, and XGBoost / LightGBM / random forest people on the other hand. There is some overlap in skills and people, of course, but I think this pattern exists. If I'm not mistaken, this should be idle_speculation's first big success in a computer vision type of domain on Kaggle. Congrats, BTW! I hope we see a detailed write-up.", "258903": "Bastiaan, at that time he was #1 (and had simillar score to private LB one). Plus he has access to huge server resources. If you looked up at Expedia (https://www.kaggle.com/c/expedia-hotel-recommendations) he said he had access to at least 700 gb RAM if i remember correctly." }, "source": "meta" }