In the last few days of this comp, I am starting to pick my next \"game\".
\nI have 3 targets for now:
I guess some recommendation experts will rush into this one after OTTO comp ends.
\nAnd there is efficiency prize, which is a very interesting idea. No matter who thought of this idea, a big thumb up to you. I am tired of ensembing, especially for the people lacking of computation resources.
\n2.GoDaddy. I am attracted by its small dataset, especially after the torture by OTTO(just kidding😂). We can finally get the relief from OOM error.
\n3.Icecube. Such a beatiful lab under the ice. And there is write up prizes. A pretty new domain to learn about.
\nI am truthly thankful for the innovations kaggle team has made recently. I like those new concepts. I think innovation will make kaggle community more active, and attract more kagglers with different taste. Good job, kaggle team.
\nLast but not least, do take serious actions against cheaters. I know there are lots of jobs to do. But please prioritize this one. Cheaters not only hurt the motivation of people doing kaggle, destroying the hard work of genuine participants, but also gradually deteriorate the reputation of kaggle medals. Just imagine, one day people do not trust your gold medal anymore, they always suspect how did you get it. Once the credibility of kaggle medals disappears, many people will lose the interest taking part into kaggle.
\nI like machine learning, I like kaggling, I am a happy kaggler. I wish the imagination above will never happen.
\nAll the best to the genuine kagglers.
Almost forget, please share your thoughts about your next \"game\".
", "messages": [ { "id": "2118561", "postDate": "01/28/2023 05:39:03", "content": "In the last few days of this comp, I am starting to pick my next \"game\".
\nI have 3 targets for now:
I guess some recommendation experts will rush into this one after OTTO comp ends.
\nAnd there is efficiency prize, which is a very interesting idea. No matter who thought of this idea, a big thumb up to you. I am tired of ensembing, especially for the people lacking of computation resources.
\n2.GoDaddy. I am attracted by its small dataset, especially after the torture by OTTO(just kidding😂). We can finally get the relief from OOM error.
\n3.Icecube. Such a beatiful lab under the ice. And there is write up prizes. A pretty new domain to learn about.
\nI am truthly thankful for the innovations kaggle team has made recently. I like those new concepts. I think innovation will make kaggle community more active, and attract more kagglers with different taste. Good job, kaggle team.
\nLast but not least, do take serious actions against cheaters. I know there are lots of jobs to do. But please prioritize this one. Cheaters not only hurt the motivation of people doing kaggle, destroying the hard work of genuine participants, but also gradually deteriorate the reputation of kaggle medals. Just imagine, one day people do not trust your gold medal anymore, they always suspect how did you get it. Once the credibility of kaggle medals disappears, many people will lose the interest taking part into kaggle.
\nI like machine learning, I like kaggling, I am a happy kaggler. I wish the imagination above will never happen.
\nAll the best to the genuine kagglers.
Almost forget, please share your thoughts about your next \"game\".
", "rawMarkdown": "In the last few days of this comp, I am starting to pick my next \"game\".\nI have 3 targets for now:\n1. Learning Equality. A nlp recommendation task. \n\nI guess some recommendation experts will rush into this one after OTTO comp ends. \n\nAnd there is efficiency prize, which is a very interesting idea. No matter who thought of this idea, a big thumb up to you. I am tired of ensembing, especially for the people lacking of computation resources.\n\n2.GoDaddy. I am attracted by its small dataset, especially after the torture by OTTO(just kidding😂). We can finally get the relief from OOM error.\n\n3.Icecube. Such a beatiful lab under the ice. And there is write up prizes. A pretty new domain to learn about.\n\nI am truthly thankful for the innovations kaggle team has made recently. I like those new concepts. I think innovation will make kaggle community more active, and attract more kagglers with different taste. Good job, kaggle team.\n\nLast but not least, do take serious actions against cheaters. I know there are lots of jobs to do. But please prioritize this one. Cheaters not only hurt the motivation of people doing kaggle, destroying the hard work of genuine participants, but also gradually deteriorate the reputation of kaggle medals. Just imagine, one day people do not trust your gold medal anymore, they always suspect how did you get it. Once the credibility of kaggle medals disappears, many people will lose the interest taking part into kaggle.\nI like machine learning, I like kaggling, I am a happy kaggler. I wish the imagination above will never happen.\nAll the best to the genuine kagglers.\n\nAlmost forget, please share your thoughts about your next \"game\".", "votes": null }, { "id": "2118589", "postDate": "01/28/2023 06:11:53", "content": "Learning Equality. Cause I am a nlper and I learned a lot about recommender system techniques in this(otto) competition. I think I can get a better score in that game~ BTW, your CV is so much strong.
", "rawMarkdown": "Learning Equality. Cause I am a nlper and I learned a lot about recommender system techniques in this(otto) competition. I think I can get a better score in that game~ BTW, your CV is so much strong.", "votes": null }, { "id": "2118605", "postDate": "01/28/2023 06:23:38", "content": "You didn't mention RSNA. I'll be there.
", "rawMarkdown": "You didn't mention RSNA. I'll be there.", "votes": null }, { "id": "2118624", "postDate": "01/28/2023 06:37:24", "content": "I browsed the discussion in RSNA, it seems it needs huge amount of compution resources which excludes me out.
", "rawMarkdown": "I browsed the discussion in RSNA, it seems it needs huge amount of compution resources which excludes me out.", "votes": null }, { "id": "2118640", "postDate": "01/28/2023 06:44:55", "content": "1 + 1 > 2 in terms of domain knowledge, you have a large chance of getting better score. Good luck!
\nRegards my CV, indeed, even a 0.0002 boost in my local CV will cause a rank boost in LB. And if my local CV predicts a 0.6035 LB score, my position is somewhere around middle of all 0.603 teams. I have to say, the host and kaggle team did an excellent job in preparing the data. They split the data to almost identical distributions, otherwise such good correlation is not expected.
Small break :)
\nThen RSNA, but I wish I went for a big break instead.
", "rawMarkdown": "Small break :)\n\nThen RSNA, but I wish I went for a big break instead.", "votes": null }, { "id": "2118915", "postDate": "01/28/2023 11:39:31", "content": "Seems a lot of GMs will go to RSNA. A tough battle begins.
", "rawMarkdown": "Seems a lot of GMs will go to RSNA. A tough battle begins.", "votes": null } ], "comments": [ { "id": 2118589, "author_name": "earlee0412", "author_url": "", "post_date": "01/28/2023 06:11:53", "content": "Learning Equality. Cause I am a nlper and I learned a lot about recommender system techniques in this(otto) competition. I think I can get a better score in that game~ BTW, your CV is so much strong.
", "votes": null, "replies": [ { "id": 2118640, "author_name": "buumoo", "author_url": "", "post_date": "01/28/2023 06:44:55", "content": "1 + 1 > 2 in terms of domain knowledge, you have a large chance of getting better score. Good luck!
\nRegards my CV, indeed, even a 0.0002 boost in my local CV will cause a rank boost in LB. And if my local CV predicts a 0.6035 LB score, my position is somewhere around middle of all 0.603 teams. I have to say, the host and kaggle team did an excellent job in preparing the data. They split the data to almost identical distributions, otherwise such good correlation is not expected.
You didn't mention RSNA. I'll be there.
", "votes": null, "replies": [ { "id": 2118624, "author_name": "buumoo", "author_url": "", "post_date": "01/28/2023 06:37:24", "content": "I browsed the discussion in RSNA, it seems it needs huge amount of compution resources which excludes me out.
", "votes": null, "replies": [] } ] }, { "id": 2118641, "author_name": "theoviel", "author_url": "", "post_date": "01/28/2023 06:48:38", "content": "Small break :)
\nThen RSNA, but I wish I went for a big break instead.
", "votes": null, "replies": [ { "id": 2118915, "author_name": "buumoo", "author_url": "", "post_date": "01/28/2023 11:39:31", "content": "Seems a lot of GMs will go to RSNA. A tough battle begins.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "2118561": "In the last few days of this comp, I am starting to pick my next \"game\".\nI have 3 targets for now:\n1. Learning Equality. A nlp recommendation task. \n\nI guess some recommendation experts will rush into this one after OTTO comp ends. \n\nAnd there is efficiency prize, which is a very interesting idea. No matter who thought of this idea, a big thumb up to you. I am tired of ensembing, especially for the people lacking of computation resources.\n\n2.GoDaddy. I am attracted by its small dataset, especially after the torture by OTTO(just kidding😂). We can finally get the relief from OOM error.\n\n3.Icecube. Such a beatiful lab under the ice. And there is write up prizes. A pretty new domain to learn about.\n\nI am truthly thankful for the innovations kaggle team has made recently. I like those new concepts. I think innovation will make kaggle community more active, and attract more kagglers with different taste. Good job, kaggle team.\n\nLast but not least, do take serious actions against cheaters. I know there are lots of jobs to do. But please prioritize this one. Cheaters not only hurt the motivation of people doing kaggle, destroying the hard work of genuine participants, but also gradually deteriorate the reputation of kaggle medals. Just imagine, one day people do not trust your gold medal anymore, they always suspect how did you get it. Once the credibility of kaggle medals disappears, many people will lose the interest taking part into kaggle.\nI like machine learning, I like kaggling, I am a happy kaggler. I wish the imagination above will never happen.\nAll the best to the genuine kagglers.\n\nAlmost forget, please share your thoughts about your next \"game\".", "2118589": "Learning Equality. Cause I am a nlper and I learned a lot about recommender system techniques in this(otto) competition. I think I can get a better score in that game~ BTW, your CV is so much strong.", "2118605": "You didn't mention RSNA. I'll be there.", "2118624": "I browsed the discussion in RSNA, it seems it needs huge amount of compution resources which excludes me out.", "2118640": "1 + 1 > 2 in terms of domain knowledge, you have a large chance of getting better score. Good luck!\nRegards my CV, indeed, even a 0.0002 boost in my local CV will cause a rank boost in LB. And if my local CV predicts a 0.6035 LB score, my position is somewhere around middle of all 0.603 teams. I have to say, the host and kaggle team did an excellent job in preparing the data. They split the data to almost identical distributions, otherwise such good correlation is not expected.", "2118641": "Small break :)\n\nThen RSNA, but I wish I went for a big break instead.", "2118915": "Seems a lot of GMs will go to RSNA. A tough battle begins." }, "source": "meta" }