Hi Boyuan, hi everyone, On top of what Andy said, I would like to add that you don’t have to commit to certain algorithms in the proposal, as long as you make the plan very clear, and you leave time for discussing alternatives, pros and cons with the community.
Since you say there is some overlap with your exam period, it wouldn’t hurt to mention also when the exam period will end. I think you are under-discussing the “self-taught learning” part of the proposal. You’re basically saying: “This is also an algorithm. Some people argue whether it’s really semi-supervised but it probably is.” This hasn’t convinced me that I want it. >From a writing point of view I think your proposal reads well, only that the >“Details” section looks as if there are some headings in there that just >render as paragraph text. I find paragraphs starting with “And” to be too >conversational for this context, so you could try to rephrase. Finally, too bad it’s not 3 months ago, I could have met you in Saarbruecken, I was just next door! Hope my comments can help strengthen your proposal, Yours, Vlad > On 24 Mar 2015, at 19:08, Boyuan Deng <bryanhsud...@gmail.com> wrote: > > Hi Andreas: > >> when I think there is a closed form solution > Yes, I remember that in some paper they first give the analytical solution to > the optimization problem, and then prove that it's the same result that > iterative version will converge to. I'll find that paper and read it again. > > I think this improvement is highly viable. Tomorrow I'll estimate the > workload and update my proposal. > > Boyuan > > > On 03/24/2015 08:35 PM, Andy wrote: >> Hi Boyuan. >> I looked over your application and it looks good so far. >> I think it could be a bit more ambitious. I know the idea page was not very >> elaborate. >> It might be interesting to improve the existing graph-based algorithms. >> There is some discussion in >> https://github.com/scikit-learn/scikit-learn/pull/3758 >> I was a bit surprised to see that the algorithms currently are iterative, >> when I think there is a closed form solution. >> Maybe you could look into that, too? >> If you have other related ideas, feel free to include them as well. >> >> Best, >> Andy >> >> >> On 03/22/2015 05:47 PM, Boyuan Deng wrote: >>> Hi all: >>> >>> This is the link to my proposal for the "Cross-validation and >>> Meta-estimators for Semi-supervised Learning" topic: >>> https://docs.google.com/document/d/1f2nfFEBk567QhKd2OJzDNM9t21Glkp0XxFgtbpy8UjI/edit?usp=sharing >>> >>> Please leave comments and help improving it! >>> >>> Also I want to contribute a bigger patch to scikit-learn in the following >>> weeks. So dear developers if you know there is a suitable issue (better if >>> related to my topic), please @bryandeng on that issue page. >>> >>> Boyuan >>> >>> On 03/19/2015 03:44 AM, Joel Nothman wrote: >>>> Hi Boyuan, >>>> >>>> It's good to hear you're an experienced scikit-learn user, and that it has >>>> worked for you. It's also pleasing to hear someone's interested in this >>>> project, because I feel the semi-supervised capabilities of scikit-learn >>>> and its API have been left half-baked. >>>> >>>> I strongly recommend, however, that as well as putting together a >>>> proposal, you try to show your practical abilities to contribute to the >>>> project so that core developers become familiar with your work. I see you >>>> have started to do so at #4409; something that involves testing and/or >>>> documentation would be ideal. >>>> >>>> - Joel >>>> >>>> On 19 March 2015 at 09:30, Boyuan Deng <bryanhsud...@gmail.com> wrote: >>>> Hi all: >>>> >>>> I am a Master's student in the European Union's Erasmus Mundus LCT >>>> program, studying natural language processing at Saarland University, >>>> Germany and also doing machine learning and information retrieval at >>>> Max-Planck Institute for Informatics, which is on the same campus. >>>> >>>> These years I've been a heavy scikit-learn user and once got top 10 in KDD >>>> Cup using scikit-learn. So first I want to thank you all the developers! >>>> And that's why I strongly want to contribute to scikit-learn during this >>>> year's GSoC. >>>> >>>> I'm interested in the topic "Cross-validation and Meta-Estimators for >>>> semi-supervised learning" and is starting to write my proposal. I'll post >>>> the link for it once it has basically taken shape this week. >>>> >>>> I'm always available if any mentor wants to contact me. >>>> >>>> Best regards, >>>> Boyuan Deng >>>> >>>> >>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Dive into the World of Parallel Programming The Go Parallel Website, >>>> sponsored >>>> by Intel and developed in partnership with Slashdot Media, is your hub for >>>> all >>>> things parallel software development, from weekly thought leadership blogs >>>> to >>>> news, videos, case studies, tutorials and more. Take a look and join the >>>> conversation now. http://goparallel.sourceforge.net/ >>>> _______________________________________________ >>>> Scikit-learn-general mailing list >>>> Scikit-learn-general@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general >>>> >>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Dive into the World of Parallel Programming The Go Parallel Website, >>>> sponsored >>>> by Intel and developed in partnership with Slashdot Media, is your hub for >>>> all >>>> things parallel software development, from weekly thought leadership blogs >>>> to >>>> news, videos, case studies, tutorials and more. Take a look and join the >>>> conversation now. >>>> http://goparallel.sourceforge.net/ >>>> >>>> >>>> _______________________________________________ >>>> Scikit-learn-general mailing list >>>> >>>> Scikit-learn-general@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general >>> >>> >>> >>> ------------------------------------------------------------------------------ >>> Dive into the World of Parallel Programming The Go Parallel Website, >>> sponsored >>> by Intel and developed in partnership with Slashdot Media, is your hub for >>> all >>> things parallel software development, from weekly thought leadership blogs >>> to >>> news, videos, case studies, tutorials and more. Take a look and join the >>> conversation now. >>> http://goparallel.sourceforge.net/ >>> >>> >>> _______________________________________________ >>> Scikit-learn-general mailing list >>> >>> Scikit-learn-general@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general >> >> >> >> ------------------------------------------------------------------------------ >> Dive into the World of Parallel Programming The Go Parallel Website, >> sponsored >> by Intel and developed in partnership with Slashdot Media, is your hub for >> all >> things parallel software development, from weekly thought leadership blogs to >> news, videos, case studies, tutorials and more. Take a look and join the >> conversation now. >> http://goparallel.sourceforge.net/ >> >> >> _______________________________________________ >> Scikit-learn-general mailing list >> >> Scikit-learn-general@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming The Go Parallel Website, sponsored > by Intel and developed in partnership with Slashdot Media, is your hub for all > things parallel software development, from weekly thought leadership blogs to > news, videos, case studies, tutorials and more. Take a look and join the > conversation now. > http://goparallel.sourceforge.net/_______________________________________________ > Scikit-learn-general mailing list > Scikit-learn-general@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general ------------------------------------------------------------------------------ Dive into the World of Parallel Programming The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
