verne.freenode.net changed the topic of #mlpack to: http://www.mlpack.org/ -- We don't respond instantly... but we will respond. Give it a few minutes. Or hours. -- Channel logs: http://www.mlpack.org/irc/
< rcurtin>
wasiq: it would be good to read those papers and be familiar with SGD-type approaches
< rcurtin>
the Hogwild! paper may also be useful
< rcurtin>
(and you could take a look at other more recent papers that cite the Hogwild! paper)
tsathoggua has quit [Quit: Konversation terminated!]
< rcurtin>
I haven't heard from awhitesong in over a year, I wonder why travis decided to build his branch
< zoq>
and sends a messesage to this channel
< rcurtin>
bizarre
timsl has quit [Quit: Page closed]
liu__ has joined #mlpack
liu__ has quit [Client Quit]
ank_95_ has quit [Quit: Connection closed for inactivity]
< wasiq>
rcurtin, alright. So basically SGD takes a random value of the input set as parameter and manages to retain a reasonable consistency in output,but with a increased error probability?
< rcurtin>
yes, it's a stochastic optimizer
< rcurtin>
it's not necessarily right to say that the error probability is increased, though
< rcurtin>
there has been a lot of theoretical analysis of SGD-like algorithms (again the Hogwild! paper has some good analysis and links to other analyses)
hercky has joined #mlpack
tsathoggua has joined #mlpack
tsathoggua has quit [Client Quit]
< wasiq>
okay.. Ill read up on it.
hercky has quit [Quit: Page closed]
Tanishq has quit [Ping timeout: 240 seconds]
zoq_ has joined #mlpack
zoq has quit [Ping timeout: 260 seconds]
Falit has joined #mlpack
drupal-contribut has joined #mlpack
cheral has joined #mlpack
drupal-contribut has left #mlpack []
abdelhadiTEMMAR has joined #mlpack
< abdelhadiTEMMAR>
hi ! i'm Abdelhadi TEMMAR :)
< abdelhadiTEMMAR>
so, how many people use this library today ?
anveshi has joined #mlpack
v3 has joined #mlpack
ank_95_ has joined #mlpack
< abdelhadiTEMMAR>
i'm interested in the neuroevolution project. I was wodering if stochastic gradient descent and CNE are synonyms ? Because "simple weight evolution over a topologically static neural network." seems to be definition of SGD for me
STW has joined #mlpack
praveench has joined #mlpack
abdelhadiTEMMAR has quit [Ping timeout: 252 seconds]
tont has joined #mlpack
anveshi has quit [Ping timeout: 252 seconds]
Tanishq has joined #mlpack
yvtheja has quit [Ping timeout: 252 seconds]
tugu has joined #mlpack
< tugu>
I want to work on the open source #mlpack and have a good knowledge of machine learning. I want to know the criterion that you judge in a potential candidate.
< Guest26255>
g++ 2.cpp -lmlpack -std=c++11 /usr/bin/ld: /tmp/ccM52eLV.o: undefined reference to symbol '_ZTIN5boost13serialization13typeid_system27extended_type_info_typeid_0E' //usr/lib/x86_64-linux-gnu/libboost_serialization.so.1.54.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status
< Guest26255>
can anybody help me ?
Tanishq has quit [Ping timeout: 276 seconds]
cache-nez has quit [Ping timeout: 268 seconds]
praveench has quit [Quit: Page closed]
Tanishq has joined #mlpack
tont has quit [Ping timeout: 240 seconds]
acbull has joined #mlpack
ayshtmr has joined #mlpack
< acbull>
Hello. Something run with the compiling.
< yvtheja>
ank_95_, You look at ideas page first :)
< yvtheja>
<yvtheja> ank_95_, You canlook at ideas page first :)
vtantia has joined #mlpack
anveshi has quit [Quit: Page closed]
ayshtmr has joined #mlpack
< ank_95_>
I have looked at it and find the project "Fast k-centers Algorithm & Implementation" interesting, Can you tell how should I proceed.
serbanM has quit [Ping timeout: 240 seconds]
yvtheja has quit [Ping timeout: 246 seconds]
yvtheja has joined #mlpack
anveshi has joined #mlpack
< Jayaganesh>
ank_95, I'm another contributor like u. U need to probably start off with "Dual-Tree Boruvka MST algorithm" and its implementation in MLpack. And even before that u'd like to read "Scalable Sparse Approximation of a Sample Mean" by Cortes and Scott-2013. U may start off with these. hope these helped u.
vt has joined #mlpack
vtantia has quit [Ping timeout: 252 seconds]
ayshtmr has left #mlpack []
cache-nez has joined #mlpack
mentekid has quit [Remote host closed the connection]
< rcurtin>
please be patient for an answer. I have to answer tens of emails and lots of bug reports
< rcurtin>
you are not "supposed" to do anything in particular; your project proposal is up to you
cache-nez has joined #mlpack
< rcurtin>
what you've suggested (Algorithm 3) is one possibility, but it is not the only possibility for what you could do for this project
asnel_ulrich has quit [Ping timeout: 246 seconds]
wasiq has joined #mlpack
rahulgoyal has quit [Ping timeout: 240 seconds]
< ranjan123>
and how can I test the sequential sgd ?
< rcurtin>
you can build mlpack_test (`make mlpack_test`) and run the test: `bin/mlpack_test -t SGDTest`
< rcurtin>
for the code that is used for testing, see src/mlpack/tests/sgd_test.cpp
< ranjan123>
hmm. thank you !
< rcurtin>
sure, glad I could be helpful
asnel_ulrich has joined #mlpack
< Rishabh>
rcurtin: Hi, I wish to apply for GSoC 2016 for the project Implement tree types. Since the project involves implementing different variations of tree data structure, I want to start contributing by implementing one of the types of tree enlisted under the project description.
< Rishabh>
hi, I was going through the code in core/tree/rectangle_tree/rectangle_tree.hpp . Rectangle tree has parameters maxLeafSize = 20,minLeafSize = 8, maxNumChildren = 5 and minNumChildren = 2. Can't understand why, am I missing out on something ?
< kirizaki>
@zoq: mlpack::backtrace doesn't need LIBBFD_IBERTY_LIBRARY and LIBBFD_OPCODES_LIBRARY in Linux
< rcurtin>
Rishabh: those are just defaults, they may not be the best for any given situation but they work okay
< kirizaki>
@zoq: including those libs was throwing some errors, I don't remember now what was it exactly
< kirizaki>
@zoq: do You want me to reproduce it?
Eigenspace has joined #mlpack
< zoq>
kirizaki: I shouldn't make a difference to include LIBBFD_IBERTY_LIBRARY and LIBBFD_OPCODES_LIBRARY as long as you don't use it. No need to reproduce the error.
Eigenspace has quit [Client Quit]
sumedhghaisas has quit [Ping timeout: 252 seconds]