langford   32

Twitter
RT : Crews will be heading to an outage affecting 2687 customers in area. Updates:…
ViewRoyal  Colwood  Langford  from twitter
march 2017 by wakemp
Why LinkedIn is Important for the Financial Professional
Why LinkedIn is Important for the Financial Professional, from Socialware Blog | Social Business Management for Financial Services http://blog.socialware.com
ifttt  googlereader  Socialware  Blog  |  Social  Business  Management  for  Financial  Services  Mike  Langford 
october 2011 by scottpierce
Scientific Method (Stargate SG1; G)
The last thing Catherine needed right now was another over-eager, starry-eyed officer on her team whose only concern was making Major before her next birthday.
sg1  carter  langford  backstory  pre-series  author:zinke 
november 2010 by mischief5
ICML 2010 Tutorial on Learning through Exploration
"This tutorial is about learning through exploration. The goal is to learn how to make decisions when the payoff of only a chosen action is observed rather than all choices. The setting we address is simpler than general reinforcement learning, because we consider situations where future rewards are not affected by past decisions, although the algorithms we discuss do have applications in this more general setting."
sequential_decisions  bandit_problems  langford  john  beygelzimer  alina 
june 2010 by shivak
A suggestion from John/Jean
Use some of the training data to grow mini-experts, one for each subset of features, then use the rest of the data to figure out how the mini-experts complement one another. Why? To avoid incoherence conditions.
ensemble_methods  machine_learning  incoherence  langford  john  audibert  jean-yves 
may 2010 by shivak
What’s the difference between gambling and rewarding good prediction?
John equates financial risk for investing with regret in online learning, and comes to boring conclusions.
investment  online_learning  langford  john 
may 2010 by shivak
Importance weighted active learning
Computationally tractable, loss-function agnostic active learning with realistic, finite-sample generalization guarantees. Generalization of the disagreement coefficient to different loss functions, and a minimax lower bound that is tighter than Kaariainen's. Empirical evaluation.
active_learning  learning_theory  beygelzimer  alina  dasgupta  sanjoy  langford  john 
may 2009 by shivak

related tags

'13  'avengers  4’  7.30pm.  active_learning  agnostic_setting  ago  alabama  alex  alina  and  andrea  anirban  arrangement  at  attenberg  audibert  author:zinke  backstory  balcan  bandit_problems  bc  beygelzimer  birmingham  blog  blogs  brian  broadcasts  brown  business  carter  cast  ceremony  charlie  choir  classification  collins  colwood  come  concentration_of_measure  county  criticism  cs  dasgupta  david  dimension_reduction  diversity  economic  ensemble_methods  filetype:pdf  financial  five  for  free  girls  googlereader  hashing  health  hear  history  hiv  if  ifttt  in  inauguration  incoherence  insurance  investment  jana  jayme  jean-yves  jefferson  john's  john  jordan  josh  katherine  kilian  learning_theory  life  literature  machine_learning  machinelearning  making  management  maria-florina  markets  massmutual  mayor  mechanism_design  media:document  mekons  mike  millennium  more  municipalities  netvouzimported  netvouzpublic  of  online_learning  overfitting  pat  pdf  pio  porn  pornstars  pre-series  proposal  reasons  reference  research  review  sanjoy  semi-supervised_learning  sequential_decisions  services  sewer  sexy  sf  sg1  sheen  smola  social  socialware  sparsity  st  stableford  statistics  suit  thanks  the  third  to  today!  tomorrow  video.  viewroyal  want  waterloo  weinberger  why'  world  writing  years  you  | 

Copy this bookmark:



description:


tags: