A General Characterization of the Statistical Query Complexity. Feldman, V. In Kale, S. & Shamir, O., editors, volume 65, of Proceedings of Machine Learning Research, pages 785–830, Amsterdam, Netherlands, 07–10 Jul, 2017. PMLR.
[Feldman17] Provides a characterization of SQ learning in terms of new "statistical dimension;" to be seen in view of the poly-factor equivalence between SQ learning and LDP-constrained/memory-constrained learnin (e.g., [KLNRS11,BDD98]). Includes applications to memory-limited streaming and communication-limited learning.

A General Characterization of the Statistical Query Complexity [link]Paper  bibtex   
@inproceedings{Feldman17,
  title = 	 {A General Characterization of the Statistical Query Complexity},
  author = 	 {Vitaly Feldman},
  pages = 	 {785--830},
  year = 	 {2017},
  editor = 	 {Satyen Kale and Ohad Shamir},
  volume = 	 {65},
  series = 	 {Proceedings of Machine Learning Research},
  address = 	 {Amsterdam, Netherlands},
  month = 	 {07--10 Jul},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v65/feldman17c/feldman17c.pdf},
  url = 	 {http://proceedings.mlr.press/v65/feldman17c.html},
  bibbase_note = {<div class="well well-small bibbase"><span class="bluecite">[Feldman17]</span> Provides a characterization of SQ learning in terms of new "statistical dimension;" to be seen in view of the poly-factor equivalence between SQ learning and LDP-constrained/memory-constrained learnin (e.g., [KLNRS11,BDD98]). Includes applications to memory-limited streaming and communication-limited learning.</div>}
}

Downloads: 0