BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Quantum Statistical Query Learning - Vojtěch Havlíček\, IBM T.J
 . Watson Research Center
DTSTART:20240125T141500Z
DTEND:20240125T153000Z
UID:TALK209839@talks.cam.ac.uk
CONTACT:Subhayan Roy Moulik
DESCRIPTION:Joint work with Louis Schatzki (UIUC) and Srinivasan Arunachal
 am (IBM Almaden).\n\nStatistical Query Learning (SQ) is a restriction of P
 AC learning in which learners form hypotheses by querying expectation valu
 es over the data distribution and labels. It is well known that SQ cannot 
 learn the concept class of parities\, even under uniform distribution. Thi
 s shows that SQ is a restriction of PAC. 

Here we study a quantum gen
 eralization of SQ\, quantum statistical query learning (QSQ) previously pr
 oposed by Arunachalam\, Grilo and Yuen.  It has been shown that QSQ can ef
 ficiently learn parities and is therefore a stronger learning model than S
 Q. It was however open how it compares to (quantum) PAC learning model.\n\
 nOur main result is a task that separates quantum PAC and QSQ. To prove it
 \, we built on Feldman‘s work on statistical query learning to derive lo
 wer bounds on QSQ learning. I will discuss our result\, give details about
  the lower bounding technique and outline some of its applications
LOCATION:MR9
END:VEVENT
END:VCALENDAR