BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Drupal iCal API//EN
X-WR-CALNAME:Events items teaser
X-WR-TIMEZONE:America/Toronto
BEGIN:VTIMEZONE
TZID:America/Toronto
X-LIC-LOCATION:America/Toronto
BEGIN:DAYLIGHT
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
DTSTART:20240310T070000
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
DTSTART:20231105T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:69eb6728207f2
DTSTART;TZID=America/Toronto:20240611T100000
SEQUENCE:0
TRANSP:TRANSPARENT
DTEND;TZID=America/Toronto:20240611T110000
URL:https://uwaterloo.ca/institute-for-quantum-computing/events/classical-v
 erification-quantum-learning
SUMMARY:Classical Verification of Quantum Learning
CLASS:PUBLIC
DESCRIPTION:CS/MATH SEMINAR MARCEL HINSCHE FROM FREIE UNIVERSITÄT BERLIN\n
 \nQuantum-Nano Centre\, 200 University Ave West\, Waterloo\, ON CA N2L 3G1
 \nZOOM ONLY\n\nQuantum data access and quantum processing can make certain
 \nclassically intractable learning tasks feasible. However\, quantum\ncapa
 bilities will only be available to a select few in the near\nfuture. Thus\
 , reliable schemes that allow classical clients to\ndelegate learning to u
 ntrusted quantum servers are required to\nfacilitate widespread access to 
 quantum learning advantages. Building\non a recently introduced framework 
 of interactive proof systems for\nclassical machine learning\, we develop 
 a framework for classical\nverification of quantum learning. We exhibit le
 arning problems that a\nclassical learner cannot efficiently solve on thei
 r own\, but that they\ncan efficiently and reliably solve when interacting
  with an untrusted\nquantum prover. Concretely\, we consider the problems 
 of agnostic\nlearning parities and Fourier-sparse functions with respect t
 o\ndistributions with uniform input marginal. We propose a new quantum\nda
 ta access model that we call \"mixture-of-superpositions\" quantum\nexampl
 es\, based on which we give efficient quantum learning algorithms\nfor the
 se tasks. Moreover\, we prove that agnostic quantum parity and\nFourier-sp
 arse learning can be efficiently verified by a classical\nverifier with on
 ly random example or statistical query access.\nFinally\, we showcase two 
 general scenarios in learning and\nverification in which quantum mixture-o
 f-superpositions examples do\nnot lead to sample complexity improvements o
 ver classical data. Our\nresults demonstrate that the potential power of q
 uantum data for\nlearning tasks\, while not unlimited\, can be utilized by
  classical\nagents through interaction with untrusted quantum entities.
DTSTAMP:20260424T125048Z
END:VEVENT
END:VCALENDAR