BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Isaac Newton Institute Seminar Series
SUMMARY:Functions of Few Coordinate Variables: Sampling Sc
hemes and Recovery Algorithms - Simon FoucartĀ (
Texas A&\;M University )
DTSTART;TZID=Europe/London:20190618T090000
DTEND;TZID=Europe/London:20190618T095000
UID:TALK126127AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/126127
DESCRIPTION:I will revisit in this talk the task of approximat
ing multivariate functions that depend on only a f
ew of their variables. The number of samples requi
red to achieve this task to a given accuracy has b
een determined for Lipschitz functions several yea
rs ago. However\, two questions of practical inter
est remain: can we provide an explicit sampling s
trategy and can we efficiently produce approximant
s? I will (attempt to) answer these questions unde
r some additional assumptions on the target functi
on. Firstly\, if it is known to be linear\, then t
he problem is exactly similar to the standard comp
ressive sensing problem\, and I will review some o
f recent contributions there. Secondly\, if the ta
rget function is quadratic\, then the problem conn
ects to sparse phaseless recovery and to jointly l
ow-rank and bisparse recovery\, for which some res
ults and open questions will be presented. Finally
\, if the target function is known to increase coo
rdinatewise\, then the problem reduces to group te
sting\, from which I will draw the sought-after sa
mpling schemes and recovery algorithms.
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR