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:Artificial Intelligence Research Group Talks (Comp
 uter Laboratory)
SUMMARY:Success\, sensitivity and unbelievable quality of 
 LLM code generation - Andrei Paleyes (University o
 f Cambridge)
DTSTART;TZID=Europe/London:20250624T130000
DTEND;TZID=Europe/London:20250624T140000
UID:TALK232729AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/232729
DESCRIPTION:Code generation is one of the most active areas of
  application of Large Language Models (LLMs). LLMs
  lower barriers to writing code and accelerate dev
 elopment process\, so much so there are now resear
 ch projects and businesses powered largely by LLM 
 code generation process. It is however important t
 o realise that the overall quality of generated pr
 ograms depends on the quality of given prompts. Fu
 nctionality and quality of generated code can be s
 ensitive to user's background and familiarity with
  software engineering.\n\nIn this talk we will dis
 cuss several aspects of LLM code generation. We wi
 ll talk about end-user software engineering paradi
 gm and how it is put to practice today. We will in
 vestigate how sensitive this approach can be to ch
 anges in the requirements. And we will see how LLM
 s can be deceptively good at solving seemingly imp
 ossible problems with code.\n\n"You can also join 
 us on Zoom":https://cam-ac-uk.zoom.us/j/8340033552
 2?pwd=LkjYvMOvVpMbabOV1MVTm8QU6DrGN7.1\n
LOCATION:Lecture Theatre 2\, Computer Laboratory\, William 
 Gates Building
CONTACT:Mateja Jamnik
END:VEVENT
END:VCALENDAR