BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Fredrik Odegaard (room: WMC 4335\, jointly hosted with Beedie) (Iv
 ey Business School (University of Western Ontario))
DTSTART:20260220T193000Z
DTEND:20260220T210000Z
DTSTAMP:20260825T060155Z
UID:SFUOR/69
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/SFUOR/69/">D
 ecoding Consumer Preferences Using Attention-Based Language Models</a>\nby
  Fredrik Odegaard (room: WMC 4335\, jointly hosted with Beedie) (Ivey Busi
 ness School (University of Western Ontario)) as part of PIMS-CORDS SFU Ope
 rations Research Seminar\n\nLecture held in ASB 10908.\n\nAbstract\nThis p
 aper proposes a new demand estimation method using attention-based languag
 e models. An encoder-only language model is trained in a two-stage process
  to analyze the natural language descriptions of used cars from a large US
 -based online auction marketplace. The approach enables semi-nonparametric
 ally estimation for the demand primitives of a structural model representi
 ng the private valuations and market size for each vehicle listing. In the
  first stage\, the language model is fine-tuned to encode the target aucti
 on outcomes using the natural language vehicle descriptions. In the second
  stage\, the trained language model's encodings are projected into the par
 ameter space of the structural model. The model's capability to conduct co
 unterfactual analyses within the trained market space is validated using a
  subsample of withheld auction data\, which includes a set of unique "zero
  shot" instances.\n\nThis is joint work with Joshua Foster.\n
LOCATION:https://researchseminars.org/talk/SFUOR/69/
END:VEVENT
END:VCALENDAR