BEGIN:VCALENDAR
VERSION:2.0
CALSCALE:GREGORIAN
PRODID:iCalendar-Ruby
BEGIN:VEVENT
CATEGORIES:Colloquium/Seminar
DESCRIPTION:Title: Large Language Models for Teacher Feedback\n\nPresenter:
  Dr. Dora Demszky\, Assistant Professor\, Education Data Science\, Graduate
  School of Education and Computer Science\, Stanford University\n\nAbstract
 : Large Language Models show unprecedented potential in scaling many aspect
 s of education — can they also facilitate high-quality instruction by servi
 ng as a professional learning tool for teachers? This talk synthesizes find
 ings from three papers that evaluate the ability of LLMs to provide feedbac
 k to educators on their discourse. Firstly\, we evaluate the ability of Gha
 tGPT to perform instructional coaching tasks\, such as using an observation
 al instrument to score instruction and provide teachers with suggestions fo
 r improvement. We show that while the model generates feedback that is rele
 vant\, it struggled with insightfulness. Secondly\, we explore the use of L
 LMs in assisting math tutors with addressing students math mistakes. Despit
 e consistent improvements over the original tutor responses\, the models' r
 emediation efforts often fell short when compared to experienced math teach
 ers\, indicating a need for keeping an expert in the loop. Lastly\, we exam
 ine LLMs ability to use growth mindset supportive language (GMSL) in math t
 eaching moments (responding to student mistakes\, introducing/debriefing ta
 sks). Encouragingly\, after prompt-engineering\, LLMs showcased the ability
  to reframe unsupportive utterances effectively\, even surpassing GMSL-trai
 ned teachers in certain evaluations. These collective insights demonstrate 
 the promise of LLMs in complementing existing teacher professional learning
  processes. However\, the current limitations also emphasize the continued 
 need for human expertise in improving and overseeing the model's implementa
 tion.\n\n\nBio: Dora Demszky is an Assistant Professor in Education Data Sc
 ience at Stanford University\, and in Computer Science (by courtesy). Her r
 esearch focuses developing and deploying tools that combine natural languag
 e processing\, linguistics and input from practitioners to facilitate equit
 able\, student-centered instruction. Her tools analyze educational discours
 e such as student-teacher interactions\, student group work and textbooks\,
  to identify features of of high-quality instruction and suggest areas of i
 mprovement for educators. Dr Demszky has received her PhD in Linguistics at
  Stanford.\n\nhttps://www.dorademszky.com/
DTEND:20230915T200000Z
DTSTAMP:20260420T094315Z
DTSTART:20230915T180000Z
GEO:40.008252;-105.268415
LOCATION:Muenzinger Psychology\, D430
SEQUENCE:0
SUMMARY:ICS Colloquium: Dr. Dora Demszky\, Assistant Professor\, Education 
 Data Science\, Graduate School of Education and Computer Science\, Stanford
  University
UID:tag:localist.com\,2008:EventInstance_43951341829969
URL:https://calendar.colorado.edu/event/ics_colloquium_dr_dora_demszky_assi
 stant_professor_education_data_science_graduate_school_of_education_and_com
 puter_science_stanford_university
END:VEVENT
END:VCALENDAR