Program · Tentative
A research day, then a working day
Day 1 is a research day with keynotes, a panel, talks and posters. Day 2 is a working day: a few working groups sharing concerns, then open discussion. Both days run 10 AM to 5 PM, with coffee and light snacks; lunch is on your own in Kendall Square.
Day 1 · Research day
Up to 500 people.
- 01Keynotes
Two keynotes: one from agent evaluation, one from a field with a longer evaluation tradition.
- 02Panel
A 45-minute panel on a question the field disagrees about, with panelists from academia and from companies that deploy agents.
- 03Group overviews
Ten invited groups present their setting, their approach and one unresolved problem, eight minutes each.
- 04Contributed talks
Eight talks from an open call, including industry case studies of evaluating agents in production.
- 05Posters and demos
60–80 posters and demos across two sessions, open to early-stage work and tools.
Day 2 · Working day
150–200 people.
- 01Working groups
A few working groups, each around one theme, where participants share what they evaluate and the problems they have run into.
- 02Open discussion
The groups come back together for an open discussion of the shared problems and disagreements.
- 03Follow-up
A public summary follows within six weeks.