Keiji AI LogoKeiji AI

Research · EMNLP 2023

AutoTrial

Research on eligibility criteria design. AutoTrial helps design a trial's inclusion and exclusion criteria with a language model, drawing on similar earlier trials and giving the reasoning behind each output.

Who created AutoTrial?

AutoTrial was written by Keiji AI co-founders Zifeng Wang and Jimeng Sun, who is Keiji AI's CEO, with Cao Xiao.

What the paper found

  • Tested on over 70,000 clinical trials.
  • Won about 60% of human evaluations against GPT-3.5 baselines, with a much smaller parameter size.

How AutoTrial relates to TrialMind

It is research, not a TrialMind module or a separately sold product. Published methods like this one inform how TrialMind works; TrialMind is the product Keiji AI sells.

See how TrialMind handles this work: Biostatistics and trial design.

The paper

Zifeng Wang, Cao Xiao, Jimeng Sun. AutoTrial: Prompting Language Models for Clinical Trial Design. EMNLP, 2023.

Details checked against the published record on 2026-11-01.

Frequently asked questions

Who created AutoTrial?
AutoTrial was written by Keiji AI co-founders Zifeng Wang and Jimeng Sun, who is Keiji AI's CEO, with Cao Xiao.
Is AutoTrial part of TrialMind?
No. It is research, not a TrialMind module or a separately sold product. Published methods like this one inform how TrialMind works; TrialMind is the product Keiji AI sells.
What did the AutoTrial paper find?
Tested on over 70,000 clinical trials. Won about 60% of human evaluations against GPT-3.5 baselines, with a much smaller parameter size.

All publications

Every paper sits under its primary research area.