Clinical Trial NLI
Natural language inference for clinical trial reports, determining whether a given statement is entailed or contradicted by clinical trial evidence. Based on SemEval-2023 Task 7 (Jullien et al.).
Configuration Fileconfig.yaml
# Clinical Trial NLI
# Based on Jullien et al., SemEval 2023
# Paper: https://aclanthology.org/2023.semeval-1.307/
# Dataset: https://github.com/ai-systems/Task-7-SemEval-2023
#
# This task asks annotators to determine whether a statement about a clinical
# trial is entailed or contradicted by the provided evidence from the trial
# report. Annotators also provide an explanation of the evidence supporting
# their judgment.
#
# NLI Labels:
# - Entailment: The statement is supported by the clinical trial evidence
# - Contradiction: The statement is contradicted by the clinical trial evidence
annotation_task_name: "Clinical Trial NLI"
task_dir: "."
data_files:
- sample-data.json
item_properties:
id_key: "id"
text_key: "text"
output_annotation_dir: "annotation_output/"
output_annotation_format: "json"
port: 8000
server_name: localhost
annotation_schemes:
- annotation_type: radio
name: nli_judgment
description: "Is the statement entailed or contradicted by the clinical trial evidence?"
labels:
- "Entailment"
- "Contradiction"
keyboard_shortcuts:
"Entailment": "1"
"Contradiction": "2"
tooltips:
"Entailment": "The statement is supported by or can be inferred from the trial evidence"
"Contradiction": "The statement conflicts with or is negated by the trial evidence"
- annotation_type: text
name: evidence_explanation
description: "Explain which part of the evidence supports your judgment"
annotation_instructions: |
You will see a statement about a clinical trial and the corresponding trial evidence.
1. Read the statement carefully.
2. Read the clinical trial evidence provided.
3. Determine whether the statement is entailed (supported) or contradicted by the evidence.
4. Write a brief explanation of which evidence supports your judgment.
html_layout: |
<div style="padding: 15px; max-width: 800px; margin: auto;">
<div style="background: #eff6ff; border: 1px solid #bfdbfe; border-radius: 8px; padding: 12px; margin-bottom: 12px;">
<strong style="color: #1e40af;">Trial ID:</strong>
<span style="font-size: 14px; margin-left: 8px; font-family: monospace;">{{trial_id}}</span>
</div>
<div style="background: #fef3c7; border: 1px solid #fde68a; border-radius: 8px; padding: 16px; margin-bottom: 16px;">
<strong style="color: #92400e;">Statement:</strong>
<p style="font-size: 16px; line-height: 1.7; margin: 8px 0 0 0;">{{text}}</p>
</div>
<div style="background: #f0f9ff; border: 1px solid #bae6fd; border-radius: 8px; padding: 16px; margin-bottom: 16px;">
<strong style="color: #0369a1;">Trial Evidence:</strong>
<p style="font-size: 15px; line-height: 1.6; margin: 8px 0 0 0;">{{evidence}}</p>
</div>
</div>
allow_all_users: true
instances_per_annotator: 50
annotation_per_instance: 2
allow_skip: true
skip_reason_required: false
Sample Datasample-data.json
[
{
"id": "clinical_nli_001",
"text": "The treatment group showed statistically significant improvement in blood pressure levels compared to the placebo group.",
"evidence": "Primary endpoint: Mean systolic blood pressure reduction was 12.3 mmHg (SD 4.1) in the treatment arm vs. 2.1 mmHg (SD 3.8) in the placebo arm (p < 0.001). The treatment demonstrated clinically meaningful blood pressure reduction at week 12.",
"trial_id": "NCT04523817"
},
{
"id": "clinical_nli_002",
"text": "No adverse events were reported in the experimental group during the study period.",
"evidence": "Safety outcomes: In the experimental group, 23 patients (15.3%) reported at least one adverse event, including headache (8.7%), nausea (4.0%), and fatigue (2.6%). Three serious adverse events were recorded, though none were deemed related to the intervention.",
"trial_id": "NCT04198726"
}
]
// ... and 8 more itemsGet This Design
Clone or download from the repository
Quick start:
git clone https://github.com/davidjurgens/potato-showcase.git cd potato-showcase/semeval/2023/task07-clinical-trial-nli potato start config.yaml
Details
Annotation Types
Domain
Use Cases
Tags
Found an issue or want to improve this design?
Open an IssueRelated Designs
Patient Forum Question Answering
Question answering and response classification for patient health forum posts, categorizing responses by type (medical advice, experience sharing, information request, emotional support). Based on SemEval-2022 Task 6 (Akhtar et al.).
Safe Biomedical NLI
Safe biomedical natural language inference task requiring annotators to determine entailment or contradiction between clinical premise-hypothesis pairs and provide reasoning. Based on SemEval-2024 Task 2 (Safe Biomedical NLI).
Argument Reasoning in Civil Procedure
Legal argument reasoning task requiring annotators to answer multiple-choice questions about civil procedure by selecting the best answer and providing legal reasoning. Based on SemEval-2024 Task 5.