Synthetic Caregiver Prompts and Intent Labels for AAC Gesture Disambiguation

A fully synthetic dataset of 11.6 KB contains simulated caregiver speech prompts and communicative intent labels for context-aware AAC systems. It models interaction scenarios where caregiver speech helps disambiguate ambiguous gestures or short symbolic responses. Created by Andrea Zismam and last updated in May 2026, it includes pipeline outputs like semantic similarity rankings and runtime measurements.

Use Cases

• Training context-aware AAC systems based on simulated caregiver prompts and intent labels.
• Evaluating semantic intent inference algorithms based on the provided similarity rankings and inferred intents.
• Benchmarking lightweight language model routing strategies for CPU-only/offline assistive AI systems.
• Studying multi-stage intent classification pipelines based on the included resolution stages and runtime measurements.

Strengths

• Dataset is fully synthetic, containing no real patient data or personally identifiable information.
• Includes outputs from a multi-stage pipeline such as semantic similarity rankings and runtime measurements.
• Released under a permissive CC-BY-4.0 license.

Limitations

• Dataset is small at 11.6 KB, indicating a limited scope.
• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

figshare

Collection Method

Synthetic generation as part of a proof-of-concept evaluation for a research paper.

Freshness

Last updated 2026-05-19 15:39:37.