Entity · paper

An Evaluation Framework for Structured Audio Captions Validated by Controlled Perturbations

paperactiveprovisionalan-evaluation-framework-for-structured-audio-captions-validated-by-controlled-perturbations-5c01d915·1 events·first seen Jul 24, 2026

Aliases: An Evaluation Framework for Structured Audio Captions Validated by Controlled Perturbations

Co-occurring entities

AudioCards

More like this (12)

Auditing Protocol-Level Shortcuts in Large Audio Language Model Judges for Speech Evaluation Block-Compositional Caption Supervision How Do Instructions Shape Speech? Cross-Attention Attribution for Style-Captioned Text-to-Speech From Black-Box to Clinical Insight: A Multi-Stage Explainable Framework for Speech-Based Cognitive Impairment Detection Acoustic Cue Alignment in Audio Language Models for Speech Emotion Recognition Audio Interaction Model DirectAudioEdit: Inversion-Free Text-Guided Audio Editing via Diffusion Prediction Contrast ENTRAP-VL: A Taxonomic Probe for Dual Contextual Entrainment in Vision-Language Models How Robust is OCR-Reasoning? Evaluating OCR-Reasoning Robustness of Vision-Language Models under Visual Perturbations Probing Low Frame Rate Degradation in Neural Audio Codecs Progressive Refinement: An Iterative Pseudo-Labeling Approach for Mandarin-English Code-Switching ASR Leveraging Audio-LLMs to Filter Speech-to-Speech Training Data

Recent events (1)

4arXiv · cs.CL·Jul 24, 2026·source ↗

Multi-axis evaluation framework for structured audio captions validated via controlled perturbations

A new arXiv preprint proposes a multi-axis evaluation framework for structured audio captions, addressing limitations of existing flat-text metrics when applied to heterogeneous multimodal outputs. The framework evaluates outputs across five axes—tag-sets, descriptions, logical reasoning, numeric measurements, and spectral profiles—combining LLM judges with deterministic computational metrics. Validation uses a controlled perturbation protocol that injects typed, graded errors into ground-truth annotations to confirm the framework distinguishes paraphrases from genuine corruptions. The work builds on the AudioCards dataset.

Evaluation and Benchmarking Multimodal Progress An Evaluation Framework for Structured Audio Captions Validated by Controlled Perturbations AudioCards