MuLA-Bench exposes conditional failure patterns that a single long-context score does not capture, and evaluates ten audio-language models and conducts pooled diagnostics on a fixed eight-model cohort.
Abstract
Long-form audio performance is often summarized by context length and aggregate accuracy, obscuring how language, evidence, and task jointly shape difficulty. We introduce MuLA-Bench: 5,038 open-ended questions over 1,769 in-the-wild recordings totaling 1,377.9 hours, covering 16 languages and eight domains. A balanced Language x Domain semantic track supports controlled comparisons, while a complementary acoustic track preserves naturally occurring non-speech evidence. Evidence-grounded generation, shortcut checks, and language-expert review provide auditable questions without translating a shared source set or injecting target sounds. We evaluate ten audio-language models and conduct pooled diagnostics on a fixed eight-model cohort. Language rankings change across domains and tasks; acoustic-semantic performance gaps vary with the requested operation; and temporal errors can persist after the correct event is identified. Long-range retrieval is comparatively strong, while precise clock alignment and factual grounding of natural acoustic events remain fragile. MuLA-Bench thus exposes conditional failure patterns that a single long-context score does not capture.
AudioICL-Bench is introduced, a diagnostic benchmark whose per-episode rules are resampled so that no correct answer is recoverable from prior knowledge, and its nine tasks are organized along two axes that separate what must be learned from demonstrations from what must be perceived in the signal, enabling failures to...
Jia-Hung Chen, Yi-Cheng Lin, Kai-Wei Chang et al.· 0 citations
FuseAlign, a transformer-based aligner trained on large-scale pseudo-labeled speech with online label correction, is introduced, a transformer-based aligner trained on large-scale pseudo-labeled speech with online label correction that substantially outperforms all baselines and remains robust under real ASR transcript...
Mithilesh Vaidya, Stephen W. Bailey, Sumukh Badam et al.· 0 citations
Long-context understanding remains a fundamental challenge for large language models, as excessively long inputs often lead models to forget salient information. This issue is even more pronounced in the speech domain, where audio, as a low-compression modality, requires substantially more embeddings than text to prese...
SEA-SpeechBench is introduced, the first large-scale multitask benchmark that evaluates speech understanding in 11 SEA languages through 97,194 samples across 99 evaluation sets and 597 hours of curated audio data, exposing critical model limitations and underscore the need for inclusive model development.
Jingyi Liao, Wenyu Zhang, Zhuo-Han Liu et al.· 1 citation
Large audio language models (LALMs) can describe what is heard, but their ability to localize when queried content occurs remains less systematically evaluated. We present TAG-Bench, a benchmark for temporal audio grounding in which a model returns every time interval that matches a natural-language query. TAG-Bench co...
Yu-Hang Dai, Xin Shu, Zeng-Xi Li et al.· 2 citations
Mizar, a 159.3M-parameter ALM, is introduced, a recipe that brings together architecture, data, and three-stage training to build Mizar, a 159.3M-parameter ALM that surpasses the previous best-performing ALM below 200M parameters on all three benchmarks.