SpectCount fine-tunes LALMs using on-the-fly synthetic signals to fix identified spectrotemporal weaknesses and boost performance on unseen auditory benchmarks.
SpectCount: Spectrotemporal Counting via Synthetic Signals Improves Large Audio Language Models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
Large audio language models (LALMs) extend large language models with an audio encoder and large-scale audio data. However, the scarcity of high-quality annotated audio data remains a fundamental bottleneck for scaling. Through probing signal detectability analysis, we identify fine-grained spectrotemporal perceptual weaknesses in a foundation LALM. To address these challenges, we propose Spectrotemporal Counting (SpectCount), a data-efficient fine-tuning approach based on fully synthetic audio signals generated on-the-fly, without relying on real-world audio, annotations, or pretrained generative models. SpectCount not only resolves the observed weaknesses but also improves performance on diverse auditory benchmarks spanning sound, music, and speech, unseen during fine-tuning. These results suggest that weakness-targeted synthetic signals provide a data-efficient path toward enhanced auditory understanding capabilities in LALMs.
fields
eess.AS 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SpectCount: Spectrotemporal Counting via Synthetic Signals Improves Large Audio Language Models
SpectCount fine-tunes LALMs using on-the-fly synthetic signals to fix identified spectrotemporal weaknesses and boost performance on unseen auditory benchmarks.