Search papers, labs, and topics across Lattice.
This paper introduces HyPASE, a hyperbolic geometry-based framework designed for parameter-efficient fine-tuning of large audio-language models specifically for speech emotion recognition tasks. By utilizing the Poincare ball model, HyPASE effectively captures the multi-granularity nature of emotion cues, leading to significant improvements in performance on standard benchmarks such as MELD and IEMOCAP. The framework's innovative components, including a Hyperbolic Geometric Adapter and an Emotion-aware Multi-capacity Cross-modal Aggregator, enable robust zero-shot cross-dataset generalization while maintaining a constrained parameter budget.
Fine-tuning large audio-language models for emotion recognition can be drastically improved by leveraging hyperbolic geometry, leading to better performance on class-imbalanced datasets.
Large Audio-Language Models (LALMs) excel at general speech understanding; however, adapting them to fine-grained tasks like Speech Emotion Recognition (SER) remains a significant bottleneck. Current Parameter-Efficient Fine-Tuning (PEFT) methods typically operate in flat Euclidean space, and this geometry fails to capture the multi-granularity nature of emotion cues, which range from low-level prosody to high-level semantics. To address this, we propose HyPASE, a hyperbolic PEFT framework for LALM-based SER. HyPASE leverages the Poincare ball model, using the hyperbolic radius as an explicit proxy for representational granularity. The framework consists of two core components: a Hyperbolic Geometric Adapter (HGA) for layer-adaptive weight modulation, and an Emotion-aware Multi-capacity Cross-modal Aggregator (EMCA) that compresses multi-scale features into compact audio prefixes. Empirical results on standard benchmarks show that HyPASE outperforms Euclidean PEFT baselines across all metrics on MELD and achieves a notable Unweighted Accuracy gain on IEMOCAP, particularly in class-imbalanced emotion recognition, with the accompanying slight Weighted Accuracy trade-off reflecting hyperbolic space's geometric prioritization of minority-class representations; furthermore, HyPASE achieves robust zero-shot cross-dataset generalization within a constrained parameter budget. By grounding the adaptation process in hyperbolic geometry, HyPASE offers a highly efficient path for LALM fine-tuning.
