Uncertainty Quantification for Multimodal Large Language Models with Incoherence-adjusted Semantic Volume
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908856167170048 |
|---|---|
| author | Lau, Gregory Kang Ruey Dao, Hieu Lin, Nicole Kan Hui Low, Bryan Kian Hsiang |
| author_facet | Lau, Gregory Kang Ruey Dao, Hieu Lin, Nicole Kan Hui Low, Bryan Kian Hsiang |
| contents | Despite their capabilities, Multimodal Large Language Models (MLLMs) may produce plausible but erroneous outputs, hindering reliable deployment. Accurate uncertainty metrics could enable escalation of unreliable queries to human experts or larger models for improved performance. However, existing uncertainty metrics have practical constraints, such as being designed only for specific modalities, reliant on external tools, or computationally expensive. We introduce UMPIRE, a training-free uncertainty quantification framework for MLLMs that works efficiently across various input and output modalities without external tools, relying only on the models' own internal modality features. UMPIRE computes the incoherence-adjusted semantic volume of sampled MLLM responses for a given task instance, effectively capturing both the global semantic diversity of samples and the local incoherence of responses based on internal model confidence. We propose uncertainty desiderata for MLLMs and provide theoretical analysis motivating UMPIRE's design. Extensive experiments show that UMPIRE consistently outperforms baseline metrics in error detection and uncertainty calibration across image, audio, and video-text benchmarks, including adversarial and out-of-distribution settings. We also demonstrate UMPIRE's generalization to non-text output tasks, including image and audio generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_24195 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Uncertainty Quantification for Multimodal Large Language Models with Incoherence-adjusted Semantic Volume Lau, Gregory Kang Ruey Dao, Hieu Lin, Nicole Kan Hui Low, Bryan Kian Hsiang Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition Machine Learning Despite their capabilities, Multimodal Large Language Models (MLLMs) may produce plausible but erroneous outputs, hindering reliable deployment. Accurate uncertainty metrics could enable escalation of unreliable queries to human experts or larger models for improved performance. However, existing uncertainty metrics have practical constraints, such as being designed only for specific modalities, reliant on external tools, or computationally expensive. We introduce UMPIRE, a training-free uncertainty quantification framework for MLLMs that works efficiently across various input and output modalities without external tools, relying only on the models' own internal modality features. UMPIRE computes the incoherence-adjusted semantic volume of sampled MLLM responses for a given task instance, effectively capturing both the global semantic diversity of samples and the local incoherence of responses based on internal model confidence. We propose uncertainty desiderata for MLLMs and provide theoretical analysis motivating UMPIRE's design. Extensive experiments show that UMPIRE consistently outperforms baseline metrics in error detection and uncertainty calibration across image, audio, and video-text benchmarks, including adversarial and out-of-distribution settings. We also demonstrate UMPIRE's generalization to non-text output tasks, including image and audio generation. |
| title | Uncertainty Quantification for Multimodal Large Language Models with Incoherence-adjusted Semantic Volume |
| topic | Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2602.24195 |