Certainly Uncertain: A Benchmark and Metric for Multimodal Epistemic and Aleatoric Awareness
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913413795414016 |
|---|---|
| author | Chandu, Khyathi Raghavi Li, Linjie Awadalla, Anas Lu, Ximing Park, Jae Sung Hessel, Jack Wang, Lijuan Choi, Yejin |
| author_facet | Chandu, Khyathi Raghavi Li, Linjie Awadalla, Anas Lu, Ximing Park, Jae Sung Hessel, Jack Wang, Lijuan Choi, Yejin |
| contents | The ability to acknowledge the inevitable uncertainty in their knowledge and reasoning is a prerequisite for AI systems to be truly truthful and reliable. In this paper, we present a taxonomy of uncertainty specific to vision-language AI systems, distinguishing between epistemic uncertainty (arising from a lack of information) and aleatoric uncertainty (due to inherent unpredictability), and further explore finer categories within. Based on this taxonomy, we synthesize a benchmark dataset, CertainlyUncertain, featuring 178K visual question answering (VQA) samples as contrastive pairs. This is achieved by 1) inpainting images to make previously answerable questions into unanswerable ones; and 2) using image captions to prompt large language models for both answerable and unanswerable questions. Additionally, we introduce a new metric confidence-weighted accuracy, that is well correlated with both accuracy and calibration error, to address the shortcomings of existing metrics. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_01942 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Certainly Uncertain: A Benchmark and Metric for Multimodal Epistemic and Aleatoric Awareness Chandu, Khyathi Raghavi Li, Linjie Awadalla, Anas Lu, Ximing Park, Jae Sung Hessel, Jack Wang, Lijuan Choi, Yejin Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition The ability to acknowledge the inevitable uncertainty in their knowledge and reasoning is a prerequisite for AI systems to be truly truthful and reliable. In this paper, we present a taxonomy of uncertainty specific to vision-language AI systems, distinguishing between epistemic uncertainty (arising from a lack of information) and aleatoric uncertainty (due to inherent unpredictability), and further explore finer categories within. Based on this taxonomy, we synthesize a benchmark dataset, CertainlyUncertain, featuring 178K visual question answering (VQA) samples as contrastive pairs. This is achieved by 1) inpainting images to make previously answerable questions into unanswerable ones; and 2) using image captions to prompt large language models for both answerable and unanswerable questions. Additionally, we introduce a new metric confidence-weighted accuracy, that is well correlated with both accuracy and calibration error, to address the shortcomings of existing metrics. |
| title | Certainly Uncertain: A Benchmark and Metric for Multimodal Epistemic and Aleatoric Awareness |
| topic | Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2407.01942 |