Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.19457 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917387223171072 |
|---|---|
| author | Jiang, Kailin Jiang, Ning Du, Yuntao Ren, Yuchen Li, Yuchen Gao, Yifan Bi, Jinhe Ma, Yunpu Li, Bin Liu, Lei Li, Qing |
| author_facet | Jiang, Kailin Jiang, Ning Du, Yuntao Ren, Yuchen Li, Yuchen Gao, Yifan Bi, Jinhe Ma, Yunpu Li, Bin Liu, Lei Li, Qing |
| contents | Large Multimodal Models (LMMs) encode rich factual knowledge via cross-modal pre-training, yet their static representations struggle to maintain an accurate understanding of time-sensitive factual knowledge. Existing benchmarks remain constrained by static designs, inadequately evaluating LMMs' ability to understand time-sensitive knowledge. To address this gap, we propose MINED, a comprehensive benchmark that evaluates temporal awareness along 6 key dimensions and 11 challenging tasks: cognition, awareness, trustworthiness, understanding, reasoning, and robustness. MINED is constructed from Wikipedia by two professional annotators, containing 2,104 time-sensitive knowledge samples spanning six knowledge types. Evaluating 15 widely used LMMs on MINED shows that Gemini-2.5-Pro achieves the highest average CEM score of 63.07, while most open-source LMMs still lack time understanding ability. Meanwhile, LMMs perform best on organization knowledge, whereas their performance is weakest on sport. To address these challenges, we investigate the feasibility of updating time-sensitive knowledge in LMMs through knowledge editing methods and observe that LMMs can effectively update knowledge via knowledge editing methods in single editing scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_19457 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MINED: Probing and Updating with Multimodal Time-Sensitive Knowledge for Large Multimodal Models Jiang, Kailin Jiang, Ning Du, Yuntao Ren, Yuchen Li, Yuchen Gao, Yifan Bi, Jinhe Ma, Yunpu Li, Bin Liu, Lei Li, Qing Computation and Language Large Multimodal Models (LMMs) encode rich factual knowledge via cross-modal pre-training, yet their static representations struggle to maintain an accurate understanding of time-sensitive factual knowledge. Existing benchmarks remain constrained by static designs, inadequately evaluating LMMs' ability to understand time-sensitive knowledge. To address this gap, we propose MINED, a comprehensive benchmark that evaluates temporal awareness along 6 key dimensions and 11 challenging tasks: cognition, awareness, trustworthiness, understanding, reasoning, and robustness. MINED is constructed from Wikipedia by two professional annotators, containing 2,104 time-sensitive knowledge samples spanning six knowledge types. Evaluating 15 widely used LMMs on MINED shows that Gemini-2.5-Pro achieves the highest average CEM score of 63.07, while most open-source LMMs still lack time understanding ability. Meanwhile, LMMs perform best on organization knowledge, whereas their performance is weakest on sport. To address these challenges, we investigate the feasibility of updating time-sensitive knowledge in LMMs through knowledge editing methods and observe that LMMs can effectively update knowledge via knowledge editing methods in single editing scenarios. |
| title | MINED: Probing and Updating with Multimodal Time-Sensitive Knowledge for Large Multimodal Models |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2510.19457 |