尽管EDL及其相关工作取得了显著成功,我们认为现有的EDL方法(第3.1节)在狄利克雷分布的构建上(定理1)和优化目标的设计上保持了一些死板的设置,这些设置虽然被广泛接受,但并不是主观逻辑框架(第2节)本质上所要求的。本节的理论分析和第5节的实验均表明,这些非必要的设置阻碍了该类方法更准确地量化不确定性。具体来说,本节我们仔细分析并放宽了EDL中的两种非必要设置,包括:(1)在模型构建中,先验权重参数被固定为类别数(第3.2节);(2)在模型优化中,传统的优化目标包含一个最小化方差的正则项,这可能会加剧过度自信(第3.3节)。需要注意的是,我们对上述EDL非必要设置的放宽都是严格遵循主观逻辑理论的。
当先验权重W被设置为零时,等式 7 中的投影概率将退化为传统的概率形式,这种形式仅依赖于各类证据的比例,而不受证据大小的影响,因为将证据按常数系数缩放不会影响投影概率。然而,当W不为零时,我们有
根据 [22],我们重点与其他基于狄利克雷的不确定性方法进行比较,包括传统的EDL[10]、I-EDL [22]、KL-PN [33]、RKL-PN [34] 和 PostN [35]。此外,我们还展示了有代表性的单前向传播方法DUQ [31] 和流行的贝叶斯不确定性方法MC Dropout [6] 的结果以供参考。
1. Choi, J., Chun, D., Kim, H., & Lee, H.-J. (2019). Gaussian yolov3: An accurate and fast object detector using localization uncertainty for autonomous driving. Proceedings of the IEEE/CVF International Conference on Computer Vision, 502–511.
2. Abdar, M., Khosravi, A., Islam, S. M. S., Acharya, U. R., & Vasilakos, A. V. (2022). The need for quantification of uncertainty in artificial intelligence for clinical data analysis: Increasing the level of trust in the decision-making process. IEEE Systems, Man, and Cybernetics Magazine, 8(3), 28–40.
3. Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. (2017). On calibration of modern neural networks. International Conference on Machine Learning, 1321–1330.
4. Blundell, C., Cornebise, J., Kavukcuoglu, K., & Wierstra, D. (2015). Weight uncertainty in neural network. International Conference on Machine Learning, 1613–1622.
5. Dusenberry, M., Jerfel, G., Wen, Y., Ma, Y., Snoek, J., Heller, K., Lakshminarayanan, B., & Tran, D. (2020). Efficient and scalable bayesian neural nets with rank-1 factors. International Conference on Machine Learning, 2782–2792.
6. Gal, Y., & Ghahramani, Z. (2016). Dropout as a bayesian approximation: Representing model uncertainty in deep learning. International Conference on Machine Learning, 1050–1059.
7. Lakshminarayanan, B., Pritzel, A., & Blundell, C. (2017). Simple and scalable predictive uncertainty estimation using deep ensembles. Advances in Neural Information Processing Systems, 30.
8. Wen, Y., Tran, D., & Ba, J. (2020). Batchensemble: An alternative approach to efficient ensemble and lifelong learning. arXiv Preprint arXiv:2002.06715.
9. Egele, R., Maulik, R., Raghavan, K., Lusch, B., Guyon, I., & Balaprakash, P. (2022). Autodeuq: Automated deep ensemble with uncertainty quantification. 2022 26th International Conference on Pattern Recognition (ICPR), 1908–1914.
10. Sensoy, M., Kaplan, L., & Kandemir, M. (2018). Evidential deep learning to quantify classification uncertainty. Advances in Neural Information Processing Systems, 31.
11. Amini, A., Schwarting, W., Soleimany, A., & Rus, D. (2020). Deep evidential regression. Advances in Neural Information Processing Systems, 33, 14927–14937.
12. Bao, W., Yu, Q., & Kong, Y. (2021). Evidential deep learning for open set action recognition. Proceedings of the IEEE/CVF International Conference on Computer Vision, 13349–13358.
13. Qin, Y., Peng, D., Peng, X., Wang, X., & Hu, P. (2022). Deep evidential learning with noisy correspondence for cross-modal retrieval. Proceedings of the 30th ACM International Conference on Multimedia, 4948–4956.
14. Chen, M., Gao, J., Yang, S., & Xu, C. (2022). Dual-evidential learning for weakly-supervised temporal action localization. European Conference on Computer Vision, 192–208.
15. Oh, D., & Shin, B. (2022). Improving evidential deep learning via multi-task learning. Proceedings of the AAAI Conference on Artificial Intelligence, 36, 7895–7903.
16. Sun, S., Zhi, S., Heikkilä, J., & Liu, L. (2022). Evidential uncertainty and diversity guided active learning for scene graph generation. The Eleventh International Conference on Learning Representations.
17. Park, Y., Choi, W., Kim, S., Han, D.-J., & Moon, J. (2022). Active learning for object detection with evidential deep learning and hierarchical uncertainty aggregation. The Eleventh International Conference on Learning Representations.
18. Sapkota, H., & Yu, Q. (2022). Adaptive robust evidential optimization for open set detection from imbalanced data. The Eleventh International Conference on Learning Representations.
19. Gao, J., Chen, M., & Xu, C. (2023). Collecting cross-modal presence-absence evidence for weakly-supervised audio-visual event perception. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 18827–18836.
20. Jøsang, A. (2001). A logic for uncertain probabilities. International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 9(03), 279–311.
21. Jøsang, A. (2016). Subjective logic (Vol. 3). Springer.
22. Deng, D., Chen, G., Yu, Y., Liu, F., & Heng, P.-A. (2023). Uncertainty estimation by fisher information-based evidential deep learning. arXiv Preprint arXiv:2303.02045.
23. Soleimany, A. P., Amini, A., Goldman, S., Rus, D., Bhatia, S. N., & Coley, C. W. (2021). Evidential deep learning for guided molecular property prediction and discovery. ACS Central Science, 7(8), 1356–1367.
24. Kandemir, M., Akgül, A., Haussmann, M., & Unal, G. (2022). Evidential turing processes. International Conference on Learning Representations.
25. Meinert, N., Gawlikowski, J., & Lavin, A. (2023). The unreasonable effectiveness of deep evidential regression. Proceedings of the AAAI Conference on Artificial Intelligence, 37, 9134–9142.
26. Chen, M., Gao, J., & Xu, C. (2023). Cascade evidential learning for open-world weakly-supervised temporal action localization. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 14741–14750.
27. Chen, M., Gao, J., & Xu, C. (2023). Uncertainty-aware dual-evidential learning for weakly-supervised temporal action localization. IEEE Transactions on Pattern Analysis and Machine Intelligence.
28. Gao, J., Chen, M., & Xu, C. (2023). Vectorized evidential learning for weakly-supervised temporal action localization. IEEE Transactions on Pattern Analysis and Machine Intelligence.
29. Ritter, H., Botev, A., & Barber, D. (2018). A scalable laplace approximation for neural networks. 6th International Conference on Learning Representations, ICLR 2018-Conference Track Proceedings, 6.
30. Izmailov, P., Vikram, S., Hoffman, M. D., & Wilson, A. G. G. (2021). What are bayesian neural network posteriors really like? International Conference on Machine Learning, 4629–4640.
31. Van Amersfoort, J., Smith, L., Teh, Y. W., & Gal, Y. (2020). Uncertainty estimation using a single deep deterministic neural network. International Conference on Machine Learning, 9690–9700.
32. Liu, J., Lin, Z., Padhy, S., Tran, D., Bedrax Weiss, T., & Lakshminarayanan, B. (2020). Simple and principled uncertainty estimation with deterministic deep learning via distance awareness. Advances in Neural Information Processing Systems, 33, 7498–7512.
33. Malinin, A., & Gales, M. (2018). Predictive uncertainty estimation via prior networks. Advances in Neural Information Processing Systems, 31.
34. Malinin, A., & Gales, M. (2019). Reverse kl-divergence training of prior networks: Improved uncertainty and adversarial robustness. Advances in Neural Information Processing Systems, 32.
35. Charpentier, B., Zügner, D., & Günnemann, S. (2020). Posterior network: Uncertainty estimation without ood samples via density-based pseudo-counts. Advances in Neural Information Processing Systems, 33, 1356–1367.
36. Bendale, A., & Boult, T. E. (2016). Towards open set deep networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1563–1572.
37. Krishnan, R., Subedar, M., & Tickoo, O. (2018). BAR: Bayesian activity recognition using variational inference. arXiv Preprint arXiv:1811.03305.
38. Chen, G., Qiao, L., Shi, Y., Peng, P., Li, J., Huang, T., Pu, S., & Tian, Y. (2020). Learning open set network with discriminative reciprocal points. Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part III 16, 507–522.
39. McAllester, D. A. (1998). Some pac-bayesian theorems. Proceedings of the Eleventh Annual Conference on Computational Learning Theory, 230–234.
40. LeCun, Y. (1998). The MNIST database of handwritten digits. Http://Yann. Lecun. Com/Exdb/Mnist/.
41. Xiao, H., Rasul, K., & Vollgraf, R. (2017). Fashion-mnist: A novel image dataset for benchmarking machine learning algorithms. arXiv Preprint arXiv:1708.07747.
42. Clanuwat, T., Bober-Irizar, M., Kitamoto, A., Lamb, A., Yamamoto, K., & Ha, D. (2018). Deep learning for classical japanese literature. arXiv Preprint arXiv:1812.01718.
43. Krizhevsky, A., Hinton, G., et al. (2009). Learning multiple layers of features from tiny images.
44. Netzer, Y., Wang, T., Coates, A., Bissacco, A., Wu, B., & Ng, A. (2018). The street view house numbers (SVHN) dataset. Technical report, Accessed 2016-08-01.[Online].
45. Vinyals, O., Blundell, C., Lillicrap, T., Wierstra, D., et al. (2016). Matching networks for one shot learning. Advances in Neural Information Processing Systems, 29.
46. Wah, C., Branson, S., Welinder, P., Perona, P., & Belongie, S. (2011). The caltech-ucsd birds-200-2011 dataset.
47. Soomro, K., Zamir, A. R., & Shah, M. (2012). UCF101: A dataset of 101 human actions classes from videos in the wild. arXiv Preprint arXiv:1212.0402.
48. Kuehne, H., Jhuang, H., Garrote, E., Poggio, T., & Serre, T. (2011). HMDB: A large video database for human motion recognition. 2011 International Conference on Computer Vision, 2556–2563.
49. Monfort, M., Pan, B., Ramakrishnan, K., Andonian, A., McNamara, B. A., Lascelles, A., Fan, Q., Gutfreund, D., Feris, R. S., & Oliva, A. (2021). Multi-moments in time: Learning and interpreting models for multi-action video understanding. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(12), 9434–9445.
50. Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv Preprint arXiv:1409.1556.
51. Yang, S., Liu, L., & Xu, M. (2021). Free lunch for few-shot learning: Distribution calibration. arXiv Preprint arXiv:2101.06395.
52. Gao, J., Zhang, T., & Xu, C. (2020). Learning to model relationships for zero-shot video classification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(10), 3476–3491.