Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
153 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Open-world Machine Learning: A Review and New Outlooks (2403.01759v2)

Published 4 Mar 2024 in cs.LG and cs.CV

Abstract: Machine learning has achieved remarkable success in many applications. However, existing studies are largely based on the closed-world assumption, which assumes that the environment is stationary, and the model is fixed once deployed. In many real-world applications, this fundamental and rather naive assumption may not hold because an open environment is complex, dynamic, and full of unknowns. In such cases, rejecting unknowns, discovering novelties, and then incrementally learning them, could enable models to be safe and evolve continually as biological systems do. This paper provides a holistic view of open-world machine learning by investigating unknown rejection, novel class discovery, and class-incremental learning in a unified paradigm. The challenges, principles, and limitations of current methodologies are discussed in detail. Finally, we discuss several potential directions for future research. This paper aims to provide a comprehensive introduction to the emerging open-world machine learning paradigm, to help researchers build more powerful AI systems in their respective fields, and to promote the development of artificial general intelligence.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (210)
  1. Multimodal biomedical ai. Nature Medicine, 28(9):1773–1784, 2022.
  2. Machine learning-informed and synthetic biology-enabled semi-continuous algal cultivation to unleash renewable fuel productivity. Nature Communications, 13(1):541, 2022.
  3. Scientific discovery in the age of artificial intelligence. Nature, 620(7972):47–60, 2023.
  4. The elements of statistical learning: data mining, inference, and prediction, volume 2. Springer, 2009.
  5. Lifelong machine learning, volume 1. Springer, 2018.
  6. Towards open world recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1893–1902, 2015.
  7. Towards robust pattern recognition: A review. Proceedings of the IEEE, 108(6):894–922, 2020.
  8. Zhi-Hua Zhou. Open-environment machine learning. National Science Review, 9(8):nwac123, 2022.
  9. Class-incremental learning: survey and performance evaluation on image classification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(5):5513–5533, 2022.
  10. A baseline for detecting misclassified and out-of-distribution examples in neural networks. In International Conference on Learning Representations, 2016.
  11. Learning by seeing more classes. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(6):7477–7493, 2022.
  12. A survey on learning to reject. Proceedings of the IEEE, 111(2):185–215, 2023.
  13. Autonovel: Automatically discovering and learning novel visual categories. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(10):6767–6781, 2021.
  14. Catastrophic interference in connectionist networks: The sequential learning problem. In Psychology of learning and motivation, volume 24, pages 109–165. Elsevier, 1989.
  15. Revisiting confidence estimation: Towards reliable failure prediction. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023.
  16. Three types of incremental learning. Nature Machine Intelligence, 4(12):1185–1197, 2022.
  17. Continual learning: Applications and the road forward. arXiv preprint arXiv:2311.11908, 2023.
  18. A comprehensive survey of continual learning: Theory, method and application. arXiv preprint arXiv:2302.00487, 2023.
  19. A probabilistic challenge for object detection. Nature Machine Intelligence, 1(9):443–443, 2019.
  20. Reliable and trustworthy machine learning for health using dataset shift detection. Advances in Neural Information Processing Systems, 34:3043–3056, 2021.
  21. Clinical applications of continual learning machine learning. The Lancet Digital Health, 2(6):e279–e281, 2020.
  22. Continual learning in medical devices: Fda’s action plan and beyond. The Lancet Digital Health, 3(6):e337–e338, 2021.
  23. OpenAI. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023.
  24. Exploring ai ethics of chatgpt: A diagnostic analysis. arXiv preprint arXiv:2301.12867, 2023.
  25. Reliable visual question answering: Abstain rather than answer incorrectly. In European Conference on Computer Vision, pages 148–166. Springer, 2022.
  26. Ai autonomy: Self-initiated open-world continual learning and adaptation. AI Magazine, 2023.
  27. Anomaly detection: A survey. ACM computing surveys (CSUR), 41(3):1–58, 2009.
  28. Deep learning for anomaly detection: A review. ACM computing surveys (CSUR), 54(2):1–38, 2021.
  29. Generalized out-of-distribution detection: A survey. arXiv preprint arXiv:2110.11334, 2021.
  30. A unified survey on anomaly, novelty, open-set, and out of-distribution detection: Solutions and future challenges. Transactions on Machine Learning Research, 2022.
  31. Toward open set recognition. IEEE transactions on pattern analysis and machine intelligence, 35(7):1757–1772, 2012.
  32. Recent advances in open set recognition: A survey. IEEE transactions on pattern analysis and machine intelligence, 43(10):3614–3631, 2020.
  33. Deep anomaly detection with outlier exposure. In International Conference on Learning Representations, 2019.
  34. Mixture outlier exposure: Towards out-of-distribution detection in fine-grained environments. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 5531–5540, 2023.
  35. Open-set recognition: A good closed-set classifier is all you need. In International Conference on Learning Representations, 2022.
  36. Openood: Benchmarking generalized out-of-distribution detection. Advances in Neural Information Processing Systems, 35:32598–32611, 2022.
  37. Feed two birds with one scone: Exploiting wild data for both out-of-distribution generalization and detection. In International Conference on Machine Learning, pages 1454–1471. PMLR, 2023.
  38. On the impact of spurious correlation for out-of-distribution detection. In Proceedings of the AAAI Conference on Artificial Intelligence, 2022.
  39. The relationship between precision-recall and roc curves. In Proceedings of the 23rd international conference on Machine learning, pages 233–240, 2006.
  40. Openood v1. 5: Enhanced benchmark for out-of-distribution detection. arXiv preprint arXiv:2306.09301, 2023.
  41. Enhancing the reliability of out-of-distribution image detection in neural networks. In International Conference on Learning Representations, 2018.
  42. Energy-based out-of-distribution detection. Advances in neural information processing systems, 33:21464–21475, 2020.
  43. A tutorial on energy-based learning. Predicting structured data, 1(0), 2006.
  44. Scaling out-of-distribution detection for real-world settings. In International Conference on Machine Learning, pages 8759–8773. PMLR, 2022.
  45. A simple unified framework for detecting out-of-distribution samples and adversarial attacks. Advances in neural information processing systems, 31, 2018.
  46. Glow: Generative flow with invertible 1x1 convolutions. Advances in neural information processing systems, 31, 2018.
  47. Ev Zisselman and Aviv Tamar. Deep residual flow for out of distribution detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13994–14003, 2020.
  48. Out-of-distribution detection with deep nearest neighbors. In International Conference on Machine Learning, pages 20827–20840. PMLR, 2022.
  49. React: Out-of-distribution detection with rectified activations. Advances in Neural Information Processing Systems, 34:144–157, 2021.
  50. Vim: Out-of-distribution with virtual-logit matching. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4921–4930, 2022.
  51. Using self-supervised learning can improve model robustness and uncertainty. Advances in neural information processing systems, 32, 2019.
  52. Mitigating neural network overconfidence with logit normalization. In International Conference on Machine Learning, pages 23631–23644. PMLR, 2022.
  53. Unsupervised representation learning by predicting image rotations. In International Conference on Learning Representations, 2018.
  54. A simple framework for contrastive learning of visual representations. In International conference on machine learning, pages 1597–1607. PMLR, 2020.
  55. Csi: Novelty detection via contrastive learning on distributionally shifted instances. Advances in neural information processing systems, 33:11839–11852, 2020.
  56. How to exploit hyperspherical embeddings for out-of-distribution detection? In The Eleventh International Conference on Learning Representations, 2023.
  57. Mos: Towards scaling out-of-distribution detection for large semantic space. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8710–8719, 2021.
  58. Openmix: Exploring outlier samples for misclassification detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12074–12083, 2023.
  59. Distance-based image classification: Generalizing to new classes at near-zero cost. IEEE transactions on pattern analysis and machine intelligence, 35(11):2624–2637, 2013.
  60. He Zhang and Vishal M Patel. Sparse representation-based open set recognition. IEEE transactions on pattern analysis and machine intelligence, 39(8):1690–1696, 2016.
  61. Towards open set deep networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1563–1572, 2016.
  62. Learning placeholders for open-set recognition. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4401–4410, 2021.
  63. Reducing network agnostophobia. Advances in Neural Information Processing Systems, 31, 2018.
  64. Deep transfer learning for multiple class novelty detection. In Proceedings of the ieee/cvf conference on computer vision and pattern recognition, pages 11544–11552, 2019.
  65. C2ae: Class conditioned auto-encoder for open-set recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2307–2316, 2019.
  66. Conditional gaussian distribution learning for open set recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13480–13489, 2020.
  67. Generative openmax for multi-class open set classification. BMVC, 2017.
  68. Generative adversarial nets. Advances in neural information processing systems, 27, 2014.
  69. Opengan: Open-set recognition via open data generation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 813–822, 2021.
  70. Difficulty-aware simulator for open set recognition. In European Conference on Computer Vision, pages 365–381. Springer, 2022.
  71. Open set learning with counterfactual images. In Proceedings of the European Conference on Computer Vision (ECCV), pages 613–628, 2018.
  72. Hybrid models for open set recognition. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part III 16, pages 102–117. Springer, 2020.
  73. Residual flows for invertible generative modeling. Advances in Neural Information Processing Systems, 32, 2019.
  74. Why normalizing flows fail to detect out-of-distribution data. Advances in neural information processing systems, 33:20578–20589, 2020.
  75. Robust classification with convolutional prototype learning. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3474–3482, 2018.
  76. Convolutional prototype network for open set recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(5):2358–2370, 2020.
  77. Learning open set network with discriminative reciprocal points. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part III 16, pages 507–522. Springer, 2020.
  78. Adversarial reciprocal points learning for open set recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(11):8065–8081, 2021.
  79. Class-specific semantic reconstruction for open set recognition. IEEE transactions on pattern analysis and machine intelligence, 45(4):4214–4228, 2022.
  80. Randaugment: Practical automated data augmentation with a reduced search space. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 702–703, 2020.
  81. Rethinking the inception architecture for computer vision. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2818–2826, 2016.
  82. The devil is in the wrongly-classified samples: Towards unified open-set recognition. In The Eleventh International Conference on Learning Representations, 2023.
  83. The semantic shift benchmark. In ICML 2022 Shift Happens Workshop, 2022.
  84. Novel class discovery: an introduction and key concepts. arXiv preprint arXiv:2302.12028, 2023.
  85. Unsupervised deep embedding for clustering analysis. In International conference on machine learning, pages 478–487. PMLR, 2016.
  86. Deep clustering for unsupervised learning of visual features. In Proceedings of the European conference on computer vision (ECCV), pages 132–149, 2018.
  87. Learning to discover novel visual categories via deep transfer clustering. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 8401–8409, 2019.
  88. Supervised knowledge may hurt novel class discovery performance. Transactions on Machine Learning Research, 2023.
  89. Open-world semi-supervised learning. In International Conference on Learning Representations, 2022.
  90. Dong-Hyun Lee et al. Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks. In Workshop on challenges in representation learning, ICML, volume 3, page 896. Atlanta, 2013.
  91. Realistic evaluation of deep semi-supervised learning algorithms. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors, Advances in Neural Information Processing Systems, volume 31. Curran Associates, Inc., 2018.
  92. Learning to cluster in order to transfer across domains and tasks. In International Conference on Learning Representations, 2018.
  93. Regularization with stochastic transformations and perturbations for deep semi-supervised learning. Advances in neural information processing systems, 29, 2016.
  94. Novel visual category discovery with dual ranking statistics and mutual knowledge distillation. Advances in Neural Information Processing Systems, 34:22982–22994, 2021.
  95. Neighborhood contrastive learning for novel class discovery. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10867–10875, June 2021.
  96. Openmix: Reviving known knowledge for discovering novel visual categories in an open world. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9462–9470, 2021.
  97. A unified objective for novel class discovery. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 9284–9292, October 2021.
  98. Unsupervised learning of visual features by contrasting cluster assignments. Advances in neural information processing systems, 33:9912–9924, 2020.
  99. Marco Cuturi. Sinkhorn distances: Lightspeed computation of optimal transport. Advances in neural information processing systems, 26, 2013.
  100. Modeling inter-class and intra-class constraints in novel class discovery. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 3449–3458, June 2023.
  101. James MacQueen et al. Some methods for classification and analysis of multivariate observations. In Proceedings of the fifth Berkeley symposium on mathematical statistics and probability, volume 1, pages 281–297. Oakland, CA, USA, 1967.
  102. K-means++ the advantages of careful seeding. In Proceedings of the eighteenth annual ACM-SIAM symposium on Discrete algorithms, pages 1027–1035, 2007.
  103. Generalized category discovery. In IEEE Conference on Computer Vision and Pattern Recognition, 2022.
  104. Supervised contrastive learning. Advances in neural information processing systems, 33:18661–18673, 2020.
  105. Xcon: Learning with experts for fine-grained category discovery. In British Machine Vision Conference (BMVC), 2022.
  106. Dynamic conceptional contrastive learning for generalized category discovery. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7579–7588, 2023.
  107. Learning semi-supervised gaussian mixture models for generalized category discovery. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 16623–16633, October 2023.
  108. Promptcal: Contrastive affinity learning via auxiliary prompts for generalized novel category discovery. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3479–3488, 2023.
  109. Visual prompt tuning. In European Conference on Computer Vision, pages 709–727. Springer, 2022.
  110. Parametric classification for generalized category discovery: A baseline study. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 16590–16600, October 2023.
  111. Emerging properties in self-supervised vision transformers. In Proceedings of the IEEE/CVF international conference on computer vision, pages 9650–9660, 2021.
  112. An extensive comparative study of cluster validity indices. Pattern recognition, 46(1):243–256, 2013.
  113. The map equation. The European Physical Journal Special Topics, 178(1):13–23, 2009.
  114. Prototypical contrastive learning of unsupervised representations. In International Conference on Learning Representations, 2021.
  115. Overcoming catastrophic forgetting in neural networks. Proceedings of the National Academy of Sciences, 114:3521 – 3526, 2017.
  116. Learning without forgetting. IEEE Transactions on Pattern Analysis and Machine Intelligence, 40:2935–2947, 2018.
  117. Memory aware synapses: Learning what (not) to forget. In Proceedings of the European Conference on Computer Vision, pages 139–154, 2018.
  118. Rahaf Aljundi. Continual learning in neural networks. arXiv preprint arXiv:1910.02718, 2019.
  119. Federated learning: Challenges, methods, and future directions. IEEE signal processing magazine, 37(3):50–60, 2020.
  120. Prototype augmentation and self-supervision for incremental learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5871–5880, 2021.
  121. An empirical investigation of catastrophic forgetting in gradient-based neural networks. arXiv preprint arXiv:1312.6211, 2013.
  122. Metaplasticity: the plasticity of synaptic plasticity. Trends in neurosciences, 19(4):126–130, 1996.
  123. Brain-inspired replay for continual learning with artificial neural networks. Nature communications, 11(1):1–14, 2020.
  124. Continual learning of context-dependent processing in neural networks. Nature Machine Intelligence, 1(8):364–372, 2019.
  125. Orthogonal gradient descent for continual learning. In International Conference on Artificial Intelligence and Statistics, pages 3762–3773, 2020.
  126. Training networks in null space of feature covariance for continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 184–193, 2021.
  127. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531, 2015.
  128. Encoder based lifelong learning. In Proceedings of the IEEE/CVF Conference on Computer vision, pages 1320–1328, 2017.
  129. Learning without memorizing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5138–5146, 2019.
  130. Paying more attention to attention: Improving the performance of convolutional neural networks via attention transfer. In International Conference on Learning Representations, 2016.
  131. Class-incremental learning via deep model consolidation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 1131–1140, 2020.
  132. More classifiers, less forgetting: A generic multi-classifier paradigm for incremental learning. In Proceedings of the European Conference on Computer Vision, pages 699–716, 2020.
  133. Overcoming catastrophic forgetting with unlabeled data in the wild. In Proceedings of the IEEE/CVF Conference on Computer vision, pages 312–321, 2019.
  134. Calibration for non-exemplar based class-incremental learning. In IEEE International Conference on Multimedia and Expo, pages 1–6. IEEE, 2021.
  135. icarl: Incremental classifier and representation learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5533–5542, 2017.
  136. Towards prior gap and representation gap for long-tailed recognition. Pattern Recognition, 133:109012, 2023.
  137. Maintaining discrimination and fairness in class incremental learning. In Proceedings of the IEEE/CVF Conference on Computer vision, pages 13205–13214, 2020.
  138. Imbalanced-learn: A python toolbox to tackle the curse of imbalanced datasets in machine learning. The Journal of Machine Learning Research, 18(1):559–563, 2017.
  139. Gdumb: A simple approach that questions our progress in continual learning. In European Conference on Computer Vision, 2020.
  140. End-to-end incremental learning. In European Conference on Computer Vision, pages 233–248, 2018.
  141. Podnet: Pooled outputs distillation for small-tasks incremental learning. In European Conference on Computer Vision, pages 86–102. Springer, 2020.
  142. Learning a unified classifier incrementally via rebalancing. In Proceedings of the IEEE/CVF Conference on Computer vision, pages 831–839, 2019.
  143. Ss-il: Separated softmax for incremental learning. In Proceedings of the IEEE/CVF International conference on computer vision, pages 844–853, 2021.
  144. Large scale incremental learning. In Proceedings of the IEEE/CVF International conference on computer vision, pages 374–382, 2019.
  145. Scail: Classifier weights scaling for class incremental learning. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, pages 1266–1275, 2020.
  146. Il2m: Class incremental learning with dual memory. In Proceedings of the IEEE/CVF international conference on computer vision, pages 583–592, 2019.
  147. Topology-preserving class-incremental learning. In European Conference on Computer Vision, pages 254–270, 2020.
  148. Model behavior preserving for class-incremental learning. IEEE Transactions on Neural Networks and Learning Systems, 2022.
  149. Co2l: Contrastive continual learning. In Proceedings of the IEEE/CVF International conference on computer vision, pages 9516–9525, 2021.
  150. On learning the geodesic path for incremental learning. In Proceedings of the IEEE/CVF International conference on computer vision, pages 1591–1600, 2021.
  151. Distilling causal effect of data in class-incremental learning. In Proceedings of the IEEE/CVF International conference on computer vision, pages 3957–3966, 2021.
  152. Packnet: Adding multiple tasks to a single network by iterative pruning. In Proceedings of the IEEE/CVF international conference on computer vision, pages 7765–7773, 2018.
  153. Overcoming catastrophic forgetting with hard attention to the task. In International conference on machine learning, pages 4548–4557. PMLR, 2018.
  154. Adaptive aggregation networks for class-incremental learning. In Proceedings of the IEEE/CVF international conference on computer vision, pages 2544–2553, 2021.
  155. Dualnet: Continual learning, fast and slow. Advances in Neural Information Processing Systems, 34:16131–16144, 2021.
  156. Der: Dynamically expandable representation for class incremental learning. In Proceedings of the IEEE/CVF international conference on computer vision, pages 3014–3023, 2021.
  157. Foster: Feature boosting and compression for class-incremental learning. In European conference on computer vision, pages 398–414. Springer, 2022.
  158. Sabine Trepte. The social media privacy model: Privacy and communication in the light of social media affordances. Communication Theory, 31(4):549–570, 2021.
  159. Engrams and circuits crucial for systems consolidation of a memory. Science, 356(6333):73–78, 2017.
  160. Biological underpinnings for lifelong learning machines. Nature Machine Intelligence, 4(3):196–210, 2022.
  161. Continual learning with deep generative replay. Advances in neural information processing systems, 30, 2017.
  162. Memory replay gans: Learning to generate new categories without forgetting. Advances in Neural Information Processing Systems, 31, 2018.
  163. Dreaming to distill: Data-free knowledge transfer via deepinversion. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8715–8724, 2020.
  164. Always be dreaming: A new approach for data-free class-incremental learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 9374–9384, 2021.
  165. Latent replay for real-time continual learning. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 10203–10209. IEEE, 2020.
  166. Remind your neural network to prevent catastrophic forgetting. In European Conference on Computer Vision, pages 466–483. Springer, 2020.
  167. Memory-efficient incremental learning through feature adaptation. In European Conference on Computer Vision, pages 699–715. Springer, 2020.
  168. Incremental learning using conditional adversarial networks. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 6619–6628, 2019.
  169. Generative feature replay with orthogonal weight modification for continual learning. In 2021 International Joint Conference on Neural Networks (IJCNN), pages 1–8. IEEE, 2021.
  170. Generative feature replay for class-incremental learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pages 226–227, 2020.
  171. Self-sustaining representation expansion for non-exemplar class-incremental learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9296–9305, 2022.
  172. Semantic drift compensation for class-incremental learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6982–6991, 2020.
  173. Bring evanescent representations to life in lifelong class incremental learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16732–16741, 2022.
  174. Class-incremental learning via dual augmentation. Advances in Neural Information Processing Systems, 34:14306–14318, 2021.
  175. Federated class-incremental learning with prototype guided transformer. arXiv preprint arXiv:2401.02094, 2024.
  176. Class incremental learning with self-supervised pre-training and prototype learning. arXiv preprint arXiv:2308.02346, 2023.
  177. Learning to prompt for continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 139–149, 2022.
  178. Dualprompt: Complementary prompting for rehearsal-free continual learning. In European Conference on Computer Vision, pages 631–648. Springer, 2022.
  179. Coda-prompt: Continual decomposed attention-based prompting for rehearsal-free continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023.
  180. Pivot: Prompting for video continual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24214–24223, 2023.
  181. Hierarchical decomposition of prompt-based continual learning: Rethinking obscured sub-optimality. In Thirty-seventh Conference on Neural Information Processing Systems, 2023.
  182. Learnability and algorithm for continual learning. In International Conference on Machine Learning, 2023.
  183. A call to reflect on evaluation practices for failure detection in image classification. In International Conference on Learning Representations, 2022.
  184. Rethinking confidence calibration for failure prediction. In European Conference on Computer Vision, pages 518–536. Springer, 2022.
  185. Unified classification and rejection: A one-versus-all framework. arXiv preprint arXiv:2311.13355, 2023.
  186. Novel class discovery without forgetting. In European Conference on Computer Vision, pages 570–586. Springer, 2022.
  187. Class-incremental novel class discovery. In European Conference on Computer Vision, pages 317–333. Springer, 2022.
  188. Incremental generalized category discovery. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 19137–19147, October 2023.
  189. On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258, 2021.
  190. Learning transferable visual models from natural language supervision. In International Conference on Machine Learning, pages 8748–8763, 2021.
  191. Segment anything. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 4015–4026, October 2023.
  192. In chatgpt we trust? measuring and characterizing the reliability of chatgpt. arXiv preprint arXiv:2304.08979, 2023.
  193. Accuracy of segment-anything model (sam) in medical image segmentation tasks. arXiv preprint arXiv:2304.09324, 2023.
  194. Preventing zero-shot transfer degradation in continual learning of vision-language models. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 19125–19136, October 2023.
  195. Graph posterior network: Bayesian predictive uncertainty for node classification. In Advances in Neural Information Processing Systems, volume 34, pages 18033–18048, 2021.
  196. Energy-based out-of-distribution detection for graph neural networks. In International Conference on Learning Representations, 2023.
  197. Overcoming catastrophic forgetting in graph neural networks. In Proceedings of the AAAI conference on artificial intelligence, volume 35, pages 8653–8661, 2021.
  198. Cglb: Benchmark tasks for continual graph learning. Advances in Neural Information Processing Systems, 35:13006–13021, 2022.
  199. Detecting everything in the open world: Towards universal object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11433–11443, 2023.
  200. Incremental object detection via meta-learning. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(12):9209–9216, 2021.
  201. Mining unseen classes via regional objectness: A simple baseline for incremental segmentation. Advances in Neural Information Processing Systems, 35:24340–24353, 2022.
  202. Endpoints weight fusion for class incremental semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7204–7213, 2023.
  203. Towards open world object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5830–5840, 2021.
  204. Incorporating neuro-inspired adaptability for continual learning in artificial intelligence. Nature Machine Intelligence, pages 1–13, 2023.
  205. Incremental adaptive learning vector quantization for character recognition with continuous style adaptation. Cognitive Computation, 10(2):334–346, 2018.
  206. Triple-memory networks: A brain-inspired method for continual learning. IEEE Transactions on Neural Networks and Learning Systems, 33(5):1925–1934, 2021.
  207. Privacy and security of big data in ai systems: A research and standards perspective. In 2019 IEEE International Conference on Big Data (Big Data), pages 5737–5743. IEEE, 2019.
  208. Towards fairness in visual recognition: Effective strategies for bias mitigation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8919–8928, 2020.
  209. Machine unlearning: A survey. ACM Computing Surveys, 2023.
  210. Forget-me-not: Learning to forget in text-to-image diffusion models. arXiv preprint arXiv:2211.08332, 2023.
Citations (12)
View on Semantic Scholar

Summary

  • The paper introduces a novel approach enabling models to detect unknown classes and adapt continuously in dynamic environments.
  • It presents methodologies like OOD detection, anomaly recognition, and combined supervised-unsupervised clustering to overcome closed-world limitations.
  • Future directions include unified frameworks and brain-inspired strategies aimed at enhancing AI adaptability and reducing catastrophic forgetting.

Open-world Machine Learning: A Comprehensive Overview

The paper "Open-world Machine Learning: A Review and New Outlooks" presents a meticulous analysis and synthesis of an emerging paradigm in machine learning that challenges the traditional closed-world assumption. This paper explores the imperative advancements needed to equip models with the capability to recognize, learn, and adapt to dynamic and unknown environments, termed as open-world machine learning (OWL). The authors provide a detailed exposition on key components of OWL—unknown rejection, novel class discovery, and class-incremental learning—highlighting the challenges, methodologies, and prospective research directions in this domain.

This document comprises a comprehensive narrative on how existing models, primarily trained under closed-world assumptions, are limited in their potential to engage with unpredictable scenarios typical of real-world applications. The authors propose that an intelligent system, much like a biological organism, should continuously evolve by identifying unknowns, discovering new class paradigms, and adapting incrementally without prior constraints.

The paper is structured into three primary segments: unknown rejection, novel class discovery, and class-incremental learning, each presenting distinct challenges and existing solutions. Unknown rejection deals with the ability of a model to recognize instances that diverge from known classes. Here, the authors detail methods like out-of-distribution (OOD) detection, anomaly detection, and open-set recognition (OSR) which are pivotal for ensuring AI systems abstain from making fallacious predictions.

For novel class discovery, the authors pivot to methods that facilitate the identification and categorization of unknown classes within new data, a step necessary before incremental learning becomes feasible. They explore various methodologies that integrate supervised and unsupervised learning paradigms to enable robust clustering and labeling of novel classes.

Class-incremental learning, the final discussed aspect, is concerned with the sustainable evolution of a model's ability to integrate new knowledge without catastrophic forgetting. This section expands upon exemplar replay, regularization techniques, and prompt-based methods that contribute to continual learning processes essential for maintaining long-term model efficacy.

A key contribution of this paper is its forward-looking perspective, offering viable future directions for OWL. These include the formulation of unified OWL frameworks, integration within foundation models, and extension to structured data and complex applications beyond classification. The authors also emphasize the prospective value of brain-inspired models and open-world unlearning, domains poised to bridge current gaps between human-like adaptability and artificial intelligence.

In exploring these facets, the authors challenge current AI paradigms, emphasizing the necessity of systems capable of real-time learning and adaptation in unpredictable environments. They advocate for a paradigm shift where AI systems emulate not merely the problem-solving aspects of intelligence but its inherent dynamism and adaptability, akin to biological systems.

Overall, the paper serves as a pivotal resource for researchers, guiding the development of sophisticated AI systems that can thrive in open-world conditions. The insights and innovations discussed lay a foundational understanding for future explorations and technological advancements in artificial general intelligence. The methodologies and potential implications highlighted in this paper are crucial for steering AI research towards achieving more aligned, safe, and continuously learning systems capable of real-world interaction.

File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets