Learning Placeholders for Open-Set Recognition
By Da-Wei Zhou, Han-Jia Ye, De-chuan Zhan. 2021
267 citations recorded by the source corpus.
Traditional classifiers are deployed under closed-set setting, with both training and test classes belong to the same set. However, real-world applications probably face the input of unknown categories, and the model will recognize them as known ones. Under such circumstances, open-set recognition is proposed to maintain classification performance on known classes and reject unknowns. The closed-set models make overconfident predictions over familiar known class instances, so that calibration and thresholding across categories become essential issues when extending to an open-set environment. To this end, we proposed to learn PlaceholdeRs for Open-SEt Recognition (Proser), which prepares for the unknown classes by allocating placeholders for both data and classifier. In detail, learning data placeholders tries to anticipate open-set class data, thus transforms closed-set training into open-set training. Besides, to learn the invariant information between target and non-target classes, we reserve classifier placeholders as the class-specific boundary between known and unknown. The proposed Proser efficiently generates novel class by manifold mixup, and adaptively sets the value of reserved open-set classifier during training. Experiments on various datasets validate the effectiveness of our proposed method.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Anomaly & Outlier Detection
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Han-Jia Ye
- De-chuan Zhan
- Ziwei Liu
- Da-Wei Zhou
- Qiwen Wang
- Vineeth Nallure Balasubramanian
- Bernt Schiele
- Jun Zhu
In Computer Vision and Pattern Recognition
- Han-Jia Ye
- De-chuan Zhan
- Ziwei Liu
- Da-Wei Zhou
- Luc Van Gool
- Bernt Schiele
- Dacheng Tao
- Ming-Hsuan Yang
In Anomaly & Outlier Detection
- De-chuan Zhan
- Da-Wei Zhou
- Han-Jia Ye
- K. J. Joseph
- Yue Zhao
- Vineeth Nallure Balasubramanian
- Ziwei Liu
- Kai Han
Citations in the corpus
- Language-driven Semantic Segmentation (2022, background): “A variety of solutions have been proposed (Zhang et al., 2020b; Liu et al., 2020a; Perera et al., 2020; Zhou et al., 2021) for open-set recognition (Scheirer et al., 2012; Geng et al., 2020).”
- Novel Class Discovery without Forgetting (2022, background): “An instance outside this set of classes may be wrongly classified into one of the known classes often with high confidence [48,62,46,18].”
- FOSTER: Feature Boosting and Compression for Class-Incremental Learning (2022, background): “The real world is constantly changing, with new concepts and categories continuously springing up [14,35,46,48].”
- Few-Shot Class-Incremental Learning by Sampling Multi-Phase Tasks (2022, background): “However, real-world applications often face the stream format data [7] and novel classes [8], [9], [10], [11], e.g., incoming hot topics in social media and new types of products in e-commerce.”
- Forward Compatible Few-Shot Class-Incremental Learning (2022, background): “For example, [14] seeks to detect new classes by learning placeholders, [5] utilizes the embedding with large margin between classes, [7] encourages class-wise orthogonality for more compact embedding.”
- Learning to Classify With Incremental New Class (2021, unknown): “An ideal system should tackle such real-world scenarios, which requires it to automatically detect novel classes and incrementally self-update with detected instances.”
- Deep Class-Incremental Learning: A Survey (2023, background): “If the test dataset contains instances from unknown classes, open-set recognition [208] and novel class discovery [216], [217] can equip”
- A Model or 603 Exemplars: Towards Memory-Efficient Class-Incremental Learning (2022, background): “In the open world, training data is often collected in stream format with new classes appearing [16, 15, 63, 65].”
- ADBench: Anomaly Detection Benchmark (2022, methodology): “…for AD as well; future benchmark can consider including: ( i ) OOD methods: MSP [65], energy-based EBO [104], and Mahalanobis distance-based MDS [92]; ( ii ) novelty detection methods: OCGAN [135] and Adversarial One-Class Classifier [154]; and ( iii ) OSR methods: OpenGAN [79] and PROSER [204].”
- Generalized Out-of-Distribution Detection: A Survey (2021, unknown): “Domain Adaptation/Generalization Domain Adaptation (DA) (Wang & Deng, 2018) and Domain Generalization (DG) (Zhou et al., 2021b) also follow “open-world” assumption.”
- CMG: A Class-Mixed Generation Approach to Out-of-Distribution Detection (2022, methodology): “( We follow the data processing method in [51, 58] to split known and unknown classes.”
- Class-Incremental Learning: A Survey (2023, methodology): “In DyTox, only task tokens are expanded for each new task, which requires much less memory than saving the whole backbone.”