Discovering Novelty via Transfer Learning

被引:0
作者
Islam, Shafkat [1 ]
Bhargava, Bharat K. [1 ]
机构
[1] Purdue Univ, W Lafayette, IN 47907 USA
来源
SEMANTIC INTELLIGENCE, ISIC 2022 | 2023年 / 964卷
关键词
Novelty detection; Transfer learning; Open-world AI;
D O I
10.1007/978-981-19-7126-6_6
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Modern-day intelligent systems, i.e., autonomous vehicles, smart manufacturing, etc., rely on distinct data-driven machine learning (ML) models for performing different safety-critical operations such as pedestrian detection, chemical plant control, among others. Since the performance of ML models depends on their generalizability capability, handling out-of-distribution data during the operational phase is of paramount importance for enhancing the adaptability of artificial intelligence (AI) systems. Hence, finding the root causes of novelty is critical for minimizing its impact on the performance of AI systems. However, detection of novelties is not trivial since each AI system operates in a specific environment or agent settings, whereas small changes cause the detection mechanism to adapt to different environmental constraints. For example, detecting novelties in identical intelligent navigation systems differs if the system is deployed in two different countries, though the operation of a navigation system is indistinguishable. In this paper, to reduce the dependability of novelty detectionmechanisms on environment's or agent's attributes, we propose transfer learning-based novelty detection mechanisms for inter-domain applications. In this regard, we analyze the importance of feature transformation to enhance novelty detection systems' transferability. The proposed detection mechanisms aim at augmenting AI systems with rapid responsiveness to novel surroundings, thus, making AI systems responsible and trustworthy. We conduct multiple experiments on state-of-the-art neural networks, i.e., ResNet50, Mobile-Net, and benchmark datasets, i.e., MNIST.
引用
收藏
页码:67 / 75
页数:9
相关论文
共 50 条
[41]   RockNet: Rockfall and Earthquake Detection and Association via Multitask Learning and Transfer Learning [J].
Liao, Wu-Yu ;
Lee, En-Jui ;
Wang, Chung-Ching ;
Chen, Po ;
Provost, Floriane ;
Hibert, Clement ;
Malet, Jean-Philippe ;
Chu, Chung-Ray ;
Lin, Guan-Wei .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2023, 61
[42]   Discovering Blind Spots in Reinforcement Learning [J].
Ramakrishnan, Ramya ;
Kamar, Ece ;
Dey, Debadeepta ;
Shah, Julie ;
Horvitz, Eric .
PROCEEDINGS OF THE 17TH INTERNATIONAL CONFERENCE ON AUTONOMOUS AGENTS AND MULTIAGENT SYSTEMS (AAMAS' 18), 2018, :1017-1025
[43]   Discovering Novelty in Spatio/Temporal Data Using One-Class Support Vector Machines [J].
Smets, Koen ;
Verdonk, Brigitte ;
Jordaan, Elsa M. .
IJCNN: 2009 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1- 6, 2009, :1551-+
[44]   Few-Shot Learning with Novelty Detection [J].
Bjerge, Kim ;
Bodesheim, Paul ;
Karstoft, Henrik .
DEEP LEARNING THEORY AND APPLICATIONS, PT I, DELTA 2024, 2024, 2171 :340-363
[45]   Transfer Learning of Surrogate Models via Domain Affine Transformation [J].
Pan, Shuaiqun ;
Vermetten, Diederick ;
Lopez-Ibanez, Manuel ;
Back, Thomas ;
Wang, Hao .
PROCEEDINGS OF THE 2024 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, GECCO 2024, 2024, :385-393
[46]   Meta-Learning and Novelty Detection for Machine Learning with Reject Option [J].
Drapal, Patricia ;
Silva-Filho, Telmo ;
Prudencio, Ricardo B. C. .
2024 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN 2024, 2024,
[47]   Discriminative transfer learning via local and global structure preservation [J].
Chao Wang ;
Hongya Tuo ;
Jiexin Wang ;
Lingfeng Qiao .
Signal, Image and Video Processing, 2019, 13 :753-760
[48]   On Physical-Layer Authentication via Online Transfer Learning [J].
Chen, Yi ;
Ho, Pin-Han ;
Wen, Hong ;
Chang, Shih Yu ;
Real, Shahriar .
IEEE INTERNET OF THINGS JOURNAL, 2022, 9 (02) :1374-1385
[49]   Transfer Learning for Personality Perception via Speech Emotion Recognition [J].
Li, Yuanchao ;
Bell, Peter ;
Lai, Catherine .
INTERSPEECH 2023, 2023, :5197-5201
[50]   Transfer Learning for Dynamical Systems Models via Autoencoders and GANs [J].
Damiani, Angelo ;
Lopez, Gustavo Viera ;
Manganini, Giorgio ;
Metelli, Alberto Maria ;
Restelli, Marcello .
2024 AMERICAN CONTROL CONFERENCE, ACC 2024, 2024, :8-14