Timezone: »
We present a novel multi-source uncertainty prediction approach that enables deep learning (DL) models to be actively trained with much less labeled data. By leveraging the second-order uncertainty representation provided by subjective logic (SL), we conduct evidence-based theoretical analysis and formally decompose the predicted entropy over multiple classes into two distinct sources of uncertainty: vacuity and dissonance, caused by lack of evidence and conflict of strong evidence, respectively. The evidence based entropy decomposition provides deeper insights on the nature of uncertainty, which can help effectively explore a large and high-dimensional unlabeled data space. We develop a novel loss function that augments DL based evidence prediction with uncertainty anchor sample identification. The accurately estimated multiple sources of uncertainty are systematically integrated and dynamically balanced using a data sampling function for label-efficient active deep learning (ADL). Experiments conducted over both synthetic and real data and comparison with competitive AL methods demonstrate the effectiveness of the proposed ADL model.
Author Information
Weishi Shi (Rochester Institute of Technology)
Xujiang Zhao (The University of Texas at Dallas)
Feng Chen (UT Dallas)
Qi Yu (Rochester Institute of Technology)
More from the Same Authors
-
2021 : A Nested Bi-level Optimization Framework for Robust Few Shot Learning »
Krishnateja Killamsetty · Changbin Li · Chen Zhao · Rishabh Iyer · Feng Chen -
2023 Poster: Actively Testing Your Model While It Learns: Realizing Label-Efficient Learning in Practice »
Dayou Yu · Weishi Shi · Qi Yu -
2023 Poster: Distributionally Robust Ensemble of Lottery Tickets Towards Calibrated Sparse Network Training »
Hitesh Sapkota · Dingrong Wang · Zhiqiang Tao · Qi Yu -
2023 Poster: Improvements on Uncertainty Quantification for Node Classification via Distance Based Regularization »
Russell Hart · Linlin Yu · Yifei Lou · Feng Chen -
2021 Poster: A Gaussian Process-Bayesian Bernoulli Mixture Model for Multi-Label Active Learning »
Weishi Shi · Dayou Yu · Qi Yu -
2021 Poster: RETRIEVE: Coreset Selection for Efficient and Robust Semi-Supervised Learning »
Krishnateja Killamsetty · Xujiang Zhao · Feng Chen · Rishabh Iyer -
2020 Poster: Uncertainty Aware Semi-Supervised Learning on Graph Data »
Xujiang Zhao · Feng Chen · Shu Hu · Jin-Hee Cho -
2020 Spotlight: Uncertainty Aware Semi-Supervised Learning on Graph Data »
Xujiang Zhao · Feng Chen · Shu Hu · Jin-Hee Cho -
2020 Poster: Dynamic Fusion of Eye Movement Data and Verbal Narrations in Knowledge-rich Domains »
Ervine Zheng · Qi Yu · Rui Li · Pengcheng Shi · Anne Haake -
2019 Poster: Integrating Bayesian and Discriminative Sparse Kernel Machines for Multi-class Active Learning »
Weishi Shi · Qi Yu