Diversified dictionaries for multi-instance learning

Journal article

Qiao, Maoying, Liu, Liu, Yu, Jun, Xu, Chang and Tao, Dacheng. (2017) Diversified dictionaries for multi-instance learning. Pattern Recognition. 64, pp. 407-416. https://doi.org/10.1016/j.patcog.2016.08.026
AuthorsQiao, Maoying, Liu, Liu, Yu, Jun, Xu, Chang and Tao, Dacheng

Multiple-instance learning (MIL) has been a popular topic in the study of pattern recognition for years due to its usefulness for such tasks as drug activity prediction and image/text classification. In a typical MIL setting, a bag contains a bag-level label and more than one instance/pattern. How to bridge instance-level representations to bag-level labels is a key step to achieve satisfactory classification accuracy results. In this paper, we present a supervised learning method, diversified dictionaries MIL, to address this problem. Our approach, on the one hand, exploits bag-level label information for training class-specific dictionaries. On the other hand, it introduces a diversity regularizer into the class-specific dictionaries to avoid ambiguity between them. To the best of our knowledge, this is the first time that the diversity prior is introduced to solve the MIL problems. Experiments conducted on several benchmark (drug activity and image/text annotation) datasets show that the proposed method compares favorably to state-of-the-art methods.

Keywordsmulti-instance learning; diversified learning; dictionary learning
JournalPattern Recognition
Journal citation64, pp. 407-416
PublisherElsevier Ltd
Digital Object Identifier (DOI)https://doi.org/10.1016/j.patcog.2016.08.026
Scopus EID2-s2.0-85007162586
Research or scholarlyResearch
Page range407-416
FunderAustralian Research Council
Publisher's version
All rights reserved
File Access Level
Output statusPublished
Publication dates
Online26 Aug 2016
Publication process dates
Accepted22 Aug 2016
Deposited10 Jun 2021
ARC Funded ResearchThis output has been funded, wholly or partially, under the Australian Research Council Act 2001
Grant IDARC/FT130101457
Permalink -


Restricted files

Publisher's version

  • 0
    total views
  • 0
    total downloads
  • 0
    views this month
  • 0
    downloads this month
These values are for the period from 19th October 2020, when this repository was created.

Export as

Related outputs

Improving stochastic block models by incorporating power-law degree characteristic
Qiao, Maoying, Yu, Jun, Bian, Wei, Li, Qiang and Tao, Dacheng. (2017) Improving stochastic block models by incorporating power-law degree characteristic. Twenty-Sixth International Joint Conference on Artificial Intelligence (IJCAI-17). Melbourne, Australia 19 - 25 Aug 2017 International Joint Conferences on Artificial Intelligence Organization. pp. 2620-2626 https://doi.org/10.24963/ijcai.2017/365
Conditional graphical lasso for multi-label image classification
Li, Qiang, Qiao, Maoying, Bian, Wei and Tao, Dacheng. (2016) Conditional graphical lasso for multi-label image classification. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, Nevada, United States of America 27 - 30 Jun 2016 Computer Vision Foundation. pp. 2977-2986 https://doi.org/10.1109/CVPR.2016.325