Please use this identifier to cite or link to this item:
https://hdl.handle.net/2440/136312
Citations | ||
Scopus | Web of Science® | Altmetric |
---|---|---|
?
|
?
|
Type: | Conference paper |
Title: | Unshuffling Data for Improved Generalization in Visual Question Answering |
Author: | Teney, D. Abbasnejad, E. van den Hengel, A. |
Citation: | Proceedings / IEEE International Conference on Computer Vision. IEEE International Conference on Computer Vision, 2021, vol.abs/2002.11894, pp.1397-1407 |
Publisher: | IEEE |
Publisher Place: | Los Alamitos, CA |
Issue Date: | 2021 |
Series/Report no.: | IEEE International Conference on Computer Vision |
ISBN: | 9781665428125 |
ISSN: | 1550-5499 |
Conference Name: | IEEE/CVF International Conference on Computer Vision (ICCV) (11 Oct 2021 - 17 Oct 2021 : Virtual Online) |
Statement of Responsibility: | Damien Teney, Ehsan Abbasnejad, Anton van den Hengel |
Abstract: | Generalization beyond the training distribution is a core challenge in machine learning. The common practice of mixing and shuffling examples when training neural networks may not be optimal in this regard. We show that partitioning the data into well-chosen, non-i.i.d. subsets treated as multiple training environments can guide the learning of models with better out-of-distribution generalization. We describe a training procedure to capture the patterns that are stable across environments while discarding spurious ones. The method makes a step beyond correlation-based learning: the choice of the partitioning allows injecting information about the task that cannot be otherwise recovered from the joint distribution of the training data. We demonstrate multiple use cases with the task of visual question answering, which is notorious for dataset biases. We obtain significant improvements on VQA-CP, using environments built from prior knowledge, existing meta data, or unsupervised clustering. We also get improvements on GQA using annotations of “equivalent questions”, and on multidataset training (VQA v2 / Visual Genome) by treating them as distinct environments. |
Rights: | ©2021 IEEE |
DOI: | 10.1109/ICCV48922.2021.00145 |
Published version: | https://ieeexplore.ieee.org/xpl/conhome/9709627/proceeding |
Appears in Collections: | Australian Institute for Machine Learning publications |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
hdl_136312.pdf | Submitted version | 2.15 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.