David Powers


2022

pdf bib
Proceedings of the 20th Annual Workshop of the Australasian Language Technology Association
Pradeesh Parameswaran | Jennifer Biggs | David Powers
Proceedings of the 20th Annual Workshop of the Australasian Language Technology Association

pdf
Contrastive Visual and Language Learning for Visual Relationship Detection
Thanh Tran | Maelic Neau | Paulo Santos | David Powers
Proceedings of the 20th Annual Workshop of the Australasian Language Technology Association

Visual Relationship Detection aims to understand real-world objects’ interactions by grounding visual concepts to compositional visual relation triples, written in the form of (subject, predicate, object). Previous works have explored the use of contrastive learning to implicitly predict the predicates from the relevant image regions. However, these models often directly leverage in-distribution spatial and language co-occurrences biases during training, preventing the models from generalizing to out-of-distribution compositions. In this work, we examine whether contrastive vision and language models pre-trained on large-scale external image and text dataset can assist the detection of compositional visual relationships. To this end, we propose a semi-supervised contrastive fine-tuning approach for the visual relationship detection task. The results show that fine-tuned models that were pre-trained on larger datasets do not yield better performance when performing visual relationship detection, and larger models can yield lower performance when compared with their smaller counterparts.

2016

pdf bib
Computational Natural Language Learning: +-20years +-Data +-Features +-Multimodal +-Bioplausible
David Powers
Proceedings of the 20th SIGNLL Conference on Computational Natural Language Learning

2008

pdf bib
Proceedings of the Australasian Language Technology Association Workshop 2008
Nicola Stokes | David Powers
Proceedings of the Australasian Language Technology Association Workshop 2008

2007

pdf
An Empirical Investigation into Grammatically Constrained Contexts in Predicting Distributional Similarity
Dongqiang Yang | David Powers
Proceedings of the Australasian Language Technology Workshop 2007

2004

pdf
Adaptive Compression-based Approach for Chinese Pinyin Input
JinHu Huang | David Powers
Proceedings of the Third SIGHAN Workshop on Chinese Language Processing

2003

pdf
Chinese Word Segmentation Based on Contextual Entropy
Jin Hu Huang | David Powers
Proceedings of the 17th Pacific Asia Conference on Language, Information and Computation

1998

pdf
Evolution and Evaluation of Document Retrieval Queries
Robert Steele | David Powers
New Methods in Language Processing and Computational Natural Language Learning