IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models
- Submitting institution
-
Birkbeck College
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 12
- Type
- E - Conference contribution
- DOI
-
10.1145/3077136.3080786
- Title of conference / published proceedings
- Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval
- First page
- 515
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- August
- Year of publication
- 2017
- URL
-
https://dl.acm.org/doi/10.1145/3077136.3080786
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
7
- Research group(s)
-
2 - Experimental Data Science
- Citation count
- 118
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper is the first successful attempt at unifying, in a principled manner, the two distinct schools of information retrieval: generative and discriminative retrieval models. It represents a novel application of GAN, which has been one of the most popular machine learning techniques in the last few years. In addition, it is one of the earliest methods in information retrieval that utilises policy gradient reinforcement learning. It won the prestigious Best Paper Award Honourable Mention at SIGIR-2017 (https://sigir.org/awards/best-paper-awards/). The source code and experimental datasets have been released (https://github.com/geek-ai/irgan).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -