Unsupervised Detection of Sub-Events in Large Scale Disasters

Authors: Chidubem Arachie, Manas Gaur, Sam Anzaroot, William Groves, Ke Zhang, Alejandro Jaimes354-361

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Through quantitative experiments on two large crisis data sets (Hurricane Harvey and the 2015 Nepal Earthquake), we demonstrate the effectiveness of our approach over the state-of-the-art. Our qualitative evaluation shows better performance compared to our baseline.
Researcher Affiliation Collaboration 1Department of Computer Science, Virginia Tech, Blacksburg, VA, USA achid17@vt.edu 2Artificial Intelligence Institute, University of South Carolina, Columbia, SC, USA mgaur@email.sc.edu 3Dataminr Inc., NY, USA {sanzaroot, wgroves, kzhang, ajaimes}@dataminr.com
Pseudocode No The paper describes the method using prose and a diagram (Figure 1), but does not include structured pseudocode or an algorithm block.
Open Source Code No The paper does not explicitly state that the source code for their methodology is released, nor does it provide a direct link to such a repository. The link provided in footnote 4 (https://github.com/Doc Now/hydrator) is for a tool used to process data, not their own implementation.
Open Datasets Yes We use publicly available tweets related to both crises from Crisis NLP.3 The resource provides both unlabeled tweet IDs (concerning privacy constraints) and a small labeled tweet corpus for both Hurricane Harvey and Nepal Earthquake. Footnote 3: https://crisisnlp.qcri.org/
Dataset Splits Yes We used 795, 461 distinct unlabeled tweets from the hydrated 4.6 million tweets together with 4, 000 (3, 027 informative and 973 un-informative) labeled tweets to train the methods... We used 635, 150 distinct unlabeled tweets from the hydrated tweets together with 3, 479 (1, 636 informative and 1, 843 un-informative) labeled tweets to train the methods.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU/GPU models, memory) used for running its experiments.
Software Dependencies No The paper mentions software like 'spaCy dependency parser', 'Gensim phrase model', and 'Fast Text' but does not provide specific version numbers for these dependencies.
Experiment Setup No The paper describes the overall method, including the choice of spectral clustering, but does not provide specific experimental setup details such as hyperparameters (e.g., learning rate, batch size, number of epochs) or other detailed training configurations.