Persuasive Influence Detection: The Role of Argument Sequencing
Authors: Christopher Hidey, Kathleen McKeown
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct experiments on Change My View , a specific sub-reddit of the Reddit social media platform, building on previous work using similar data (Tan et al. 2016; Wei, Liu, and Li 2016). |
| Researcher Affiliation | Academia | Christopher Hidey Department of Computer Science Columbia University New York, NY 10027 chidey@cs.columbia.edu Kathleen Mc Keown Department of Computer Science Columbia University New York, NY 10027 kathy@cs.columbia.edu |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Code and data for our experiments is available to the research community.1 1https://github.com/chridey/cmv |
| Open Datasets | Yes | We use a dataset derived from the Change My View subreddit... Tan et al. (2016) collected threads... segmented this data into submissions before and after 2015/05/08. This process resulted in 18,363 and 2,263 discussion trees, respectively, for train and test. The resulting training set has 19516 examples (14849 negative and 4667 positive). The test set contains 2465 examples (1836 negative and 629 positive). ... Code and data for our experiments is available to the research community.1 1https://github.com/chridey/cmv |
| Dataset Splits | No | The paper mentions, 'We stop training after 30 epochs and perform early stopping on a validation set,' but it does not provide specific numerical sizes or percentages for how the validation set was split from the training data. |
| 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 components and libraries like 'GloVe vectors', 'Penn Discourse Tree Bank (PDTB) relations', 'Frame Net', 'LSTM', 'MLP', and 'ReLU', but it does not specify any version numbers for these dependencies. |
| Experiment Setup | Yes | We use binary cross-entropy as the loss function and stochastic gradient descent with a mini-batch size of 100 and Nesterov momentum with a coefficient of 0.9. Word embeddings are initialized with pre-trained 300-dimensional GloVe vectors. ... We stop training after 30 epochs and perform early stopping on a validation set. ... For the recurrent and hidden dimensions, we tested values of 50, 100, 200, and 300. For dropout (Srivastava et al. 2014) and word dropout (Iyyer et al. 2015), we used values of 0.25, 0.5, and 0.75 and determined whether to use 1 or 2 hidden layers. We use ReLU as the non-linearity in Equations 3 and 8. We evaluated the number of iterations for the memory networks and found that performance increases up to 3 iterations and begins decreasing after 3. We limit the maximum length of each sentence to 32 words and the maximum length of a post to 40 sentences. Words occurring fewer than 5 times in the training set (including the original post, title, and response) were removed. |