Argument Mining from Speech: Detecting Claims in Political Debates
Authors: Marco Lippi, Paolo Torroni
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | To explore this hypothesis, we develop a machine learning classifier and train it on an original dataset based on the 2015 UK political elections debate. ... The final sections are devoted to experimental evaluation and discussion of results and future directions. ... Experiments Methodology We run experiments on our dataset by employing several different settings. ... Table 1 summarizes the main results obtained by all the considered models. |
| Researcher Affiliation | Academia | Marco Lippi DISI, University of Bologna, Italy marco.lippi3@unibo.it Paolo Torroni DISI, University of Bologna, Italy paolo.torroni@unibo.it |
| Pseudocode | No | The paper does not contain any pseudocode or clearly labeled algorithm blocks. It provides a system pipeline diagram in Figure 1. |
| Open Source Code | No | The paper does not provide or explicitly mention making its source code available. It only states: 'Let the 7-party debate dataset we release be an incentive for others to take on the challenge and start on this promising new line of research.' |
| Open Datasets | Yes | We based our study on an original corpus constructed from material available online from the 7-party leader s debate of April 2, 2015 which preceded the last UK general elections. ... The entire corpus with annotated claims can be downloaded from http://argumentationmining.disi.unibo.it |
| Dataset Splits | Yes | We employed Support Vector Machines and exploited a 10-fold cross validation on the dataset of each candidate. We employed both linear and rbf kernel, with parameters (C for linear kernel, C and γ for rbf) chosen by an inner 5-fold cross validation on the training set. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used to run the experiments, such as CPU/GPU models or memory specifications. |
| Software Dependencies | No | The paper mentions 'Google Speech APIs' and the 'Rasta Mat library' but does not provide specific version numbers for these or any other software dependencies, which is required for reproducibility. |
| Experiment Setup | No | The paper states that parameters were 'chosen by an inner 5-fold cross validation on the training set' but does not provide the concrete hyperparameter values (e.g., specific C or gamma values) that were selected or used for the final reported results. |