Argument Mining

Argumentative Fallacies

The detection and classification of argumentative fallacies. Fallacy constitute an important aspect of argumentation because they require substantial reasoning capabilities to be spotted, thus, representing a valuable challenge for machine learning models.

Multimodality

We have worked in evaluating the combination of audio and text modalities. Does audio modality provide any benefit in addressing argument mining tasks?

Dialogues

We have developed a benchmark for argumentative dialogues on scientific papers. Nonetheless, nowadays, the focus is more on developing commercial tools rather than on developing benchmarks.

Interpretability

The development of interpretable models for argumentation. A current focus is on extracting local explanations from texts to acquire insights about data. For instance, which kind of patterns are associated with certain argumentative components?

Reasoning in LLMs

LLMs and reasoning are hot topics that are currently vastly being investigated. Though, there are still few attempts that aim to use argumentation as a resource for assessing reasoning in LLMs.