Seojin Bang

TL;DR: The model ATM-TCR is presented which uses a multi-head self-attention mechanism to capture biological contextual information and improve generalization performance and a novel application of the attention map from the model is presented to improve out-of-sample performance by demonstrating on recent SARS-CoV-2 data.

TL;DR: The authors classify argumentative relations based on four logical and theory-informed mechanisms between two statements, namely (i) factual consistency, (ii) sentiment coherence, (iii) causal relation, and (iv) normative relation.

TL;DR: The authors presented a large-scale analysis of sentence attackability in online arguments and demonstrated that machine learning models can automatically detect attackable sentences in arguments, significantly better than several baselines and comparably well to laypeople.

TL;DR: This paper proposed the variational information bottleneck for interpretation (VIBI), a system-agnostic interpretable method that provides a brief but comprehensive explanation of a black-box decision system, which adopts an information theoretic principle, information bottleneck principle, as a criterion for finding such explanations.