Abstract
The study of questions in the setting of dialogical interactions in corporate communication has the purpose of understanding and capitalizing on the opinions that the questioner has with respect to the questioned company. Particularly, financial analysts have the maximal incentive to be right in their forecasts about the company’s performance, but they are also incentivized and expected to maintain a good relationship with the management– and therefore, not to be too challenging in their questions. While avoiding overt adversarialness, analysts adopt alternative strategies to seek the desired information; among which modulating the cornering quality of questions. This paper presents a way of measuring such cornering property, automatically extracting feature scores, and comparing the results with a manually annotated gold standard. Results encourage further research along this stream, particularly towards the study of replies and their degree of answerhood with respect to the cornering quality of the prompting question.
Original language | English |
---|---|
Title of host publication | Proceedings of the Eighth Financial Technology and Natural Language Processing and the 1st Agent AI for Scenario Planning |
Pages | 107-118 |
Publication status | Published - 3 Aug 2024 |
Event | Eighth Financial Technology and Natural Language Processing and the 1st Agent AI for Scenario Planning - International Convention Center Jeju, Jeju Island, Korea, Republic of Duration: 3 Aug 2024 → 9 Aug 2024 https://ijcai24.org/ |
Conference
Conference | Eighth Financial Technology and Natural Language Processing and the 1st Agent AI for Scenario Planning |
---|---|
Abbreviated title | IJCAI 2024 |
Country/Territory | Korea, Republic of |
City | Jeju Island |
Period | 3/08/24 → 9/08/24 |
Internet address |