Trouble on the Horizon: Forecasting the Derailment of Online Conversations as they Develop (D19-1)
Copied to clipboard
| Challenge: | Recent efforts focused on detecting antisocial behavior after the fact . a forecasting model needs to capture the flow of the conversation, not individual comments . real conversations have an unknown horizon; therefore a practical forecasting system needs to assess the risk . |
| Approach: | They propose a conversational forecasting model that learns conversational dynamics and exploits it to predict derailment as the conversation develops. |
| Outcome: | The proposed model outperforms state-of-the-art models at forecasting derailment . it learns an unsupervised representation of conversational dynamics and exploits it to predict future derailments . |
Similar Papers
A Theoretically Grounded Approach to Summarizing Conversation Dynamics for Forecasting the Derailment of Online Conversations (2026.acl-long)
Copied to clipboard
| Challenge: | Recent work on conversation derailment prediction relies on linguistic features rooted in linguistic and social theories. |
| Approach: | They propose a system that predicts from the start of a conversation whether it will derail into toxic exchanges. |
| Outcome: | The proposed system achieves 10% performance increase over baseline and 6.47% increase on benchmark dataset. |
Dynamic Forecasting of Conversation Derailment (2021.emnlp-main)
Copied to clipboard
| Challenge: | a pretrained language encoder can predict derailment in online conversations . this is a useful task for detecting and preventing abusive language . |
| Approach: | They extend a task to predict derailment in online conversations by using a pretrained language encoder. |
| Outcome: | The proposed task outperforms previous approaches in terms of performance and quality. |
Can LLMs Be Efficient Predictors of Conversational Derailment? (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Conversational derailment is a common issue on online platforms due to toxic or inappropriate remarks. |
| Approach: | They prompt pre-trained large language models to predict conversational derailment without fine-tuning . they compare chain-of-thought reasoning and few-shot exemplars to predict derailments . |
| Outcome: | The proposed model predicts conversational derailment without task-specific fine-tuning without fine-cuning. |
Wait! There’s a Way Out: A Decision Mechanism for Forecasting Conversational Derailment (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches make decision to "trigger" based on the estimated likelihood of derailment given the preceding utterances, implicitly assuming that the conversation’s future trajectory is fixed. |
| Approach: | They propose a method for decoupling the decision to trigger from derailment likelihood estimation. |
| Outcome: | The proposed method is inspired by the first human baseline on this task, which shows that humans achieve dramatically lower false positive rates by selectively deferring their decision to trigger when they anticipate that tension is likely to subside. |
How did we get here? Summarizing conversation dynamics (2024.naacl-long)
Copied to clipboard
Yilun Hua, Nicholas Chernogor, Yuzhe Gu, Seoyeon Jeong, Miranda Luo, Cristian Danescu-Niculescu-Mizil
| Challenge: | Throughout a conversation, the way participants interact with each other is in constant flux. |
| Approach: | They propose to summarize conversations by constructing human-written summaries and exploring automated baselines. |
| Outcome: | The summarizing tools help both humans and automated systems forecast toxic behavior in conversations. |
Conversations Gone Awry: Detecting Early Signs of Conversational Failure (P18-1)
Copied to clipboard
Justine Zhang, Jonathan Chang, Cristian Danescu-Niculescu-Mizil, Lucas Dixon, Yiqing Hua, Dario Taraborelli, Nithum Thain
| Challenge: | Prior work focused on characterizing and detecting content exhibiting antisocial online behavior. |
| Approach: | They propose a task of predicting from the very start of a conversation whether it will get out of hand. |
| Outcome: | The proposed framework can detect early warning signs of antisocial behavior in online conversations. |
Multimodal Conversation Modelling for Topic Derailment Detection (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing work on analysing textual dialogues that derailed into toxic content ignores visual information, such as images and videos. |
| Approach: | They propose a new multimodal conversational architecture that utilises visual and conversational contexts to classify comments for derailment. |
| Outcome: | The proposed approach outperforms existing methods and is more robust to textual noise. |
Deal, or no deal (or who knows)? Forecasting Uncertainty in Conversations using Large Language Models (2024.findings-acl)
Copied to clipboard
| Challenge: | Effective interlocutors account for the uncertain goals, beliefs, and emotions of others. |
| Approach: | They propose to calibrate language models to better represent outcome uncertainty . they propose to use two methods to calibrated small open-source models . |
| Outcome: | The proposed fine-tuning strategies can calibrate smaller open-source models to beat pre-trained models 10x their size. |
Dynamic Online Conversation Recommendation (2020.acl-main)
Copied to clipboard
| Challenge: | Existing models that assume static user interests are unable to capture the temporal aspects of user interactions and interest changes over time. |
| Approach: | They propose a neural architecture to exploit changes of user interactions and interests over time to predict which discussions they are likely to enter. |
| Outcome: | The proposed model outperforms state-of-the-art models that assume static user interests and handle future conversations that are unseen during training time. |
Response-conditioned Turn-taking Prediction (2023.findings-acl)
Copied to clipboard
| Challenge: | Traditionally, turn-taking is done using a simple silence threshold, but more modern approaches use cues known to be important in human-human turn-shifts. |
| Approach: | They propose a turn-taking and response-ranking model that conditions the end-of-turn prediction on conversation history and what the next speaker wants to say. |
| Outcome: | The proposed model outperforms the baseline model in a variety of metrics. |