Conversational Thread Extraction and Topic Detection in Text‐Based Chat

Adams, Paige

doi:10.1002/9780470588222.ch6

Cited by 6 publications

(5 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The technical word feature was included because it improves our development classification score slightly, but it does not have a significant effect on overall performance. Adams (2008) attempts to add more semantic dimensions learned via Latent Dirichlet Allocation, and similarly finds no improvement.…”

Section: Classificationmentioning

confidence: 99%

“…Our results on new conversation detection suggest that a high-performance classifier for this task could improve results substantially. It is also interesting to consider, given the weakness of our technical words feature and the disappointing results using Latent Dirichlet Allocation from Adams (2008), how semantic similarity might be usefully modeled.…”

Section: Future Workmentioning

confidence: 99%

See 1 more Smart Citation

Disentangling Chat

Elsner¹,

Charniak²

2010

Computational Linguistics

View full text Add to dashboard Cite

When multiple conversations occur simultaneously, a listener must decide which conversation each utterance is part of in order to interpret and respond to it appropriately. We refer to this task as disentanglement. We present a corpus of Internet Relay Chat dialogue in which the various conversations have been manually disentangled, and evaluate annotator reliability. We propose a graph-based clustering model for disentanglement, using lexical, timing, and discourse-based features. The model's predicted disentanglements are highly correlated with manual annotations. We conclude by discussing two extensions to the model, specificity tuning and conversation start detection, both of which are promising but do not currently yield practical improvements.

show abstract

Section: Classificationmentioning

confidence: 99%

Section: Future Workmentioning

confidence: 99%

Disentangling Chat

Elsner¹,

Charniak²

2010

Computational Linguistics

View full text Add to dashboard Cite

show abstract

“…Conversation disentanglement has been a challenging task for a long time, researchers are working on this problem for decades. This task was earlier known as conversation management (Traum, Robinson, and Stephan 2004), thread detection (Shen et al 2006), and thread extraction (Adams and Martel 2010). Conversation disentanglement research started with using handcrafted features (Elsner and Charniak 2008;Mehri and Carenini 2017;Jiang et al 2018) as input to a simple statistical classifier for reply-to link prediction.…”

Section: Related Workmentioning

confidence: 99%

End-to-End Deep Reinforcement Learning for Conversation Disentanglement

Bhukar¹,

Kumar²,

Raghu³

et al. 2023

AAAI

View full text Add to dashboard Cite

Collaborative Communication platforms (e.g., Slack) support multi-party conversations which contain a large number of messages on shared channels. Multiple conversations intermingle within these messages. The task of conversation disentanglement is to cluster these intermingled messages into conversations. Existing approaches are trained using loss functions that optimize only local decisions, i.e. predicting reply-to links for each message and thereby creating clusters of conversations. In this work, we propose an end-to-end reinforcement learning (RL) approach that directly optimizes a global metric. We observe that using existing global metrics such as variation of information and adjusted rand index as a reward for the RL agent deteriorates its performance. This behaviour is because these metrics completely ignore the reply-to links between messages (local decisions) during reward computation. Therefore, we propose a novel thread-level reward function that captures the global metric without ignoring the local decisions. Through experiments on the Ubuntu IRC dataset, we demonstrate that the proposed RL model improves the performance on both link-level and conversation-level metrics.

show abstract

“…In this task, all comments are split into a number of threads. After that RTS could be exploited to discover the tree structure of the threads [ 5 , 10 , 16 ].…”

Section: Related Workmentioning

confidence: 99%

“…A list of dialog act labels and an approach for modeling dialogue acts have been proposed in conversational speech [ 30 – 32 ]. Detection of dialogue act labels for each post is suitable for thread detection [ 5 , 16 ] and finding relevant answers in forums [ 8 ].…”

Section: Related Workmentioning

confidence: 99%

A Supervised Approach to Predict the Hierarchical Structure of Conversation Threads for Comments

Balali

Faili

Asadpour

2014

The Scientific World Journal

View full text Add to dashboard Cite

User-generated texts such as comments in social media are rich sources of information. In general, the reply structure of comments is not publicly accessible on the web. Websites present comments as a list in chronological order. This way, some information is lost. A solution for this problem is to reconstruct the thread structure (RTS) automatically. RTS predicts a semantic tree for the reply structure, useful for understanding users' behaviours and facilitating follow of the actual conversation streams. This paper works on RTS task in blogs, online news agencies, and news websites. These types of websites cover various types of articles reflecting the real-world events. People with different views participate in arguments by writing comments. Comments express opinions, sentiments, or ideas about articles. The reply structure of threads in these types of websites is basically different from threads in the forums, chats, and emails. To perform RTS, we define a set of textual and nontextual features. Then, we use supervised learning to combine these features. The proposed method is evaluated on five different datasets. The accuracy of the proposed method is compared with baselines. The results reveal higher accuracy for our method in comparison with baselines in all datasets.

show abstract

Conversational Thread Extraction and Topic Detection in Text‐Based Chat

Cited by 6 publications

References 12 publications

Disentangling Chat

Disentangling Chat

End-to-End Deep Reinforcement Learning for Conversation Disentanglement

A Supervised Approach to Predict the Hierarchical Structure of Conversation Threads for Comments

Contact Info

Product

Resources

About