Challenge: Qualities of a conversation are dependent on how interactions combine to form a “shape” of the conversation.
Approach: They propose a similarity measure to capture differences in conversation dynamics and assess its sensitivity to the topic of the conversation.
Outcome: The proposed measure captures differences in conversation dynamics and assesses its sensitivity to the topic of the conversation.

Similar Papers

We’ve had this conversation before: A Novel Approach to Measuring Dialog Similarity (2021.emnlp-main)

Copied to clipboard

Challenge: Dialogs are a building block of human natural language interactions.
Approach: They propose a new edit distance metric for dialog similarity analysis using conversation semantics, conversation flow, and the participants.
Outcome: The proposed method outperforms existing methods on two publicly available datasets and is better aligned with human perception of conversation similarity.
Probing the Robustness of Trained Metrics for Conversational Dialogue Systems (2022.acl-short)

Copied to clipboard

Challenge: Existing methods for evaluating conversational dialogue systems have been shown to be inefficient and instabile.
Approach: They propose an adversarial method to stress-test trained metrics for evaluation of conversational dialogue systems using Reinforcement Learning.
Outcome: The proposed method outperforms existing methods and can be applied to stress-test trained metrics for conversational dialogue systems.
TaskDiff: A Similarity Metric for Task-Oriented Conversations (2023.emnlp-main)

Copied to clipboard

Challenge: Popularity of chatGPT and Llama 2 has led to a race to build custom task-oriented conversational assistants in enterprise domains like finance and retail.
Approach: They propose a conversational similarity metric that uses different dialogue components to compute similarity.
Outcome: Experiments on a benchmark dataset show that the proposed metric outperforms existing approaches and is more robust than previous approaches.
Computational Analysis of Conversation Dynamics through Participant Responsivity (2025.emnlp-main)

Copied to clipboard

Challenge: Growing literature explores toxicity and polarization in discourse, with comparatively little work on characterizing what makes dialogue prosocial and constructive.
Approach: They develop and evaluate methods for quantifying responsivity through semantic similarity of speaker turns and large language models to identify the relation between two speaker turns.
Outcome: The proposed method is based on semantic similarity of speaker turns and large language models to identify the relation between two speaker turns.
Evaluation and Facilitation of Online Discussions in the LLM Era: A Survey (2025.emnlp-main)

Copied to clipboard

Challenge: Recent advances in LLMs enable artificial facilitation agents to not only moderate content, but also actively improve the quality of interactions.
Approach: They propose a taxonomy on discussion quality evaluation and a new taxonomies for intervention and facilitation strategies.
Outcome: The proposed methods synthesize ideas from Natural Language Processing (NLP) and Social Sciences to provide a taxonomy on discussion quality evaluation, and a roadmap of good practices and future research directions.
Can You Follow Me? Testing Situational Understanding for ChatGPT (2023.emnlp-main)

Copied to clipboard

Challenge: Existing studies have identified certain SU limitations in non-chatbot Large Language models, but the extent and causes of these limitations are not well understood.
Approach: They propose a synthetic environment for SU testing in chat-oriented models . they test models' ability to track and enumerate environment states .
Outcome: The proposed environment allows for controlled and systematic testing of SU in chat-oriented models, and to better understand underlying causes for performance patterns.
A Collection of Pragmatic-Similarity Judgments over Spoken Dialog Utterances (2024.lrec-main)

Copied to clipboard

Challenge: a new study uses pragmatic similarity measures to train speech synthesizers . the average inter-judge correlation between utterance pairs was 0.45.
Approach: they use a re-enactment of a recorded dialog to create pragmatic similarity . they use 9 judges to listen to 220 utterance pairs and rate them on a continuous scale .
Outcome: The results show that human judges listened to 220 utterance pairs and rated them on a continuous scale.
Characterizing Similarities and Divergences in Conversational Tones in Humans and LLMs by Sampling with People (2024.acl-long)

Copied to clipboard

Challenge: Existing taxonomies or text corpora suffer from experimenter bias and are not representative of real-world distributions.
Approach: They propose an iterative method for simultaneously eliciting conversational tones and sentences . they run 50 iterations with human participants and GPT-4 and obtain a dataset of sentences and frequent conversational tone.
Outcome: The proposed method can be used to characterize the differences between humans and LLMs.
DialogueTRM: Exploring Multi-Modal Emotional Dynamics in a Conversation (2021.findings-emnlp)

Copied to clipboard

Challenge: Existing studies focus on the self and inter-personal dependencies in multi-modal conversations, but they ignore the temporal and spatial dependencies.
Approach: They propose a Dialogue Transformer for simultaneously modeling the intra-modal and inter-modal emotion dynamics.
Outcome: The proposed models outperform the state-of-the-art on three benchmark datasets.
Assessing Dialogue Systems with Distribution Distances (2021.findings-acl)

Copied to clipboard

Challenge: Existing evaluation metrics focus on turnlevel quality, which is not well suited for open-end dialogue tasks.
Approach: They propose to measure the performance of a dialogue system by computing the distributionwise distance between its generated conversations and real-world conversations.
Outcome: The proposed metrics correlate better with human judgments than existing metrics on dialogue systems.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations