Unsupervised Model for Topic Viewpoint Discovery in Online Debates Leveraging Author Interactions
Bibliographic record
Abstract
Online debate forums provide a valuable resource for textual discussions about controversial social and political issues. Discovering the viewpoints and their discourse or arguments from such resources is important for policy and decision makers. In order to detect the stance, most of the existing methods rely on expensively obtained human annotations and propose supervised solutions. In this work, we introduce a purely unsupervised Author Interaction Topic Viewpoint model (AITV) for viewpoint identification at the post and the discourse levels. The model favors "heterophily" over "homophily" when encoding the nature of the authors' interactions in online debates. It assumes that the difference in viewpoints breeds interactions, unlike similar studies based on social network analysis, which hypothesize that similar viewpoints encourage interactions. We evaluate the model's viewpoint identification and clustering accuracies at the author and post levels. Experiments are held on six corpora about four different controversial issues, extracted from two online debate forums. AITV's results show a better performance in terms of viewpoint identification at the post level than the state-of-the-art supervised methods in terms of stance prediction, even though it is unsupervised. It also outperforms a recently proposed topic model for viewpoint discovery in social networks and achieves close results to a weakly guided unsupervised method in terms of author level viewpoint identification. Our results highlight the importance of encoding "heterophily" for purely unsupervised viewpoint identification in the context of online debates. We also carry out a brief qualitative evaluation of the discourse modeling in terms of Topic-Viewpoint word clusters. AITV shows encouraging results suggesting an accurate discovery of the viewpoints and topics' discourses.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.006 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".