Research map / Learning theory and algorithms / Prediction and decision losses
Linguistic Fingerprints of Internet Censorship: The Case of Sina Weibo
Kei Yin Ng, Anna Feldman, Jing Peng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use linguistic, sentiment, semantic, readability, frequency, and follower features in the MLP | MLP censorship classifier · Scraped Weibo posts | accuracy: 80.36 | Improved |
| Use linguistic, sentiment, semantic, readability, frequency, and follower features in the MLP | MLP censorship classifier · Zhu et al.'s dataset | accuracy: 87.63 | Improved |
| Exclude LIWC and word embeddings from the feature set (ablation) | MLP censorship classifier · Both datasets | accuracy: 75.4 -> 88.50 | Improved |
| Add at least three hidden layers (ablation) | MLP censorship classifier on both datasets · Both datasets | accuracy | Improved |
| Add additional hidden layers beyond three (ablation) | MLP censorship classifier · Both datasets | accuracy | No clear effect |
| Use LIWC features only (ablation) | Censorship classifier · Weibo datasets | accuracy: <60% | Hurt |
| Use CRIE features only (ablation) | Censorship classifier · Weibo datasets | accuracy: <60% | Hurt |
| Use word embeddings only (ablation) | Censorship classifier · Weibo datasets | accuracy: <60% | Hurt |
| Exclude CRIE features (ablation) | Censorship classifier · Weibo datasets | accuracy: <60% | Hurt |
| Train Naive Bayes, Logistic Regression, or SVM instead of the MLP (ablation) | Censorship classifier · Weibo datasets | accuracy: <65% | Hurt |
| Have non-expert humans predict censorship instead of using the classifier | Censorship prediction · Crowdsourced Weibo posts | accuracy: 23.83 | Hurt |
| Use verb categories to distinguish censored from uncensored posts (ablation) | Linguistic censorship analysis · Censored and uncensored Weibo posts | difference between classes | No clear effect |
Closest papers by meaning
- A Multi-Task Benchmark for Abusive Language Detection in Low-Resource Settings (NeurIPS 2025 · similarity 0.84)
- CHASM: Unveiling Covert Advertisements on Chinese Social Media (NeurIPS 2025 · similarity 0.83)
- Quantifying Political Polarization through the Lens of Machine Translation and Vicarious Offense (AAAI 2024 · similarity 0.82)
- When Neutral Summaries Are Not That Neutral: Quantifying Political Neutrality in LLM-Generated News Summaries (Student Abstract) (AAAI 2025 · similarity 0.81)
- Multi-Modal Discussion Transformer: Integrating Text, Images and Graph Transformers to Detect Hate Speech on Social Media (AAAI 2024 · similarity 0.81)
- SocialStigmaQA: A Benchmark to Uncover Stigma Amplification in Generative Language Models (AAAI 2024 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
