AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Incorporating Discriminator in Sentence Generation: a Gibbs Sampling Method

Jinyue Su, Jiacheng Xu, Xipeng Qiu, Xuanjing Huang

AAAI 2018 · Large language models › Language, safety and interpretability · approach: Controlled text generation · 30 citations (OpenAlex)

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 changedWhereMeasuredEffect
Use Gibbs Sampling instead of Beam SearchConstrained sentence generation · SST-2BLEU-4 (w/ penalty): 1.290 -> 2.060Improved
Use Gibbs Sampling instead of Beam SearchConstrained sentence generation · SST-2BLEU-4 (w/o penalty): 6.059 -> 8.302Improved
Use Gibbs Sampling instead of Reject SamplingConstrained sentence generation · SST-2BLEU-4 (w/ penalty): 0.384 -> 2.060Improved
Use Gibbs Sampling instead of Reject SamplingConstrained sentence generation · SST-2BLEU-4 (w/o penalty): 4.029 -> 8.302Improved
Use Gibbs Sampling instead of Beam SearchConstrained sentence generation · ProductBLEU-4 (w/ penalty): 0.218 -> 1.476Improved
Use Gibbs Sampling instead of Beam SearchConstrained sentence generation · ProductBLEU-4 (w/o penalty): 4.556 -> 4.750Improved
Use Gibbs Sampling instead of Reject SamplingConstrained sentence generation · ProductBLEU-4 (w/ penalty): 0.369 -> 1.476Improved
Use Gibbs Sampling instead of Reject SamplingConstrained sentence generation · ProductBLEU-4 (w/o penalty): 3.323 -> 4.750Improved
Use Gibbs Sampling instead of Random generationConstrained sentence generation · SST-2BLEU-4 (w/ penalty): 4.338 -> 2.060Hurt
Use Gibbs Sampling instead of Random generationConstrained sentence generation · SST-2BLEU-4 (w/o penalty): 5.888 -> 8.302Improved
Use Gibbs Sampling instead of Random generationConstrained sentence generation · ProductBLEU-4 (w/ penalty): 2.093 -> 1.476Hurt
Use Gibbs Sampling instead of Random generationConstrained sentence generation · ProductBLEU-4 (w/o penalty): 2.854 -> 4.750Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy