AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Unsupervised Post-Processing of Word Vectors via Conceptor Negation

Tianlin Liu, Lyle Ungar, João Sedoc

AAAI 2019 · Architectures and efficiency › RNNs and normalization · approach: Learned embeddings for retrieval

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
Apply Conceptor Negation instead of the strongest reported baselineGloVe word vectors · SimLex-999Spearman correlation × 100: 44.97 -> 48.53Improved
Apply Conceptor Negation instead of the strongest reported baselineWord2Vec word vectors · STS 2012Pearson correlation × 100: 57.67 -> 54.31Hurt
Apply Conceptor Negation instead of the strongest reported baselineWord2Vec word vectors · STS 2013Pearson correlation × 100: 57.98 -> 59.17Improved
Apply Conceptor Negation instead of the strongest reported baselineGloVe word vectors · SICKPearson correlation × 100: 67.85 -> 66.42Hurt
Use Conceptor Negation for concept categorizationGloVe word vectors · APpurity × 100: 87.19 -> 90.95Improved
Use Conceptor Negation for concept categorizationWord2Vec word vectors · ESSLLI 2008purity × 100: 100.0 -> 100.0No clear effect
Initialize the Neural Belief Tracker with CN vectors instead of the strongest baselineGloVe word vectors · WOZ 2.0 foodgoal accuracy: 86.4 -> 88.8Improved
Initialize the Neural Belief Tracker with CN vectors instead of the strongest baselineWord2Vec word vectors · WOZ 2.0 foodgoal accuracy: 84.7 -> 78.5Hurt
Estimate the conceptor using only words appearing at least 200 times in the Wikipedia dump (ablation)Conceptor Negation · Word-vector post-processingperformanceImproved
Apply ABTT or CN to FastText vectors (ablation)FastText · Word-vector post-processingperformanceNo clear effect
Use ABTT or CN for text classification (ablation)Word vectors · Text classificationperformanceNo clear effect

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