AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Offline bilingual word vectors, orthogonal transformations and the inverted softmax

Samuel L. Smith, David H. P. Turban, Steven Hamblin, Nils Y. Hammerla

ICLR 2017 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 257 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 the SVD alignment with inverted softmax and dimensionality reduction instead of Mikolov's linear mappingBilingual word translation with the expert word dictionary · English to Italianprecision@1: 0.338 -> 0.431Improved
Add inverted softmax retrieval to SVD alignment (ablation)Bilingual word translation with the expert word dictionary · English to Italianprecision@1: 0.369 -> 0.417Improved
Add dimensionality reduction after inverted-softmax retrieval (ablation)Bilingual word translation with the expert word dictionary · English to Italianprecision@1: 0.417 -> 0.431Improved
Add inverted softmax retrieval to SVD alignment (ablation)Bilingual word translation with the expert word dictionary · Italian to Englishprecision@1: 0.322 -> 0.373Improved
Add dimensionality reduction after inverted-softmax retrieval (ablation)Bilingual word translation with the expert word dictionary · Italian to Englishprecision@1: 0.373 -> 0.380Improved
Replace the expert word dictionary with an identical-character-string pseudo-dictionarySVD bilingual word alignment · English to Italianprecision@1: 0.010 -> 0.399Improved
Train alignment using paired Europarl sentence vectors instead of a word dictionarySVD bilingual word alignment · English to Italianprecision@1: 0.313 -> 0.428Improved
Use inverted softmax instead of nearest-neighbor retrieval (ablation)SVD sentence-vector alignment with a phrase dictionary · Italian to English, 200k candidate sentencesprecision@1: 0.656 -> 0.486Hurt
Use paired sentence vectors as the training dictionarySVD bilingual word alignment · English to Italian, 200k candidate sentencesprecision@1: 0.268 -> 0.431Improved
Use paired sentence vectors as the training dictionarySVD bilingual word alignment · Italian to English, 200k candidate sentencesprecision@1: 0.473 -> 0.656Improved

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