Robert L. Mercer
Robert L. Mercer contributed to statistical speech and language research at IBM. He coauthored A Statistical Approach to Machine Translation in 1990. With Peter Brown and Stephen and Vincent Della Pietra, he published the detailed 1993 account of translation-model parameter estimation. It presents five models and probabilistic word alignments learned from bilingual sentence pairs, using French and English Canadian parliamentary data. ACL elected him a fellow in 2011 for speech recognition and statistical translation.
- Works
- 31
- Citations
- 12,285
- h-index
- 20
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
1990
A shared statistical translation approach
Mercer coauthored A Statistical Approach to Machine Translation with Peter Brown and six other researchers. The paper appeared in 1990. The collaboration continued in later translation-model work.
-
1993
Learning translation model parameters
With Brown, Stephen Della Pietra and Vincent Della Pietra, Mercer described five statistical translation models. The paper estimates model parameters from pairs of sentences that translate one another. All four authors were affiliated with IBM T.J. Watson Research Center.
-
1993
Finding word relationships across languages
The same paper defines probabilistic word-by-word alignments and a suboptimal algorithm to seek a likely alignment. Its training data uses French and English proceedings of the Canadian Parliament.
-
2011
Recognition in computational linguistics
ACL elected Mercer a fellow in 2011 for work on speech recognition and statistical machine translation.
Sources and credits
Photo: We have no published photo of Robert L. Mercer yet. Send us a free-licensed photo.