no code implementations • 10 Dec 2021 • Bryan Eikema, Germán Kruszewski, Hady Elsahar, Marc Dymetman
We show that we can sample from such EBMs with arbitrary precision at the cost of sampling efficiency.
1 code implementation • 10 Aug 2021 • Bryan Eikema, Wilker Aziz
The mode and other high-probability translations found by beam search have been shown to often be inadequate in a number of ways.
no code implementations • COLING 2020 • Bryan Eikema, Wilker Aziz
We argue that the evidence corroborates the inadequacy of MAP decoding more than casts doubt on the model and its training algorithm.
1 code implementation • WS 2019 • Bryan Eikema, Wilker Aziz
We present a deep generative model of bilingual sentence pairs for machine translation.