Christian Fuegen

Christian joined Facebook as a research scientist in 2013 due to the acquisition of Mobile Technologies. At Mobile Technologies, he was one of the core developers of “Jibbigo”, an on-device speech-to-speech translator for mobile devices, where he worked from 2007 until 2013, first as a research scientist and later as director of research. Christian received a doctorate for his work on simultaneous translation systems of lectures and speeches from University of Karlsruhe (TH) in 2008 in which he developed core components for a first-ever simultaneous lecture translation system, including innovations in speech recognition, segmentation, adaptation and work on real-time and latency requirements for simultaneous speech translation systems. At Facebook, his team is focusing on advancing the state-of-art of speech recognition and human machine interaction by executing research on self/ unsupervised learning, domain transfer learning, on-device modeling, multi-microphone recognition, or acoustic event detection.

Christian's Publications

May 01, 2020

RESEARCH

NLP

From Senones to Chenones: Tied Context-Dependent Graphemes for Hybrid Speech Recognition

There is an implicit assumption that traditional hybrid approaches for automatic speech recognition (ASR) cannot directly model graphemes and need to rely on phonetic lexicons to get competitive performance, especially on English which has poor…

Duc Le, Xiaohui Zhang, Weiyi Zhang, Christian Fuegen, Geoffrey Zweig, Michael L. Seltzer,

May 01, 2020

May 01, 2020

RESEARCH

NLP

Joint Grapheme and Phoneme Embeddings for Contextual End-to-End ASR

End-to-end approaches to automatic speech recognition, such as Listen-Attend-Spell (LAS), blend all components of a traditional speech recognizer into a unified model. Although this simplifies training and decoding pipelines, a unified model is…

Zhehuai Chen, Mahaveer Jain, Yongqiang Wang, Michael L. Seltzer, Christian Fuegen,

May 01, 2020

May 01, 2020

RESEARCH

NLP

Towards End-to-End Spoken Language Understanding

Spoken language understanding system is traditionally designed as a pipeline of a number of components. First, the audio signal is processed by an automatic speech recognizer for transcription or n-best hypotheses. With the recognition results,…

Dmitriy Serdyuk, Yongqiang Wang, Christian Fuegen, Anuj Kumar, Baiyang Liu, Yoshua Bengio,

May 01, 2020