Department of Engineering

Machine Intelligence Laboratory

Zoi Roupakia

Zoi Roupakia

Position: PhD student in speech recognition


Supervisor: Prof. Mark Gales

Zoi Roupakia is a speech scientist and a Ph.D. candidate in Engineering in the Speech Research Group of the Machine Intelligence Laboratory in the University of Cambridge. She is working under the supervision of Professor Mark Gales. She holds an MPhil in Computer, Speech, Text and Internet Technology from the same university and a 5-year degree in Electrical and Computer Engineering from Aristotle University of Thessaloniki.

She also joined Google London for a 6-month internship working in the Research team of Speech Synthesis group. As a result of the project, a patent application entitled "Methods and Systems for Voice Conversion" was filed. Part of the project is in the product line, in the Google TTS app (version 3.8).

For more information, her CV is provided here.

Research Interests

  • Large vocabulary continuous speech recognition
  • Speaker adaptation
  • Kernel methods
  • Model selection
  • Voice conversion


  • Yiannis Agiomyrgiannakis, Zoi Roupakia, "Methods and Systems for Voice Conversion", Google Inc., Patent (filed), US 14/631464 (2015)

Publications - Theses

  • Yiannis Agiomyrgiannakis, Zoi Roupakia, "Voice morphing that improves TTS quality using an optimal dynamic frequency warping-and-weighting transform", accepted in ICASSP (2016)
  • Zoi Roupakia, Anton Ragni, Mark Gales, "Rapid Nonlinear Speaker Adaptation for Large-Vocabulary Continuous Speech Recognition", accepted for Proceedings in Interspeech (2012)
  • Zoi Roupakia, Mark Gales, "Kernel Eigenvoices (Revisited) for Large-Vocabulary Speech Recognition", accepted for IEEE Signal Processing Letters (2011)
  • Zoi Roupakia, "Nonlinear methods for Speech Processing", First year report, Cambridge University (2008)
  • Zoi Roupakia, "Nonlinear Projection Schemes", MPhil thesis, Cambridge University (2007)
  • Zoi Roupakia, "Development of a web portal for smart information searching and improvement of businesses competitive advantage (Smart B2B services)", Thesis, Aristotle University of Thessaloniki (2006)


Follow Speechstones , a hub of ideas, articles, research and tools for speech recognition and synthesis

Contact Information

Zoi Roupakia
Baker Building, Room 458
Engineering Department
Trumpington Street, Cambridge Email:
CB2 1PZ, UK Tel: +44 (0)1223 765 152

Real Time Web Analytics