Implementing SRI’s Pashto speech-to-speech translation system on a smartphone

Citation

J. Zheng, A. Mandal, X. Lei, M. Frandsen, N. F. Ayan, D. Vergyri, W. Wang, M. Akbacak and K. Precoda, “Implementing SRI’s Pashto speech-to-speech translation system on a smartphone,” in Proc. 2010 IEEE Spoken Language Technology Workshop (SLT), pp. 133–138.

Abstract

We describe our recent effort implementing SRI’s UMPC-based Pashto speech-to-speech (S2S) translation system on a smart phone running the Android operating system. In order to maintain very low latencies of system response on computationally limited smart phone platforms, we developed efficient algorithms and data structures and optimized model sizes for various system components. Our current Android-based S2S system requires less than one-fourth the system memory and significantly lower processor speed with a sacrifice of 15 pct. relative loss of system accuracy, compared to a laptop-based platform.

Index Terms— speech-to-speech translation, mobile computing, smart phone, Android

↓ Download

Read more from SRI

  • An arid, rural Nevada landscape

    Can AI help us find valuable minerals?

    SRI’s machine learning-based geospatial analytics platform, already adopted by the USGS, is poised to make waves in the mining industry.

  • Two students in a computer lab

    Building a lab-to-market pipeline for education

    The SRI-led LEARN Network demonstrates how we can get the best evidence-based educational programs to classrooms and students.

  • Code reflected in a man's eyeglasses

    LLM risks from A to Z

    A new paper from SRI and Brazil’s Instituto Eldorado delivers a comprehensive update on the security risks to large language models.