About Me
Soheil Khorram is a Research Associate in Computer Science and Engineering at the University of Texas at Dallas. His main expertise is in machine learning and speech processing including speech enhancement/separation, affective computing, emotion recognition, computational paralinguistics and speech synthesis.
In 2006, Soheil ranked second among 8998 participants of the Iranian nationwide graduate school entrance exam in the field of Computer Engineering and he started an M.Sc. program at the Sharif University of Technology (SUT), Tehran, Iran. He received his M.Sc. and Ph.D. in Computer Engineering from SUT in 2008 and 2015, respectively. His Master’s had a strong emphasis on signal processing and speech enhancement to improve speech recognition systems and his Ph.D. was focused on developing and enhancing text-to-speech systems. During his Ph.D., in 2013, he was awarded a scholarship from the ASR Gooyesh Pardaz Company to participate in a research program at the Center for Speech Technology Research (CSTR), the University of Edinburgh, where he worked as a member of the Simple4All project under the supervision of Prof. Simon King to improve acoustic modeling of the statistical parametric speech synthesis systems.
After his Ph.D program, he started a postdoc at the university of Michigan (UMICH) and worked under the supervision of Prof. Emily Mower Provost and Prof. Melvin McInnis. He was with the UMICH from December 2015 to October 2018 and he worked on affective computing and computational paralinguistics as a part of the PRIORI project. In UMICH, he also supervised 22 university students in a span of two years in charge of collecting the PRIORI Emotion Dataset. He is currently working at the CRSS lab, the University of Texas at Dallas, under the supervision of Prof. John Hansen and his main focus is on speech enhancement/separation systems specifically improving speech intelligibility for cochlear implant users.
Research Interests
- Deep Learning (CNN, LSTM/GRU, Deformable CNN, GAN, Attention, Transformer)
- Speech Processing (Emotion Recognition/Affective Computing, Speech Separation, Speech Enhancement, Speech Synthesis)
- Machine Learning, Pattern Recognition, Graphical Models (HMM, CRF)
Leadership and Awards
- Managed 18 students in the task of labeling speaker/gender information of the UTD forensic speaker recognition project in 2019.
- Supervised 22 students of the University of Michigan in a 2-year period (07/2016–08/2018) for annotating emotion labels of the PRIORI Emotion Dataset. The collected dataset contains 42 hours of telephonic conversations of subjects with bipolar disorder.
- Received two conference travel awards from the University of Michigan in 2016.
- Received a scholarship from the ASR Gooyesh Pardaz Company to participate in a research program at the University of Edinburgh (CSTR lab) in 2013.
- Earned distinction for being qualified and competent project manager in ASR-Gooyesh Pardaz Company in 2012.
- Ranked first among Computer Architecture M.Sc. students of the Sharif University of Technology in 2009.
- Ranked second among Computer Hardware Engineering B.Sc. students of the Shahid Bahonar University in 2006.
- Ranked second among 8998 participants of the Iranian nationwide graduate school entrance exam in the field of Computer Engineering (Computer Architecture), in 2006.