We are very pleased to announce that 4 submissions authored by members of AGH’s Signal Processing Group were accepted for presentation at the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2025), which will take place in Hyderabad, India from April 06 to April 11, 2025.
ICASSP paper presentation:
Mateusz Barański, Jan Jasiński, Julitta Bartolewska, Stanisław Kacprzak, Marcin Witkowski, and Konrad Kowalczyk, “Investigation of Whisper ASR Hallucinations Induced by Non-Speech Audio”, accepted for IEEE ICASSP 2025, paper ID: 5502.
SPS Journal presentations:
Magdalena Rybicka, Jesus Villalba, Thomas Thebaud, Najim Dehak and Konrad Kowalczyk, “End-to-End Neural Speaker Diarization With Non-Autoregressive Attractors”, paper ID: 8586.
Mateusz Guzik and Konrad Kowalczyk, “On Ambisonic Source Separation With Spatially Informed Non-Negative Tensor Factorization”, paper ID: 8587.
Mieszko Fraś and Konrad Kowalczyk, “Reverberant Source Separation Using NTF With Delayed Subsources and Spatial Priors”, paper ID: 8588.
This year DSP team will present our research into:
1. Hallucinations in neural automatic speech recognition,
2. Neural speaker diarization – joint work with Center for Language and Speech Processing at Johns Hopkins University,
3. Ambisonic audio signal separation,
4. Joint speech separation and dereverberation.