Papers by Peter Mihajlik
Is Spoken Hungarian Low-resource?: A Quantitative Survey of Hungarian Speech Data Sets (2024.lrec-main)
Copied to clipboard
Peter Mihajlik, Katalin Mády, Anna Kohári, Fruzsina Sára Fruzsina, Gábor Kiss, Tekla Etelka Gráczi, A. Seza Doğruöz
| Challenge: | Existing data sets in Hungarian are limited in quality and quality . however, it is difficult to train a modern automatic speech recognition system with thousands of hours of transcribed speech. |
| Approach: | They propose to analyze available speech data sets in Hungarian in five categories . they estimate that the available data sets are 2800 hours across 7500 speakers . |
| Outcome: | The available data sets in spoken Hungarian are compared to other languages and are estimated to be 2800 hours in size . however, their distribution and alignment to real-life tasks are far from optimal indicating the need for larger-scale natural language speech data sets. |
BEA-Base: A Benchmark for ASR of Spontaneous Hungarian (2022.lrec-1)
Copied to clipboard
| Challenge: | Hungarian is spoken by 15 million people, yet, easily accessible Automatic Speech Recognition (ASR) benchmark datasets are practically unavailable. |
| Approach: | They propose to use a subset of the BEA spoken Hungarian database to assess ASR, primarily for conversational AI applications. |
| Outcome: | The proposed framework achieves 45% reduction in recognition error rate compared to classical approach without external language model or additional supervised data. |