Papers by Piotr Żelasko
An Application for Building a Polish Telephone Speech Corpus (L18-1)
Copied to clipboard
| Challenge: | Specifically, we describe a tool designed to improve our Automatic Speech Recognition system performance. |
| Approach: | They propose to build a tool for speech corpus collection of a specific domain content. |
| Outcome: | The proposed tool can be used to gather 63 hours of speech recordings across several domains and achieve lower WER in two grammar-based speech recognition tasks. |
What Helps Transformers Recognize Conversational Structure? Importance of Context, Punctuation, and Labels in Dialog Act Recognition (2021.tacl-1)
Copied to clipboard
| Challenge: | Existing punctuation in the transcripts has a massive effect on the models’ performance, and specific label set specificity does not affect dialog act segmentation performance. |
| Approach: | They apply two pre-trained transformer models to a conversation transcript as a sequence of dialog acts and achieve strong results on Switchboard Dialog Act and Meeting Recorder Dialog Act corpora. |
| Outcome: | The proposed models achieve 8.4% and 14.2% error rates on the Switchboard Dialog Act and Meeting Recorder Dialog Act corpora. |
WER we are and WER we think we are (2020.findings-emnlp)
Copied to clipboard
Piotr Szymański, Piotr Żelasko, Mikolaj Morzy, Adrian Szymczak, Marzena Żyła-Hoppe, Joanna Banaszczak, Lukasz Augustyniak, Jan Mizgajski, Yishay Carmiel
| Challenge: | Recent reports of very low word error rates (WERs) achieved by modern automatic speech recognition systems are skepticism towards the accuracy of modern systems. |
| Approach: | They propose to use a dataset to test automatic speech recognition systems . they propose guidelines for creating real-life datasets with high quality annotations . |
| Outcome: | The proposed system achieves 81% of accuracy on human-chatbot interactions compared to the best reported results on human conversations and public benchmarks. |
Expanding Abbreviations in a Strongly Inflected Language: Are Morphosyntactic Tags Sufficient? (L18-1)
Copied to clipboard
| Challenge: | In this paper, the problem of recovery of morphological information lost in abbreviated forms is addressed . correct inflected form of expanded abbrevation can be deduced from context words . |
| Approach: | They propose a deep bidirectional LSTM network with tag embedding to predict abbreviated words . they train on 10 million words from the Polish Sejm Corpus and achieve 74.2% prediction accuracy . |
| Outcome: | The proposed model achieves 74.2% accuracy on a smaller but more general corpus of Polish words. |
Why Aren’t We NER Yet? Artifacts of ASR Errors in Named Entity Recognition in Spontaneous Speech Transcripts (2023.acl-long)
Copied to clipboard
Piotr Szymański, Lukasz Augustyniak, Mikolaj Morzy, Adrian Szymczak, Krzysztof Surdyk, Piotr Żelasko
| Challenge: | despite advances in language models, the transcript of spontaneous human-human conversations remains an insurmountable challenge for most models. |
| Approach: | They examine the relationship between ASR and NER errors which limit NER models' ability to recover entity mentions from spontaneous speech transcripts. |
| Outcome: | The proposed model fails even if no word errors are introduced by the ASR . the proposed model's performance deteriorates when applied to the ASL outputs . |