THE CONFERENCE BEGINS IN —
PROGRAMMEThe time of the programme events is given in CEST (Central European Summer Time) or UTC+2. In the sessions where there are online presentations ( or ), a link will be published right in this programme (next to the assigned room) before the respective session begins.
| 08:00 – 12:00 | Registration (open on Tue and Wed, too) — @Venue Entrance | |
| 12:30 – 14:00 | Lunch Break (this lunch is not included in the fee) | |
| 14:00 – 14:30 |
Official Opening of the 28th International Conference TSD2025 — @Main Lecture Hall
Elmar Nöth (TSD2025 Programme Committee Chairman) |
|
| 14:30 – 16:10 |
Oral Session 1
— @Main Lecture Hall
Chairperson: Ivandre Paraboni Jakub Šmíd: Large Language Models for Czech Aspect-Based Sentiment Analysis Jakub Šmíd: Few-shot Cross-lingual Aspect-Based Sentiment Analysis with Sequence-to-Sequence Models Chenhao Chen: CRNLI: A Textual Entailment Dataset in the Chemistry Domain Adnan Ahmad: Parameter vs. Sample Efficiency in Multi-intent Recognition for Dialogue Understanding: Benchmarking Small |
Oral Session 2
— @Lecture Hall B
& Online <log in to see links>
Chairpersons: Sebastian Bayerl & Rafa Orozco Marko Čechovič: Corpus of Cross-lingual Dialogues with Minutes and Detection of Misunderstandings Vladislav Stankov: ParCzech4Speech: A New Speech Corpus Derived from Czech Parliamentary Data Sebastian Peter Bayerl: Multilingual Stutter Event Detection for English, German, and Mandarin Speech Melissa Torgbi: Inclusive ASR for Critical Public Services: Debiasing with Actor-Simulated Speech |
| 16:10 – 16:40 | Coffee Break | |
| 16:40 – 17:30 |
Oral Session 3
— @Main Lecture Hall
Chairperson: Paula Perez Ivandre Paraboni: Tracking Mental Health Indicators on Social Media Before and After Diagnosis Petr Zelina: Computing Patient Similarity Based on Unstructured Clinical Notes |
Oral Session 4
— @Lecture Hall B
Chairperson: Šárka Zikánová Anastasia Zhukova: Efficient Domain-adaptive Continual Pretraining for the Process Industry in the German Language Adam Mištera: Enhancing Masked Language Modeling in BERT Models Using Pretrained Static Embeddings |
| 18:00 – 22:00 | Welcome Reception Meeting point: At the Conference Venue |
|
| 08:00 – 9:00 | Registration — @Venue Entrance | |
| 09:00 – 10:00 |
Keynote Talk — @Main Lecture Hall
Heidi Christensen: From Code to Clinic: Making Speech-Based AI for Cognitive Health Work in the Real World ( link to the presentation) |
|
| 10:00 – 10:30 | Coffee Break | |
| 10:30 – 12:10 |
Oral Session 5
— @Main Lecture Hall
Chairperson: Heidi Christensen Yan Meng: Robust Disfluency Labeling in Spontaneous Speech: Insights from Diverse Hungarian Corpora Including Mentally Ill Speakers Ishaan Mahapatra: Systematic FAIRness Assessment of Open Voice Biomarker Datasets for Mental Health and Neurodegenerative Diseases Daniel Escobar-Grisales: Verb Motility Dynamics Reveals Cognitive Impairment in Parkinson's Disease: A Speech-Language Fusion Approach Terry Yi Zhong: RECA-PD: A Robust Explainable Cross-Attention Method for Speech-based Parkinson's Disease Classification |
Oral Session 6
— @Lecture Hall B
& Online <log in to see links>
Chairperson: Juan Camilo Vásquez Kesego Mokgosi: Synthesising Cross-Speaker Data for Low-Resource Pathological Speech Recognition with PEFT Dalai Mengke: How Far Can Synthetic Speech Go? Enhancing ASR in Low-Resource Scenarios via Voice Cloning Duygu Altinok: Mind the Gap: Entity-Preserved Context-Aware ASR for Structured Transcriptions Duygu Altinok: Boosting CTC-Based ASR Using LLM-Based Intermediate Loss Regularization |
| 12:30 – 13:45 | Lunch Break (lunch included in the fee) | |
| 13:45 – 14:45 |
Keynote Talk — @Main Lecture Hall
Bernd Möbius: Information Density and Phonetic Variation ( link to the presentation) |
|
| 14:45 – 15:30 | Coffee Break | |
| 15:30 – 16:30 |
Poster Session 1
— Main Lecture Hall
Aleš Pražák: Lightweight Target-Speaker-Based Overlap Transcription for Practical Streaming ASR Yanis Labrak: An Empirical Analysis of Discrete Unit Representations in Speech Language Modeling Pre-training Janine Rugayan: Optimizing ASR Models with Semantic Information Lukáš Matějů: Efficient Enhancement of Norwegian ASR Model Marie Kunešová: An Exploration of ECAPA-TDNN and x-vector Speaker Representations in Zero-shot Multi-speaker TTS Yuxuan Zhang: Beyond Static Emotions: Leveraging Multitask Learning to Model Dynamics of Dimensional Affect in Speech Christopher Simic: Combining Temporal Visual Dynamics and Audio Representations for Robust Speaker Identification Abner Hernandez: Enhancing ASR Accuracy for Speakers with Parkinson's Disease Using Instruction-Tuned LLMs Tomáš Lebeda: Automatic Cognitive Disorder Detection through Semantic Analysis of Verbal Image Descriptions Esau Villatoro-Tello: Unifying Global and Near-Context Biasing in a Single Trie Pass Dominik Wagner: Vocoder-Free Non-Parallel Conversion of Whispered Speech With Masked Cycle-Consistent Generative Adversarial Networks Andreas Rouvalis: Enhancing Detection of Parkinson-induced Dysarthria with Cross-lingual Transfer Learning Jan Tupý: Detection of Cognitive Disorders Using ASR-Based Nonsense Words Repetition Mykhailo Danilevskyi: Towards an Accurate Domain-Specific ASR: Transcription for Pathology Michal Novák: Automated Speaking Assessment for L2 Learners of Czech Matthias Busch: When Silence Speaks: Understanding Open-Ended Responses via LLMs in Therapeutic Voice Interaction |
|
| 17:10 – 18:10 |
TSD Programme Committee Meeting — @Lecture Hall B
& Online <log in to see links>
(TSD Programme Committee members only)
|
|
| 18:30 – 22:00 |
Conference Dinner at Entla's Keller Meeting point: At the Entla's Keller |
|
| 08:00 – 9:00 | Registration — @Venue Entrance | |
| 09:00 – 10:30 |
Students Meet Experts — @Main Lecture Hall
Host/Moderator: TBA |
|
| 10:30 – 11:00 | Coffee Break | |
| 11:00 – 12:00 |
Poster Session 2
— Main Lecture Hall
Manying Zhang: Product Recommendation with Prospect Theoretic Self-Aligned LLM Systems Petr Pechman: Refining Czech GEC: Insights from a Multi-Experiment Approach Haoyang Chen: Scale-Free Characteristics of Legal Texts and the Limitations of LLMs Andrei-Alexandru Manea: Investigating the Effect of Parallel Data in the Cross-Lingual Transfer for Vision-Language Encoders Vikram Ramanarayanan: Toward Quantifying How The Burden of Problems Reported By Patients Evolves in Parkinson's Disease Patrik Stano: Evaluating Prompt-Based and Fine-Tuned Approaches to Czech Anaphora Resolution Erolcan Er: Multilingual Implicit Discourse Relation Recognition via Abstract Object-Enhanced Chain-of-Thought Prompting Michael Neumann: Leveraging Fine-Tuned State-of-the-Art LLMs for Symptom Classification of Patient-Reported Problems in Parkinson's Disease |
|
| 12:30 – 13:45 | Lunch Break (lunch included in the fee) | |
| 13:45 – 15:25 |
Oral Session 7
— @Main Lecture Hall
Chairperson: Bernd Möbius Daiqi Liu: Audio–Vision Contrastive Learning for Phonological Class Recognition Camilo Vasquez: Emotion-Aware Speech-Driven Facial Avatar Animation via Joint Blendshape Prediction and Emotion Recognition Daniel Tihelka: Sentences vs Phrases in Neural Speech Synthesis: the Phrases Strike Back Lukáš Vladař: Evaluating Phoneme-Level Pretraining in Czech Text-to-Speech Synthesis |
Oral Session 8
— @Lecture Hall B
Chairperson: Aleš Horák Jonathan Jordan: Plant in Cupboard, Orange on Rably, Inat Aphone. Benchmarking Incremental Learning of Situation and Language Model using a Text-Simulated Situated Environment Zsolt Szántó: Knowledge Representation Approaches for Educational Question Generation Šárka Zikánová: Gold Data and Multiple Understanding of Discourse Relations Keara Schaaij: Towards Stable and Personalised Profiles for Lexical Alignment in Spoken Human-Agent Dialogue |
| 15:25 – 16:00 | Coffee Break | |
| 16:00 – 17:00 |
Keynote Talk — @Main Lecture Hall (Online)
& Online <log in to see links>
Shrikanth Narayanan: Speech-centered Machine Intelligence and Possibilities for Human Health and Wellbeing ( link to the presentation) |
|
| 18:00 – 19:00 | Dinner (not included in the fee) | |
| 09:00 – 10:40 |
Oral Session 9
— @Main Lecture Hall
& Online <log in to see links>
Chairperson: Elmar Nöth Yassin Terraf: TOSD-Net: A CNN-Transformer Architecture for Robust Frame-Level Overlapping Speech Detection in Diverse Acoustic Conditions Felix Herron: Implicit Speaker Group Encoding in Self-supervised Speech Recognition Models Alper Karamanlioglu: Multilingual Domain Adaptation for Speech Recognition Using LLMs Simen Dymbe: Using Cross-attention For Conversational ASR Over The Telephone |
Oral Session 10
— @Lecture Hall B
Chairperson: Jana Straková Jana Straková: Flexing in 73 Languages: A Single Small Model for Multilingual Inflection Vlasta Ohlídalová: Are We There yet? A Thorough Evaluation of POS Tagging on Czech Michal Olbrich: Morphological Segmentation with Neural Networks: Performance Effects of Architecture, Data Size, and Cross-Lingual Transfer in Seven Languages Kertu Saul: Automatic Semantic Tagging of Estonian Spatial Adverbials for Valency Pattern Mining |
| 10:40 – 11:10 | Coffee Break | |
| 11:10 – 11:40 |
Closing Ceremony of the 28th International Conference TSD2025 — @Main Lecture Hall
Awards for the Best Presentations & Closing Words: Elmar Nöth & TSD2025 Organizing Committee Members |
|
| 12:00 – 13:00 | Lunch Break (lunch included in the fee) | |
| 13:00 – ... | Conference Trip to Nürnberg
Meeting point: In front of the Conference Venue |
|
To ensure that everything runs smoothly, we invite all online presenters to join a technical try-out session on Monday, 25 August 2025, at 10:00 AM CEST (UTC+2) using the Zoom link to the Main Lecture Hall: <log in to see links>.
This will be an opportunity to:
