Keywords:
- accent embedding
- Accented speech
- Accentual mismatch
- acoustic generators
- Acoustic model adaptation
- acoustic modeling
- adaptation
- ADS-B data
- Aho-Corasick algorithm
- air surveillance data
- Air traffic control
- air traffic control communications
- air traffic controller
- air traffic controller’s workload
- air traffic management
- Air-Traffic Communication (ATC)
- Alzheimer's disease
- AM
- Anti-spoofing
- Arithmetic Coding
- Artificial intelligence
- Artificial Neural Networks
- ASR
- ASR robustness
- Assistant Based Speech Recognition
- association rules
- audio and voice analysis
- Audio Coding
- audiobook
- Automatic Depression Detection
- Automatic Speech Recognition
- automatic speech recognition (ASR)
- automatic speech recognition and understanding
- automatic speech understanding
- batch norm
- batch normalization
- bayesian fusion
- Benchmark
- Benchmarking
- BERT
- bias
- bias aware
- bias word error rate
- BNF
- Building Blocks
- call sign detection
- Call-sign Detection
- Call-sign Recognition
- chunking
- claim verification
- Clinical Interviews
- Command Prediction Model
- command recognition rate
- Confidence Measure (CM)
- contact-center speech
- Contextual Adaptation
- contextual ASR
- contextual biasing
- Contextualisation and adaptation of ASR
- conversational agents
- conversational AI
- conversational hallucination
- conversational modeling
- Convolutional Neural Networks
- crime scene dataset
- Criminal investigations
- Cross-modal Alignment
- Cross-modal Attentio
- Cross-modal Attention
- Customization of model
- data analysis
- Data Selection
- dataset
- deep learning
- Deep learning for speech
- deep MLPs
- Deep neural network
- deep neural networks
- Delays
- Depression Corpora
- depression detection
- dialogue
- dialogue evaluation
- dialogue flow
- dialogue simulation
- diarization
- digital mental health
- direction of arrival
- direction-of-arrival estimation
- Discourse Annotation
- Discriminative features
- DISPLACE-2
- dnn
- DOA estimation
- domain adaptation
- Domain Classification
- dropout
- Dual mode encoder
- ECAPA-TDNN embedding
- electronic flight strips
- embedding
- Encoding
- end-to-end
- end-to-end ASR
- entity linking
- entity recognition
- entity-aware evaluation
- Entropy Coding
- Environmental mismatch
- Estimation
- explainability
- F1 score
- face verification
- fact checking
- factual reporting
- Feature extraction
- fine-tuning
- finite-state transducers
- flow-aware metrics
- FM
- fmllr
- Forensics
- Foundation Models
- Frequency Domain Linear Prediction (FDLP)
- fvae-lora
- gaming
- GDPR
- GMM
- GPU decoding
- Graph Convolutional Network (GCN)
- Graph Convolutional Networks
- Graph Neural Networks
- hallucination
- hallucination detection
- high-definition video-conferencing
- HTK
- Huffman Coding
- human factors
- Human-Computer Interaction
- human-robot interaction
- hybrid system
- i-vector
- i-vectors
- information verification
- Integration of prior knowledge
- Intent Classification
- inter-task fusion
- Interpretability
- Interpretable Models
- Iterative learning
- KeyWord Spotting (KWS)
- Keyword spotting detection
- KL-HMM
- knowledge distillation
- lan- guage identification
- language identification
- Language IDentification (LID)
- language modeling
- Language Models
- Language Production
- Language targets
- Large Language Models
- Large Vocabulary Continuous Speech Recognition (LVCSR)
- latent space factorization
- Lattice-Free MMI
- LEA
- legal framework
- LID
- likelihood-based encoding
- limited training data
- Linear prediction
- LLM
- LLM evaluation
- LLM-based ASR
- local speaker segmentation
- logistic regression
- LoRA
- Low resource language
- low-rank adaptation
- low-resource
- LVCSR
- machine learning
- Machine Translation
- media bias
- Mental Lexicon
- MFCC
- microphone arrays
- Microphones
- model adaptation
- model fine-tuning
- multi-face tracking
- multi-lingual automatic speech recognition
- multi-lingual SAD
- Multi-modal Approach
- multi-modal database
- multi-task
- multilingual acoustic modeling
- Multilingual automatic speech recognition
- Multimodal machine translation
- multimodal signal processing
- multiple remote tower
- multiple sound sources
- multiple speaker detection
- multitask acoustic modeling
- multitask learning
- multitask training
- named entity recognition
- Natural language processing
- network analysis
- network output
- neural nets
- neural network
- neural network-based sound source localization methods
- neural networks
- news media
- node weighted graphs
- non-native speech
- online speech recognition
- OOV-word recognition
- open-architecture distributed system
- OpenSky Network
- Operant Motive Test
- orchestration
- OSINT
- Out- Of-Language (OOL) detection
- out-of-domain
- Out-Of-Language (OOL) detection
- parametric speech synthesis
- parametric synthesis
- perceptual evaluation of audio quality (PEAQ)
- personal data processing
- personas
- PLDA
- Position measurement
- procedural hallucination
- procedural similarity
- prompt projection
- Prompting
- pseudo-labelling
- Psycholinguistics
- rare word recognition
- Rare-word integration
- Raw Speech
- real-time ASR
- real-time audio processing
- real-time processing
- real-time speech recognition
- recurrent neural network
- reinforcement learning
- reliability estimation
- Representation and Processing
- reproducibility
- resources and evaluation
- Robots
- Robust Automatic Speech Recognition
- ROXANNE
- ROXSD
- saftety
- scenario management
- self-supervised learning
- self-supervised pre-training
- semi-supervised learning
- Semi-supervised training
- sensor fusion
- sentence embeddings
- Sentiment Analysis
- SGMM
- SGMM adaptation
- shallow fusion
- signal processing
- simulated Call Detail Records (CDRs)
- simultaneous detection
- single sound source
- situation awareness
- sound mixtures
- sound source localization
- spatial spectrum-based approaches
- speaker adaptation
- Speaker change detection
- speaker clustering
- Speaker Diarization
- Speaker identification
- speaker recognition
- speaker role classification
- speaker role detection
- speaker role identification
- speaker turn detection
- speaker verification
- Speech activity detection
- speech coding
- speech dataset
- speech decoding
- speech LLMs
- speech meta-data
- speech quality evaluations
- speech recognition
- speech synthesis
- speech understanding
- Speech-to-LLM alignment
- speech-to-text alignment
- spoken dialogue systems
- Spoken Language Understanding
- Spoken Term Detection (STD)
- spurious correlation robustness
- streaming ASR
- streaming transducer
- Subs-ace Gaussian Mixture Models
- subspace Gaussian mixture models
- supervised adaptation
- Supervised Autoencoders
- supervision
- synthetic dialogue
- synthetic NCD dataset
- System Combination
- Tandem
- task-oriented dialog
- task-oriented dialogue
- Text classification
- text denoising
- Text fine-tuning
- Text Representation
- text to speech
- Text-based speaker diarization
- text-to-speech
- text-to-speech synthesis
- topic modelling
- tower utterances
- TRACY · Law Enforcement Agencies · Suspect Detection· Non-Content Data· Social Influence Analysis· Link Prediction
- TRACY· Non-Content data· Law Enforcement Agencies · Suspect Detection· Mobile Signaling Data· ROXANNE
- training
- transfer learning
- transformer transducer
- transformers
- TTS
- Under-resourced data
- under-resourced languages
- under-resourced speech recognition
- unsupervised learning
- user identity linkage
- verification
- Very low bit rate speech coding
- voice-activity detection
- wav2vec 2.0
- wav2vec2
- weakly-supervised learning.
- Web data
- weighted finite state transducer
- WFST
- whisper
- Whisper models
- Word Consensus Networks
- Word-Confusion-Networks
- XLS-R
- XLSR
- XLSR-Transducer
- Zipformer
Publications of Petr Motlicek sorted by recency
| SDialog: A Python Toolkit for End-to-End Agent Building, User Simulation, Dialog Generation, and Evaluation, , , , , , , , , and , in: EACL, 2026 |
|
| Generating Synthetic Doctor-Patient Conversations for Long-form Audio Summarization, , , , , , , , , , , , , and , in: Interspeech 2026, 2026 |
[DOI] |
| How to Leverage Synthetic Speech for LLM-Based ASR Systems?, , , , , , , , , , , and , in: IEEE Spoken Language Technology 2026, 2026 |
[URL] |
| Geometric Latent Reasoning Induces Shorter Generations in LLMs, , , , and , in: The Fortieth Annual Conference on Neural Information Processing Systems, 2026 |
[URL] |
| Abstract Compression of Conversational Audio Context for LLM-based ASR, , , , , , , , , and , in: IEEE Spoken Language Technology 2026, 2026 |
[URL] |
| When Synthetic Speech Is All You Have: Better Call GRPO, , , , , , , and , in: IEEE Spoken Language Technology 2026, 2026 |
[URL] |
| Modelling Thematic Structure in Clinical Dialogues for Interpretable Depression Detection, , , , , , and , in: Text, Speech, and Dialogue. TSD 2026. Lecture Notes in Computer Science, 2026 |
[DOI] [URL] |
| Select or Rewrite: Oracle-Guided Fine-Tuning for Generative ASR Post-Editing from N-Best Hypotheses, , , , , and , Idiap-RR-04-2026 |
[URL] |
| When Consistency Becomes Bias: Interviewer Effects in Semi-Structured Clinical Interviews, , , , , , and , in: LREC 2026, Palma de Mallorca, Spain, ELRA Language Resource Association, 2026 |
[DOI] [URL] |
| Closing the Speech-Text Gap with Limited Audio for Effective Domain Adaptation in LLM-Based ASR, , , , , , , , , , , and , in: INTERSPEECH 2026, 2026 |
|
| Context Projector: Complementary Keyword and Dialogue Context Embeddings for LLM-based ASR, , , , , , , , , , and , in: INTERSPEECH 2026, 2026 |
|
| No Audio, No Transcripts, No Problem: LLM-based ASR Adaptation Using Only Domain Documents, , , , , , , and , in: IEEE SLT, 2026 |
|
| Avoiding Catastrophic Forgetting in Text-Only Adaptation of LLM-based ASR via Multi-View Text Denoising, , , , , , , , and , in: INTERSPEECH 2026, 2026 |
|
| Improving Domain-Specific ASR for Soccer Broadcast Commentary, and , Idiap-RR-03-2026 |
|
| Évaluation de la reconnaissance automatique de la parole par les grands modèles de langage génératifs, , , , , , , , and , in: EvalLLM2026 : Atelier sur l'evaluation des modeles generatifs (LLM), le RAG et challenges, 2026 |
|
| Reducing Prompt Sensitivity in LLM-based Speech Recognition Through Learnable Projection, , , , , , , , , and , in: ICASSP 2026, 2026 |
|
| Text-only adaptation in LLM-based ASR through text denoising, , , , , , , , , , and , in: ICASSP, 2026 |
|
| Leveraging Untranscribed Data for End-to-End Speech and Callsign Recognition in Air-Traffic Communication, , , , and , in: SESAR Innovation Days 2025 (https://www.sesarju.eu/SIDS2025), Eurocontrol, Bled, Slovenia, 2025 |
[URL] |
| Latent Space Factorization in LoRA, , , , and , in: 39th Conference on Neural Information Processing Systems, 2025 |
[URL] |
| Better Semi-supervised Learning for Multi-domain ASR Through Incremental Retraining and Data Filtering, , , , , , , , , , , , and , in: Interspeech 2025, Rotterdam, The Netherlands, pages 3618--3622, 2025 |
[DOI] [URL] |
| Unifying Global and Near-Context Biasing in a Single Trie Pass., , , , , , , , , , , and , in: Text, Speech, and Dialogue. TSD 2025. Lecture Notes in Computer Science, Springer, Springer, 2025 |
[DOI] [URL] |
| TokenVerse++: Towards Flexible Multitask Learning with Dynamic Task Activation, , , , , , , , , , and , in: 2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), IEEE, 2025 |
|
| Enhancing Speaker Diarization using Correlation-Based Clustering Initialization, , , and , Idiap-RR-09-2025 |
|
| Accelerating Criminal Investigations with TRACY, , , , , , , and , in: 16th EAI International Conference on Digital Forensics & Cyber Crime, 2025 |
|
| Improving ASR and Callsign Detection in Air Traffic Control Speech using Whisper Prompting, , and , Idiap-RR-04-2025 |
|
| Autocrime - open multimodal platform for combating organized crime, , , , , , , , , , , , , , , , , and , in: Forensic Science International: Digital Investigation, 54, 2025 |
[DOI] [URL] |
| Efficient Data Selection for Domain Adaptation of ASR Using Pseudo-Labels and Multi-Stage Filtering, , , , , , , , , , , , and , in: Proc. Interspeech, 2025 |
|
| TRACY Canvas: A Criminal Network Visualization Tool, , , , , and , Idiap-RR-03-2025 |
|
| Performance Evaluation of SLAM-ASR: The Good, the Bad, the Ugly, and the Way Forward, , , , , , , , , and , in: SALMA Workshop, Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Hyderabad, India, IEEE, 2025 |
[URL] |
| XLSR-Transducer: Streaming ASR for Self-Supervised Pretrained Models, , , , , , , and , in: Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Hyderabad, India, IEEE, 2025 |
[DOI] [URL] |
| Speech Data Selection for Efficient ASR Fine-Tuning using Domain Classifier and Pseudo-Label Filtering, , , , , , , , , , , and , in: 2025 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2025), 2025 |
[DOI] [URL] |
| ATCO2 corpus: A Large-Scale Dataset for Research on Automatic Speech Recognition and Natural Language Understanding of Air Traffic Control Communications, , , , , , , , , , , and , in: Journal of Data-centric Machine Learning Research, 2024 |
[URL] |
| Minimum effort adaptation of automatic speech recognition system in air traffic management, , , , , , and , in: European Journal of Transport and Infrastructure Research, 24(4 (2024)):133–153, 2025 |
[DOI] [URL] |
| TEAM SWITZERLAND SUBMISSION TO NIST SRE24 SPEAKER RECOGNITION EVALUATION, , , , , , , , , , , , and , Idiap-RR-10-2025 |
|
| Fast Streaming Transducer ASR Prototyping via Knowledge Distillation with Whisper, , , , , , , , and , in: Findings of the Association for Computational Linguistics: EMNLP 2024, pages 16747–16762, Association for Computational Linguistics (ACL), 2024 |
[DOI] [URL] |
| TokenVerse: Towards Unifying Speech and NLP Tasks via Transducer-based ASR, , , , , , , , and , in: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 20988–20995, Association for Computational Linguistics (ACL), 2024 |
[DOI] [URL] |