Workshop Program

Each lecture will be 60 minutes long, followed by a 15-minute Q&A session. A lecture can be presentation oriented and hands-on (30% presentation, 70% hands-on). Participants needs to install Git, SPTK, diffsptk, and Nkululeko on their own computers (Linux, OSX, Windows WSL). The workshop will be conducted in English but questions can be asked in Japanese. Coffee and tea will be provided during the breaks but lunch will not be provided.

Day 1: September 16, 2025

9:30am Opening Remarks + Git
Bagus Tris Atmaja (NAIST)
10:45am Coffee & Tea Break
11:00am Machine Learning Speaker Characteristics
Felix Burkhardt (audEERING, TU Berlin)
12:15pm Lunch break
1:30pm Fundamental of Speech Signal Processing
Takenori Yoshimura (Nitech)
2:45pm Coffee & Tea Break
3:00pm Speech Processing with SPTK and diffsptk
Takenori Yoshimura (Nitech)

Day 2: September 17, 2025

9:30am Overview of Nkululeko
Felix Burkhardt (audEERING, TU Berlin)
10:45am Coffee & Tea Break
11:00am Speech data preparation with AudFormat and CSV
Felix Burkhardt (audEERING, TU Berlin)
12:15pm Lunch break
1:30pm Data balancing, Scaling, and Optimization
Bagus Tris Atmaja (NAIST)
2:45pm Coffee & Tea Break
3:00pm Acoustic Features and Classifiers
Felix Burkhardt (audEERING, TU Berlin)

Day 3: September 18, 2025

9:30am Data splitting, segmentation, and augmentation
Felix Burkhardt (audEERING, TU Berlin)
10:45am Coffee & Tea Break
11:00am Pathological Voice Detection and Ensemble Learning
Bagus Tris Atmaja (NAIST)
12:15pm Lunch break
1:30pm Multi database and cross database evaluation
Felix Burkhardt (audEERING, TU Berlin)
2:45pm Coffee & Tea Break
3:00pm Pathological Speech and Speech Disorder
Dhany Arifianto (ITS)

Day 4: September 19, 2025

10:45am Coffee & Tea Break
11:00am Voice-based COVID and TB detection
Dhany Arifianto (ITS)
12:15pm Lunch break
11:00pm Nkululeko for Bias Detection and Mitigation
Felix Burkhardt (audEERING, TU Berlin)
11:30pm Finding Important Features for Dementia
Bagus Tris Atmaja