About Workshop
This workshop focuses on advanced data analytics techniques for graphene-based Surface Acoustic Wave (SAW) gas sensors. Participants will learn signal preprocessing methods like noise filtering and drift correction, followed by anomaly detection using autoencoders for identifying sensor faults. The workshop also covers transfer learning for adapting models to new gas analytes, providing hands-on experience in enhancing sensor performance and developing real-time, adaptive sensor systems for various applications.
Aim
The workshop aims to teach advanced data analytics for graphene-based SAW gas sensors, focusing on signal preprocessing, anomaly detection with autoencoders, and transfer learning for adapting models to new gas analytes.
Structure
📅 Day 1 – Signal Preprocessing for SAW Gas Sensors
- Overview of Surface Acoustic Wave (SAW) Gas Sensors
- Signal characteristics of graphene-based sensing devices
- Preprocessing needs: Noise filtering, drift correction, baseline alignment
- Algorithms for signal smoothing, peak alignment, and normalization
- Discussion on preprocessing pipelines in current sensor research
📅 Day 2 – Anomaly Detection Using Autoencoders
- Introduction to unsupervised anomaly detection in sensor systems
- Autoencoders: Concept, architecture, and training workflow
- Sensor-specific anomaly types: drift, spikes, signal loss
- Research trends: Denoising autoencoders, reconstruction error analysis
- Application of autoencoders in fault detection and early warning systems
📅 Day 3 – Hands-On: Transfer Learning & Adaptation to New Analytes
- Quick recap of signal preprocessing and autoencoder workflow
- Introduction to transfer learning for cross-analyte generalization
- ⚙️ Hands-On Activities:
- Load and visualize preprocessed SAW sensor data
- Use a pretrained autoencoder for anomaly detection
- Fine-tune model on data from a new gas analyte
- Evaluate model adaptation performance
Important Dates
Registration Ends
8 PM
Workshop Dates
2025-09-28
9 PM
9 PM
What You Will Gain
Outcomes
-
Expertise in signal preprocessing and anomaly detection.
-
Hands-on experience with autoencoders and transfer learning.
-
Proficiency in enhancing sensor performance through data analytics.
-
Readiness for careers in sensor technology, data science, and machine learning.
Who Should Attend
This workshop is intended for researchers, engineers, and students in sensor technology, data science, and materials science. It is ideal for those interested in gas sensor systems, signal processing, and machine learning applications. A basic understanding of machine learning concepts and programming is recommended but not required.