| | 29 | |
| | 30 | == Week2 Tutorials == |
| | 31 | - Python tutorial: https://www.w3schools.com/python/ |
| | 32 | - How to run Python code: https://www.knowledgehut.com/blog/programming/run-python-scripts |
| | 33 | - Jupyter notebook tutorial: https://www.dataquest.io/blog/jupyter-notebook-tutorial/ |
| | 34 | - Video tutorial (Optional): Neural Networks and Deep Learning: https://www.coursera.org/learn/neural-networks-deep-learning |
| | 35 | |
| | 36 | == Week 3 Activities == |
| | 37 | - Setup the TensorFlow environment and run the Python code sample for a basic neural network. |
| | 38 | - Read the paper “X-Vectors: Robust DNN Embeddings for Speaker Recognition” (IEEE ICASSP 2018). |
| | 39 | |
| | 40 | |
| | 41 | == Week 4 Activities == |
| | 42 | - Understand the speaker recognition system (X-Vector) and time-delay neural network. |
| | 43 | - Learn MFCC feature and extract the MFCC feature using TensorFlow. |
| | 44 | |
| | 45 | == Week 5 Activities == |
| | 46 | - Study the Python code samples for X-Vector and implement X-Vector. |
| | 47 | - Learn how to use X-Vector and feed the extracted MFCC features into X-Vector. |
| | 48 | |
| | 49 | == Week 6 Activities == |
| | 50 | - Read the paper “Practical Adversarial Attacks Against Speaker Recognition Systems” (HotMobile 2020). |
| | 51 | - Understand the untargeted and targeted attacks against speaker recognition systems. |
| | 52 | |
| | 53 | == Week 7 Activities == |
| | 54 | - Understand the Fast Gradient Sign Method (FGSM) for the untargeted attack. |
| | 55 | - Study the code samples for Practical Adversarial Attacks Against Speaker Recognition Systems. |
| | 56 | |
| | 57 | == Week 8 Activities == |
| | 58 | - Develop an untargeted attack that can generate adversarial samples based on the sample code and tutorial. |
| | 59 | - Evaluate the performance of the adversarial samples on the voice assistant system (X-Vector). |
| | 60 | |
| | 61 | == Week 9 Activities == |
| | 62 | - Debug and fine-tune the untargeted adversarial machine learning algorithm to achieve better performance. |
| | 63 | - Develop a targeted attack that can spoof the X-Vector and misclassify the input audio signals as targeted speakers. |
| | 64 | |
| | 65 | == Week 10 Activities == |
| | 66 | - Debug and fine-tune the developed targeted attack method. |
| | 67 | - If time allows, simulate the room impulse response (RIR) and integrate it into the developed attack methods. |
| | 68 | |
| | 69 | == Week 11 Activities == |
| | 70 | - Fine-tune the developed targeted and untargeted attack methods. |
| | 71 | - Summarize and prepare for the open house presentation. |
