Reinforce-Aligner: Reinforcement alignment search for robust end-to-end text-to-speech

Engineering

• Internals 100%
• Experimental Result 50%
• Processing Step 50%
• Feedforward 50%
• Search Method 50%
• Reinforcement Learning 50%

Keyphrases

• End-to-end Text-to-speech 100%
• Trained Agent 20%
• Cumulative Reward 20%
• Text-to-speech Synthesis 20%
• Duration Prediction 20%

Computer Science

• Speech Synthesis 20%
• Optimal Duration 20%