Teacher forcing

How it is used

Used in sequence models. During generation the model may instead receive its own previous outputs, creating different conditions.

Example sentences

1. With teacher forcing, the decoder receives the correct previous word during training.

   The next prediction is conditioned on the reference sequence rather than the model's last guess.

2. The model performed well with teacher forcing but struggled when generating from its own earlier predictions.

   Errors in generated history can create conditions that clean training histories did not expose.

Common collocations

• use teacher forcing
• teacher forcing during training

Related terms

• trainTo teach a model by showing it many examples and repeatedly adjusting its parameters so that its outputs get closer to the correct answers.
• neural networkA model made of many connected units arranged in layers, loosely inspired by the brain. Each connection has a weight that is adjusted during training.

More terms: Machine Learning and Data Processing

• artificial intelligence
• machine learning
• model
• dataset
• training data
• label
• supervised learning
• deep learning
• loss function
• overfitting
• test set
• accuracy