Google

The function $$r(\mathbf{x}, \mathbf{y}, \bar{\mathbf{y}}) = -1$$ assigns a constant negative value. This function's arguments are the vectors $$\mathbf{x}$$ and $$\mathbf{y}$$, and the average vector $$\bar{\mathbf{y}}$$. In machine learning, a value of $$-1$$ often signifies a negative reward, a penalty, or indicates a negative sample.

Formula for a Negative Reward Function $$r(\mathbf{x}, \mathbf{y}, \bar{\mathbf{y}})$$

Based on the provided scenario, what is the primary signal the model receives during training, and what is the likely impact on the model's ability to distinguish between high-quality and low-quality story segments? Explain your reasoning.

Evaluating a Reward Function for Story Generation

In a machine learning system, a penalty is assigned to every generated output segment using the function `r(x, y, ȳ) = -1`, where `x` is the input, `y` is the generated output, and `ȳ` is the average of all possible outputs. If the system generates an output segment, what is the value of the penalty it receives according to this function?

Learn Before

Related