Google

A function, denoted as $s$, that takes a sequence of $k$ variables, $\bar{y}_1, \ldots, \bar{y}_k$, as its input. The notation $\bar{y}_i$ represents the $i$-th variable in the sequence, where the bar (overline) can signify a specific property, such as being an average or a predicted value.

Function of a Sequence of Overlined Variables

In a system that generates a sequence of 'k' items, let each ȳᵢ (for i from 1 to k) represent a summary of the system's state after generating the i-th item. A scoring function is then defined as s(ȳ₁, ..., ȳₖ), which outputs a single numerical value. Based on this structure, what is the primary role of the function 's'?

An AI system generates a 5-step sequence of cooking instructions. To ensure the final recipe is logical, the system uses a function, s(ȳ₁, ȳ₂, ȳ₃, ȳ₄, ȳ₅), to calculate an overall quality score for the entire 5-step sequence. In this function, each ȳᵢ represents a summary of the recipe's state after instruction 'i' has been generated. Based on this scenario, propose a specific, plausible meaning for both the overall function 's' and for the individual variable 'ȳ₃'.

Interpreting a Sequence Evaluation Function

A text generation system is evaluating two possible three-word sequences. For each sequence, it has calculated a confidence score for each word, denoted by $\bar{y}_i$ for the $i$-th word. The system uses a function, $s(\bar{y}_1, \bar{y}_2, \bar{y}_3) = \sum_{i=1}^{3} \bar{y}_i$, to calculate a total score for the entire sequence. A higher total score is considered better. Given the scores below, which sequence should the system prefer, and what is its total score? Explain your reasoning.

Learn Before

Related