liveBook · Manning

Coarse-grained array

As explained in Sutskever's List

A coarse-grained array is a lower-resolution grid whose entries summarize local neighborhoods of a finer grid.

A coarse-grained array turns a detailed grid into a lower-resolution view of its larger patterns. In the cream-and-coffee mixing example, each entry averages a small neighborhood, then is quantized into labels such as coffee, cream, or mixed. This removes microscopic speckles while preserving features such as the boundary between the two fluids. section 8.1.1

The array represents a macrostate for estimating its description length: it is compressed with a standard compressor, while entropy is estimated from the full, fine-grained state. The two views focus on different scales. section 8.1.1

Figure 8.2 Illustration of the construction of the coarse-grained array using an example grain size of 3. The values of the shaded cells on the left are averaged to produce the value of the shaded cell on the right. Used with permission granted by the lead author (Scott Aaronson).
figure

The figure illustrates how averaging a neighborhood with grain size 3 produces one coarse-grained cell. section 8.1.1

How Sutskever's List covers it

How deep it goes
Explains neighborhood averaging and quantization, then contrasts the resulting array with fine-grained compression for entropy estimation.
Examples
Uses a cream-and-coffee grid, including a grain-size-3 illustration, to show neighborhood averages becoming coarse cells.
What you can do after
Understand how local averaging and quantization preserve large-scale mixing patterns while discarding microscopic randomness.
How it connects
Connects macrostate representation to entropy estimation, Kolmogorov complexity, and data compression.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Sutskever's List. The section links point to the parts of the book it draws on.