Aggregate difficult to divide

• When you want to divide the elements of a set by type and then perform an aggregate f on them, but the cost of the division is high
• If f is a function that can be realized by repeated binary operations
  • Examples: sum, min
• Aggregation can be done without dividing the data in advance

python

for x in S:
    subgroup[typeof(x)].append(x)
for t, xs in subgroup:
    result[t] = f(xs)

python

for x in S:
    result[typeof(x)].update(x)

• Related to attention mechanism?

---

This page is auto-translated from /nishio/分割困難な集計 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.