← All parts of this equation

Equation 4 · Part 1 · Shipping an On-Device AI Feature That Actually Works in the Field

Symbol E

E[C]=clocal+(1−p) ccloud.E[C] = c_{\mathrm{local}} + (1-p)\, c_{\mathrm{cloud}}.
EE

What this part means

E is part of the quantity the equation computes from the expression on the right.

Its job in the formula

E is part of the quantity the equation computes from the expression on the right.

The passage around this formula

That example is useful precisely because it exposes what a “local-agent-with-fallback” architecture actually costs, which is easy to gloss over in a design review. Escalation is additive, not a substitution: a request that ultimately needs the cloud still pays for the local attempt first. Write clocalc_{\mathrm{local}} for the cost — in latency, energy, or both — of attempting the request locally, ccloudc_{\mathrm{cloud}} for the marginal cost of the network round trip once escalation is triggered, and p for the probability a request is judged solvable without escalating. If the local attempt always runs before the decision to escalate is made, the expected cost per request is E[C]=clocal+(1−p) ccloudE[C] = c_{\mathrm{local}} + (1-p)\, c_{\mathrm{cloud}}. The…

Read this part in the article →

Learn the underlying idea

A variable is a named place for a value. Its letter is a local label: x can mean position in one formula and a data point in another.

Open the illustrated variables: a letter stands for a value guide →

See this notation across published equations →

Sources cited in the article section

These citations provide research context; check each source for the exact claim it supports.