Home › Hard problems › Edge AI and serving cost

Edge AI and serving cost: AI on small devices and cheaper answers

Trusting a model running on someone else's device

Solo: possibleOne person can test a narrow slice of it.

The approved modelchecked and signedinstalledsomeone else's device?swapped ortampered with?The answerlooks finechanged on the way??you, far away:how can you tell?
Once a model runs on someone else's device, it is hard to prove it is still the approved model, and that its answers were not changed.

Why it is still open

How do you prove a device runs the approved, untampered model and that its answers weren't changed?

An experiment one person can run in 8 weeks

Build a demo that signs each answer with the model's fingerprint and test which tampering it catches.

Verdict: Solo: possible

A narrow proof of concept; full solutions need hardware support.

The solo test behind the verdict

  1. Can it be measured? There is a number that shows you did better, such as an error rate, a catch rate or a cost per answer.
  2. Is it cheap to try? It runs on a laptop, free open models and open data, or under about ₹50,000 of rented GPU time.
  3. Is there a public yardstick? A benchmark or dataset others use, so your result can be compared and believed.
  4. Would a small win get noticed? A lab, a funder or a customer would care about a 10% improvement, not only a full solution.

Four yeses means Solo: strong. Two or three means Solo: possible. Fewer means Follow, don't build.

PreviousEnergy per answer

All 79 hard problems · Working on this, or know of a result we missed? basant@cleandataengineering.ai