Skip to content
Chad edited this page Sep 23, 2017 · 1 revision

🤔 Risks

  • Scope creep into different problem types in speech recognition.
  • Data for this problem does not exist or is too expensive.
  • We impose too many constraints on the problem (in order to solve it), and the product becomes unviable.
  • Our accuracy suffers from lack of data.
Clone this wiki locally