- Multi-party computation (MPC) allows multiple parties to jointly compute a function while preserving the privacy of their inputs. This can be done using techniques like secret sharing so that no single party learns the other's private inputs. - Differential privacy adds precisely calibrated noise to statistical outputs to prevent an adversary from reconstructing individuals' private inputs. It provides output privacy and is used by companies like Apple. - Secure multi-party computation, homomorphic encryption, and zero-knowledge proofs are cryptographic techniques for input privacy in statistical analysis and machine learning while preserving data privacy.