Questions: Garbled Circuits

5 questions to test your understanding

Score: 0 / 5
Question 1 Short Answer

In a garbled gate, the garbler encrypts four entries: one for each combination of input labels. The evaluator has one label per input wire and can decrypt exactly one entry. Why can't the evaluator try all four entries to learn extra information?

Think about your answer, then reveal below.
Question 2 Multiple Choice

The 'free XOR' optimization allows XOR gates to be evaluated without any ciphertext or communication. How does this work?

AXOR gates are removed from the circuit during preprocessing
BThe garbler chooses a global random offset R. For every wire, the two labels differ by XOR with R (label_1 = label_0 XOR R). XOR gate output labels are computed as the XOR of input labels: if label_a and label_b are inputs, the output label is label_a XOR label_b, which automatically encodes the XOR of the underlying bits due to the algebraic relationship with R
CXOR gates use a special hash function that requires no encryption
DThe evaluator already knows XOR results from the input labels without computation
Question 3 True / False

A garbled circuit can only be evaluated once. If the evaluator could evaluate it on two different inputs, they could learn the garbler's input.

TTrue
FFalse
Question 4 Multiple Choice

Half-gates (Zahur-Rosulek-Evans 2015) reduce the cost of a garbled AND gate from 3 ciphertexts to 2. What is the conceptual idea?

AHalf-gates split the circuit in half, garbling each half independently
BAn AND gate is decomposed into two 'half-gates': one where the garbler knows one input bit, and one where the evaluator knows one input bit. Each half-gate requires only one ciphertext (using the free XOR technique internally), totaling 2 ciphertexts per AND gate
CHalf-gates use elliptic curve operations to compress the gate representation
DThe optimization skips half the AND gates by approximating the circuit
Question 5 True / False

Garbled circuits achieve security against semi-honest adversaries directly. Achieving security against malicious adversaries requires additional techniques.

TTrue
FFalse