1: \begin{abstract}

2: 

3: Bidirectional relaying, where a relay helps two user nodes

4: to exchange messages has been an active area of recent research. In

5: the compute-and-forward strategy for bidirectional relaying, the relay

6: computes a function of the two messages using the naturally-occurring

7: sum of symbols simultaneously transmitted by user nodes in a Gaussian

8: Multiple Access Channel (MAC), and the computed function value is

9: forwarded to the user nodes in an ensuing broadcast phase. In this

10: work, we consider the Gaussian MAC in bidirectional relaying with the

11: messages taking values in a finite Abelian group and the relay

12: computing the sum within the group under an additional secrecy

13: constraint for protection against a honest but curious relay. The

14: secrecy constraint is that, while the relay should decode the group

15: sum, the individual message of each user should be perfectly secure

16: from the relay. We exploit the symbol addition that occurs in a

17: Gaussian MAC to design explicit modulations at the user nodes that

18: achieve independence between the received symbols at the relay

19: and each of the two individual transmitted messages under an average

20: transmit power constraint. We provide a lattice coding strategy for

21: reliable computation of the group sum at the relay with perfect

22: secrecy, and study rate versus average power trade-offs in the

23: large-dimension regime. Our results for secure

24: compute-and-forward are significant because we achieve perfect

25: security with finite average transmit power, and this has been done

26: using a novel approach involving Fourier-analytic tools.

27: \end{abstract}

28: