In this paper, a novel system utilizing a hybrid relay-intelligent reflecting surface (HR-IRS) to boost the security performance of directional modulation (DM) is established. In particular, the malicious attacker works in full-duplex (FD) mode and it will eavesdrop on confidential message (CM) as well as send malicious jamming. To maximize the secrecy rate (SR), a joint problem of optimizing the receive beamforming, transmit beamforming, power allocation (PA) factor, and phase shift matrix (PSM) of HR-IRS is formulated. Since the optimization problem is un-convex and the variables are coupled with each other, we address this problem by iteratively optimizing these variables. First, the receive beamforming is designed based on the generalized Rayleigh-Ritz theorem. Then, the transmit beamforming and PA factor are optimized via Dinkelbach’s Transform and successive convex approximation methods. And for PSM, two strategies, called separate optimization of PSM (SO-PSM) and joint optimization of PSM (JO-PSM), are proposed. Thus, two iterative schemes are proposed accordingly, namely maximizing SR based on SO-PSM (Max-SR-SOP) and maximizing SR based on JO-PSM (Max-SR-JOP). The former has a better performance and the latter has a lower complexity. Simulation results show that given a sufficient power budget of HR-IRS, the proposed Max-SR-SOP and Max-SR-JOP can enable HR-IRS-aided DM network to obtain a higher SR than that aided by passive IRS.

Proteins power a vast array of functional processes in living cells. The capability to create new proteins with designed structures and functions would thus enable the engineering of cellular behavior and development of protein-based therapeutics and materials. Structure-based protein design aims to find structures that are designable (can be realized by a protein sequence), novel (have dissimilar geometry from natural proteins), and diverse (span a wide range of geometries). While advances in protein structure prediction have made it possible to predict structures of novel protein sequences, the combinatorially large space of sequences and structures limits the practicality of search-based methods. Generative models provide a compelling alternative, by implicitly learning the low-dimensional structure of complex data distributions. Here, we leverage recent advances in denoising diffusion probabilistic models and equivariant neural networks to develop Genie, a generative model of protein structures that performs discrete-time diffusion using a cloud of oriented reference frames in 3D space. Through in silico evaluations, we demonstrate that Genie generates protein backbones that are more designable, novel, and diverse than existing models. This indicates that Genie is capturing key aspects of the distribution of protein structure space and facilitates protein design with high success rates. Code for generating new proteins and training new versions of Genie is available at https://github.com/aqlaboratory/genie.