Kernel-elastic autoencoder for molecular design.

We introduce the kernel-elastic autoencoder (KAE), a self-supervised generative model based on the transformer architecture with enhanced performance for molecular design. KAE employs two innovative loss functions: modified maximum mean discrepancy (m-MMD) and weighted reconstruction (LWCEL). The m-MMD loss has significantly improved the generative performance of KAE when compared to using the traditional Kullback-Leibler loss of VAE, or standard maximum mean discrepancy. Including the weighted reconstruction loss LWCEL, KAE achieves valid generation and accurate reconstruction at the same time, allowing for generative behavior that is intermediate between VAE and autoencoder not available in existing generative approaches. Further advancements in KAE include its integration with conditional generation, setting a new state-of-the-art benchmark in constrained optimizations. Moreover, KAE has demonstrated its capability to generate molecules with favorable binding affinities in docking applications, as evidenced by AutoDock Vina and Glide scores, outperforming all existing candidates from the training dataset. Beyond molecular design, KAE holds promise to solve problems by generation across a broad spectrum of applications.

Photonic crystal fiber-based surface plasmon resonance sensor with selective analyte channels and graphene-silver deposited core.

We propose a surface plasmon resonance (SPR) sensor based on photonic crystal fiber (PCF) with selectively filled analyte channels. Silver is used as the plasmonic material to accurately detect the analytes and is coated with a thin graphene layer to prevent oxidation. The liquid-filled cores are placed near to the metallic channel for easy excitation of free electrons to produce surface plasmon waves (SPWs). Surface plasmons along the metal surface are excited with a leaky Gaussian-like core guided mode. Numerical investigations of the fiber's properties and sensing performance are performed using the finite element method (FEM). The proposed sensor shows maximum amplitude sensitivity of 418 Refractive Index Units (RIU-1) with resolution as high as 2.4 × 10(-5) RIU. Using the wavelength interrogation method, a maximum refractive index (RI) sensitivity of 3000 nm/RIU in the sensing range of 1.46-1.49 is achieved. The proposed sensor is suitable for detecting various high RI chemicals, biochemical and organic chemical analytes. Additionally, the effects of fiber structural parameters on the properties of plasmonic excitation are investigated and optimized for sensing performance as well as reducing the sensor's footprint.

Double Brillouin frequency shift through circulation of odd-order Stokes signal.

We demonstrate a simple configuration for generating a double Brillouin frequency shift through the circulation of an odd-order Brillouin Stokes signal. It is operated based on cascaded Brillouin scattering in single-mode optical fibers that behave as the Brillouin gain media. A four-port circulator is incorporated into the setup to circulate the odd-order Brillouin Stokes signal in the fiber. It thus initiates a higher order Brillouin Stokes signal, which is double Brillouin frequency downshifted from the input signal. For the 5 km long fiber, the Brillouin pump power at 23 mW gives a clean output spectrum with 30 dB sideband suppression ratio. The output signal is 0.174 nm or approximately 21.7 GHz downshifted from the input signal.