Reinforcement fine-tuning has instrumental enhanced the instruction-following and reasoning abilities of large language models. In this work, we explore the applications of reinforcement fine-tuning to the autoregressive transformer-based materials generative model CrystalFormer (arXiv:2403.15734) using discriminative machine learning models such as interatomic potentials and property prediction models. By optimizing reward signals-such as energy above the convex hull and material property figures of merit-reinforcement fine-tuning infuses knowledge from discriminative models into generative models. The resulting model, CrystalFormer-RL, shows enhanced stability in generated crystals and successfully discovers crystals with desirable yet conflicting material properties, such as substantial dielectric constant and band gap simultaneously. Notably, we observe that reinforcement fine-tuning enables not only the property-guided novel material design ability of generative pre-trained model but also unlocks property-driven material retrieval from the unsupervised pre-training dataset. Leveraging rewards from discriminative models to fine-tune materials generative models opens an exciting gateway to the synergies of the machine learning ecosystem for materials.
This paper presents the design and characterization of a rectangular microstrip patch antenna array optimized for operation within the Ku-band frequency range. The antenna array is impedance-matched to 50 Ohms and utilizes a microstrip line feeding mechanism for excitation. The design maintains compact dimensions, with the overall antenna occupying an area of 29.5x7 mm. The antenna structure is modelled on an R03003 substrate material, featuring a dielectric constant of 3, a low-loss tangent of 0.0009, and a thickness of 1.574 mm. The substrate is backed by a conducting ground plane, and the array consists of six radiating patch elements positioned on top. Evaluation of the designed antenna array reveals a resonant frequency of 18GHz, with a -10 dB impedance bandwidth extending over 700MHz. The antenna demonstrates a high gain of 7.51dBi, making it well-suited for applications in 5G and future communication systems. Its compact form factor, cost-effectiveness, and broad impedance and radiation coverage further underscore its potential in these domains.
Dielectrics are crucial for technologies like flash memory, CPUs, photovoltaics, and capacitors, but public data on these materials are scarce, restricting research and development. Existing machine learning models have focused on predicting scalar polycrystalline dielectric constants, neglecting the directional nature of dielectric tensors essential for material design. This study leverages multi-rank equivariant structural embeddings from a universal neural network potential to enhance predictions of dielectric tensors. We develop an equivariant readout decoder to predict total, electronic, and ionic dielectric tensors while preserving O(3) equivariance, and benchmark its performance against state-of-the-art algorithms. Virtual screening of thermodynamically stable materials from Materials Project for two discovery tasks, high-dielectric and highly anisotropic materials, identifies promising candidates including Cs2Ti(WO4)3 (band gap $E_g=2.93 \mathrm{eV}$, dielectric constant $\varepsilon=180.90$) and CsZrCuSe3 (anisotropic ratio $\alpha_r = 121.89$). The results demonstrate our model's accuracy in predicting dielectric tensors and its potential for discovering novel dielectric materials.
The recent proliferation of personal wireless communication devices is driving the need for multi-band frequency selective components including multiplexers and dual-band filters. This paper presents a simple technique for transforming a single-band bandpass filter (BPF) into a dual-band BPF. A second order (two-pole) single-band bandpass filter was chosen for this research, giving rise to a fourth order (four-pole) dual-band bandpass filter after the proposed filter transformation. Both filters were then implemented using the compact U-shaped microstrip resonator for improved device miniaturization. The proposed work features a centre frequency of 1.4 GHz for the single-band bandpass filter, with a span of 3.4% fractional bandwidth. The dual-band bandpass filter operates at 1.35 and 1.45 GHz. The design implementation employs the commercially available Rogers RT/Duroid 6010LM substrate, having a dissipation factor of 0.0023, dielectric constant of 10.7, diel thickness (h) of 1.27 mm, and top/bottom cladding of 35 microns. The results reported for the theoretical and practical designs show good agreement and improved performance when compared to similar research works in literature. The practical responses of the prototype dual-band BPF indicate a good return loss of better than 18 dB across both bands, and an insertion loss of better than 0.1 dB.
Microstrip three-way (that is, 4.8 dB) integrated filtering power divider (FPD) is presented in this paper. The proposed FPD evenly distributes an input power signal into three equal output signals. The design incorporates balanced signal power division, and filtering technology for the removal of unwanted frequency elements and aimed at enhancing signal quality and efficiency in the radiofrequency (RF) front-end of communication systems. Microstrip folded-arms square open-loop resonator (FASOLR) is employed in the design implementation to achieve compact size. The proposed FPD features a 2.6 GHz centre frequency, with a 0.03 fractional bandwidth. The implementation is carried out on Rogers RT/Duroid 6010LM substrate with a dielectric constant of 10.7, a thickness of 1.27 mm and a loss tangent of 0.0023. The good agreement between the theoretical and practical results verifies the effectiveness of the FPD in delivering equal power outputs at the three output ports, and at the same time filtering out unwanted frequencies. The practical results of the prototype FPD indicate a good return loss of better than 15.5 dB and an insertion loss of better than 4.77+0.34 dB. The design prototype achieved compact size of 0.31 {\lambda}g x 0.18 {\lambda}g. {\lambda}g is the guided wavelength for the microstrip line impedance at the centre frequency of the 3-way equal filtering power divider.
We present a novel ship wake simulation system for generating S-band Synthetic Aperture Radar (SAR) images, and demonstrate the use of such imagery for the classification of ships based on their wake signatures via a deep learning approach. Ship wakes are modeled through the linear superposition of wind-induced sea elevation and the Kelvin wakes model of a moving ship. Our SAR imaging simulation takes into account frequency-dependent radar parameters, i.e., the complex dielectric constant ($\varepsilon$) and the relaxation rate ($\mu$) of seawater. The former was determined through the Debye model while the latter was estimated for S-band SAR based on pre-existing values for the L, C, and X-bands. The results show good agreement between simulated and real imagery upon visual inspection. The results of implementing different training strategies are also reported, showcasing a notable improvement in accuracy of classifier achieved by integrating real and simulated SAR images during the training.
Materials with high-dielectric constant easily polarize under external electric fields, allowing them to perform essential functions in many modern electronic devices. Their practical utility is determined by two conflicting properties: high dielectric constants tend to occur in materials with narrow band gaps, limiting the operating voltage before dielectric breakdown. We present a high-throughput workflow that combines element substitution, ML pre-screening, ab initio simulation and human expert intuition to efficiently explore the vast space of unknown materials for potential dielectrics, leading to the synthesis and characterization of two novel dielectric materials, CsTaTeO6 and Bi2Zr2O7. Our key idea is to deploy ML in a multi-objective optimization setting with concave Pareto front. While usually considered more challenging than single-objective optimization, we argue and show preliminary evidence that the $1/x$-correlation between band gap and permittivity in fact makes the task more amenable to ML methods by allowing separate models for band gap and permittivity to each operate in regions of good training support while still predicting materials of exceptional merit. To our knowledge, this is the first instance of successful ML-guided multi-objective materials optimization achieving experimental synthesis and characterization. CsTaTeO6 is a structure generated via element substitution not present in our reference data sources, thus exemplifying successful de-novo materials design. Meanwhile, we report the first high-purity synthesis and dielectric characterization of Bi2Zr2O7 with a band gap of 2.27 eV and a permittivity of 20.5, meeting all target metrics of our multi-objective search.
The conventional surface reflection method has been widely used to measure the asphalt pavement layer dielectric constant using ground-penetrating radar (GPR). This method may be inaccurate for in-service pavement thickness estimation with dielectric constant variation through the depth, which could be addressed using the extended common mid-point method (XCMP) with air-coupled GPR antennas. However, the factors affecting the XCMP method on thickness prediction accuracy haven't been studied. Manual acquisition of key factors is required, which hinders its real-time applications. This study investigates the affecting factors and develops a modified XCMP method to allow automatic thickness prediction of in-service asphalt pavement with non-uniform dielectric properties through depth. A sensitivity analysis was performed, necessitating the accurate estimation of time of flights (TOFs) from antenna pairs. A modified XCMP method based on edge detection was proposed to allow real-time TOFs estimation, then dielectric constant and thickness predictions. Field tests using a multi-channel GPR system were performed for validation. Both the surface reflection and XCMP setups were conducted. Results show that the modified XCMP method is recommended with a mean prediction error of 1.86%, which is more accurate than the surface reflection method (5.73%).
Resistive random-access memory (RRAM) is a promising candidate for next-generation memory devices due to its high speed, low power consumption, and excellent scalability. Metal oxides are commonly used as the oxide layer in RRAM devices due to their high dielectric constant and stability. However, to further improve the performance of RRAM devices, recent research has focused on integrating artificial intelligence (AI). AI can be used to optimize the performance of RRAM devices, while RRAM can also power AI as a hardware accelerator and in neuromorphic computing. This review paper provides an overview of the combination of metal oxides-based RRAM and AI, highlighting recent advances in these two directions. We discuss the use of AI to improve the performance of RRAM devices and the use of RRAM to power AI. Additionally, we address key challenges in the field and provide insights into future research directions
Noise and uncertainty are usually the enemy of machine learning, noise in training data leads to uncertainty and inaccuracy in the predictions. However, we develop a machine learning architecture that extracts crucial information out of the noise itself to improve the predictions. The phenomenology computes and then utilizes uncertainty in one target variable to predict a second target variable. We apply this formalism to PbZr$_{0.7}$Sn$_{0.3}$O$_{3}$ crystal, using the uncertainty in dielectric constant to extrapolate heat capacity, correctly predicting a phase transition that otherwise cannot be extrapolated. For the second example -- single-particle diffraction of droplets -- we utilize the particle count together with its uncertainty to extrapolate the ground truth diffraction amplitude, delivering better predictions than when we utilize only the particle count. Our generic formalism enables the exploitation of uncertainty in machine learning, which has a broad range of applications in the physical sciences and beyond.
This paper compiles measured complex dielectric constants of benign and hazardous materials used for developing algorithms for personnel scanners at airports at 30 GHz. The materials are grouped into broad classifications of potential threat by mapping the materials to the complex dielectric constant plane.
The multilayer structure is a promising technique used to minimize the size of planar microstrip filters. In the flexible design and incorporation of other microwave components, multilayer band-pass filter results in better and enhanced dimensions. This paper introduces a microstrip fifth-generation (5G) low-frequency band of 2.52-2.65 GHz using a parallel-coupled line (PCL) Bandpass filter and multilayer (ML) hairpin bandpass filter. The targeted four-pole resonator has a center frequency of 2.58 GHz with a bandwidth of 130 MHz. The filters are designed with a 0.1 dB passband ripple with a Chebyshev response. The hairpin-line offers compact filter design structures. Theoretically, they can be obtained by bending half-wavelength resonator resonators with parallel couplings into a "U" shape. The proposed configuration of the parallel-coupled line resonator is used to design the ML band-pass filter. The FR4 substrate with a dielectric constant ({\epsilon}r) of 4.3 and 1.6 mm thickness was used. A comparative analysis between the simulated insertion loss and the reflection coefficient of substrates RO3003 and FR4 was performed to validate the efficiency of the proposed filter design. Simulation of PCL filter is accomplished using computer simulation technology (CST) and an advanced design system (ADS) software. The PCL Bandpass filter was experimentally validated and a total tally between simulation results and measured results were achieved demonstrating a well-measured reflection coefficient. The simulated results obtained by the hairpin ML bandpass filter show that the circuit performs well in terms of Scattering(S) parameters and the filter size is significantly reduced.
To describe international trade flows, we propose the coulomb force formulation, in which the magnitude of the charge represents gross domestic product (GDP) and the distance between countries is the bilateral distance, the product of spatial distance and "dielectric constant," rather than the spatial distance as used in the gravitation model, allowing it to be time dependent. The "dielectric constant" is influenced by factors such as warfare, transportation disruptions, trade agreements, social, geography, politics, culture, and others. The GDP and distance power parameters were estimated using data from high-GDP countries' export-import transactions. We also developed a trade strength distribution equation that fits World Bank data reasonably well over a decade.
This paper presents the comparison of various neural networks and algorithms based on accuracy, quickness, and consistency for antenna modelling. Using MATLAB Nntool, 22 different combinations of networks and training algorithms are used to predict the dimensions of a rectangular microstrip antenna using dielectric constant, height of substrate, and frequency of oper-ation as input. Comparison and characterization of networks is done based on accuracy, mean square error, and training time. Algorithms, on the other hand, are analyzed by their accuracy, speed, reliability, and smoothness in the training process. Finally, these results are analyzed, and recommendations are made for each neural network and algorithm based on uses, advantages, and disadvantages. For example, it is observed that Reduced Radial Bias network is the most accurate network and Scaled Conjugate Gradient is the most reliable algorithm for electromagnetic modelling. This paper will help a researcher find the optimum network and algorithm directly without doing time-taking experimentation.
Solving physics problems for which we know the equations, boundary conditions and symmetries can be done by deep learning. The constraints can be either imposed as terms in a loss function or used to formulate a neural ansatz. In the present case study, we calculate the induced field inside and outside a dielectric cube placed in a uniform electric field, wherein the dielectric mismatch at edges and corners of the cube makes accurate calculations numerically challenging. The electric potential is expressed as an ansatz incorporating neural networks with known leading order behaviors and symmetries and the Laplace's equation is then solved with boundary conditions at the dielectric interface by minimizing a loss function. The loss function ensures that both Laplace's equation and boundary conditions are satisfied everywhere inside a large solution domain. We study how the electric potential inside and outside a quasi-cubic particle evolves through a sequence of shapes from a sphere to a cube. The neural network being differentiable, it is straightforward to calculate the electric field over the whole domain, the induced surface charge distribution and the polarizability. The neural network being retentive, one can efficiently follow how the field changes upon particle's shape or dielectric constant by iterating from any previously converged solution. The present work's objective is two-fold, first to show how an a priori knowledge can be incorporated into neural networks to achieve efficient learning and second to apply the method and study how the induced field and polarizability change when a dielectric particle progressively changes its shape from a sphere to a cube.
Biomolecular condensates such as membraneless organelles, underpinned by liquid-liquid phase separation (LLPS), are important for physiological function, with electrostatics -- among other interaction types -- being a prominent force in their assembly. Charge interactions of intrinsically disordered proteins (IDPs) and other biomolecules are sensitive to the aqueous dielectric environment. Because the relative permittivity of protein is significantly lower than that of water, the interior of an IDP condensate is a relatively low-dielectric regime, which, aside from its possible functional effects on client molecules, should facilitate stronger electrostatic interactions among the scaffold IDPs. To gain insight into this LLPS-induced dielectric heterogeneity, addressing in particular whether a low-dielectric condensed phase entails more favorable LLPS than that posited by assuming IDP electrostatic interactions are uniformly modulated by the higher dielectric constant of the pure solvent, we consider a simplified multiple-chain model of polyampholytes immersed in explicit solvents that are either polarizable or possess a permanent dipole. Notably, simulated phase behaviors of these systems exhibit only minor to moderate differences from those obtained using implicit-solvent models with a uniform relative permittivity equals to that of pure solvent. Buttressed by theoretical treatments developed here using random phase approximation and polymer field-theoretic simulations, these observations indicate a partial compensation of effects between favorable solvent-mediated interactions among the polyampholytes in the condensed phase and favorable polyampholyte-solvent interactions in the dilute phase, often netting only a minor enhancement of overall LLPS propensity from the very dielectric heterogeneity that arises from the LLPS itself. Further ramifications of this principle are discussed.
A fast and accurate predictive tool for polymer properties is demanding and will pave the way to iterative inverse design. In this work, we apply graph convolutional neural networks (GCNN) to predict the dielectric constant and energy bandgap of polymers. Using density functional theory (DFT) calculated properties as the ground truth, GCNN can achieve remarkable agreement with DFT results. Moreover, we show that GCNN outperforms other machine learning algorithms. Our work proves that GCNN relies only on morphological data of polymers and removes the requirement for complicated hand-crafted descriptors, while still offering accuracy in fast predictions.