All-optical computing based on convolutional neural networks

The rapid development of information technology has fueled an ever-increasing demand for ultrafast and ultralow-en-ergy-consumption computing.Existing computing instruments are pre-dominantly electronic processors,which use elec-trons as information carriers and possess von Neumann architecture featured by physical separation of storage and pro-cessing.The scaling of computing speed is limited not only by data transfer between memory and processing units,but also by RC delay associated with integrated circuits.Moreover,excessive heating due to Ohmic losses is becoming a severe bottleneck for both speed and power consumption scaling.Using photons as information carriers is a promising alternative.Owing to the weak third-order optical nonlinearity of conventional materials,building integrated photonic com-puting chips under traditional von Neumann architecture has been a challenge.Here,we report a new all-optical comput-ing framework to realize ultrafast and ultralow-energy-consumption all-optical computing based on convolutional neural networks.The device is constructed from cascaded silicon Y-shaped waveguides with side-coupled silicon waveguide segments which we termed“weight modulators”to enable complete phase and amplitude control in each waveguide branch.The generic device concept can be used for equation solving,multifunctional logic operations as well as many other mathematical operations.Multiple computing functions including transcendental equation solvers,multifarious logic gate operators,and half-adders were experimentally demonstrated to validate the all-optical computing performances.The time-of-flight of light through the network structure corresponds to an ultrafast computing time of the order of several picoseconds with an ultralow energy consumption of dozens of femtojoules per bit.Our approach can be further expan-ded to fulfill other complex computing tasks based on non-von Neumann architectures and thus paves a new way for on-chip all-optical computing.

Structural transformation and energy analysis for pile-up dislocations at triple junction of grain boundary

An energy model for the structure transformation of pile-ups of grain boundary dislocations(GBD)at the triple-junction of the grain boundary of ultrafine-grain materials was proposed.The energy of the pile-up of the GBD in the system was calculated by the energy model,the critical geometric and mechanical conditions for the structure transformation of head dislocation of the pile-up were analyzed,and the influence of the number density of the dislocations and the angle between Burgers vectors of two decomposed dislocations on the transformation mode of head dislocation was discussed.The results show when the GBD is accumulated at triple junction,the head dislocation of the GBD is decomposed into two Burgers vectors of these dislocations unless the angle between the two vectors is less than 90°,and the increase of applied external stress can reduce the energy barrier of the dislocation decomposition.The mechanism that the ultrafine-grained metal material has both high strength and plasticity owing to the structure transformation of the pile-up of the GBD at the triple junction of the grain boundary is revealed.

Ultra-wideband surface plasmonic bandpass filter with extremely wide upper-band rejection

An ultra-wideband bandpass filter(BPF)with a wide out-of-band rejection based on a surface plasmonic waveguide(SPW)slotline with ring grooves is designed and analyzed.A paired microstrip-to-slotline transition is designed for quasiTEM to TM mode conversion by using a microstrip line with a circular pad and the slotline with the same circular slot.The mode conversion between the TM and the surface plasmon polariton(SPP)mode is realized by using a gradient slotline with ring grooves and an impedance matching technique.The upper cut-off frequencies of the passband can be adjusted by using these proposed SPP units,while the lower frequencies of the passband are created by using the microstrip-toslotline transitions to give an ultra-wideband BPF.The dispersion curves of SPP units,electric field distribution,and the transmission spectra of the proposed ultra-wideband bandpass filter are all calculated and analyzed by the finite-difference time-domain(FDTD)method.The simulated results show that the presented filter has good performance including a wide3-dB bandwidth of 149%from 0.57 GHz to 3.93 GHz,an extremely wide 40-dB upper-band rejection from 4.2 GHz to18.5 GHz,and low loss and high selectivity in the passband.To prove the design validity,a prototype of the BPF has been manufactured and measured,showing a reasonable agreement with simulation results.The unique features of the proposed BPF may make it applicable for integrated circuit and plasmonic devices in microwave or THz frequency ranges.

A novel polarization converter based on the band-stop frequency selective surface

A dual-passband single-polarized converter based on the band-stop frequency selective surface(FSS)with a low radar cross-section(RCS)is designed in this article.The unit cell of the proposed converter is formed by a polarization layer attached to the band-stop frequency selective surface.The simulation results reveal that the co-polarization reflection coefficients below-10 d B are achieved in 3.82–13.64 GHz with a 112.4%fractional bandwidth(the ratio of the signal bandwidth to the central frequency).Meanwhile,a polarization conversion band is realized from 8.14 GHz to 9.27 GHz with a polarization conversion ratio which is over 80%.Moreover,the 1 d B transmission window is obtained in two nonadjacent bands of 3.42–7.02 GHz and 10.04–13.91 GHz corresponding to the relative bandwidths of 68.9%and 32.3%,respectively.Furthermore,the radar cross-section of the designed structure can be reduced in the wideband from 2.28 GHz to 14 GHz,and the 10 d B RCS reduction in the range of 4.10–13.35 GHz is achieved.In addition,the equivalent circuit model of this converter is established,and the simulation results of the Advanced Design System(ADS)match well with those of CST Microwave Studio(CST).The archetype of the designed converter is manufactured and measured.The experiment results match the simulation results well,which proves the reliability of the simulation results.

Advanced all-optical classification using orbitalangular-momentum-encoded diffractive networks

As a successful case of combining deep learning with photonics,the research on optical machine learning has recently undergone rapid development.Among various optical classification frameworks,diffractive networks have been shown to have unique advantages in all-optical reasoning.As an important property of light,the orbital angular momentum(OAM)of light shows orthogonality and mode-infinity,which can enhance the ability of parallel classification in information processing.However,there have been few all-optical diffractive networks under the OAM mode encoding.Here,we report a strategy of OAM-encoded diffractive deep neural network(OAM-encoded D2NN)that encodes the spatial information of objects into the OAM spectrum of the diffracted light to perform all-optical object classification.We demonstrated three different OAM-encoded D2NNs to realize(1)single detector OAM-encoded D2NN for single task classification,(2)single detector OAM-encoded D2NN for multitask classification,and(3)multidetector OAM-encoded D2NN for repeatable multitask classification.We provide a feasible way to improve the performance of all-optical object classification and open up promising research directions for D2NN by proposing OAMencoded D2NN.

Phosphorylated NFS1 weakens oxaliplatin-based chemosensitivity of colorectal cancer by preventing PANoptosis

Metabolic enzymes have an indispensable role in metabolic reprogramming,and their aberrant expression or activity has been associated with chemosensitivity.Hence,targeting metabolic enzymes remains an attractive approach for treating tumors.