Enhanced mutual information neural estimators for optical fiber communication.

The accurate estimation of mutual information (MI) plays a vital role in understanding channel capacity and optimizing the performance of optical communications. While MI computations for the additive white Gaussian noise (AWGN) channel are well-established, they fall short when dealing with the challenges posed by nonlinear optical fiber channels due to an unknown channel model. For the first time, to our knowledge, this Letter introduces a mutual information neural estimator (MINE) for MI estimation in optical fiber communications. We propose an enhanced MINE (E-MINE), achieved by enlarging the training batch size to improve estimation accuracy and stability. Our findings reveal that the E-MINE achieves highly accurate estimations in the AWGN channel and maintains strong consistency with symbol-by-symbol MI estimations, comparable to Monte Carlo (MC) methods based on a Gaussian distribution in long-haul optical fiber channels. Furthermore, with multi-symbol estimation, the E-MINE yields a 0.16 bits/4D-symbol improvement in our experiments. We anticipate that our findings will drive further research in the field, opening new possibilities for enhancing communication systems design and performance using deep learning techniques.

High-fidelity sub-petabit-per-second self-homodyne fronthaul using broadband electro-optic combs.

With the exponential growth in data density and user ends of wireless networks, fronthaul is tasked with supporting aggregate bandwidths exceeding thousands of gigahertz while accommodating high-order modulation formats. However, it must address the bandwidth and noise limitations imposed by optical links and devices in a cost-efficient manner. Here we demonstrate a high-fidelity fronthaul system enabled by self-homodyne digital-analog radio-over-fiber superchannels, using a broadband electro-optic comb and uncoupled multicore fiber. This self-homodyne superchannel architecture not only offers capacity boosting but also supports carrier-recovery-free reception. Our approach achieves a record-breaking 15,000 GHz aggregated wireless bandwidth, corresponding to a 0.879 Pb/s common public radio interface (CPRI) equivalent data rate. Higher-order formats up to 1,048,576 quadrature-amplitude-modulated (QAM) are showcased at a 100 Tb/s class data rate. Furthermore, we employ a packaged on-chip electro-optic comb as the sole optical source to reduce the cost, supporting a data rate of 100.5 Tb/s with the 1024-QAM format. These demonstrations propel fronthaul into the era of Pb/s-level capacity and exhibit the promising potential of integrated-photonics implementation, pushing the boundaries to new heights in terms of capacity, fidelity, and cost.

Learnable digital signal processing: a new benchmark of linearity compensation for optical fiber communications.

The surge in interest regarding the next generation of optical fiber transmission has stimulated the development of digital signal processing (DSP) schemes that are highly cost-effective with both high performance and low complexity. As benchmarks for nonlinear compensation methods, however, traditional DSP designed with block-by-block modules for linear compensations, could exhibit residual linear effects after compensation, limiting the nonlinear compensation performance. Here we propose a high-efficient design thought for DSP based on the learnable perspectivity, called learnable DSP (LDSP). LDSP reuses the traditional DSP modules, regarding the whole DSP as a deep learning framework and optimizing the DSP parameters adaptively based on backpropagation algorithm from a global scale. This method not only establishes new standards in linear DSP performance but also serves as a critical benchmark for nonlinear DSP designs. In comparison to traditional DSP with hyperparameter optimization, a notable enhancement of approximately 1.21 dB in the Q factor for 400 Gb/s signal after 1600 km fiber transmission is experimentally demonstrated by combining LDSP and perturbation-based nonlinear compensation algorithm. Benefiting from the learnable model, LDSP can learn the best configuration adaptively with low complexity, reducing dependence on initial parameters. The proposed approach implements a symbol-rate DSP with a small bit error rate (BER) cost in exchange for a 48% complexity reduction compared to the conventional 2 samples/symbol processing. We believe that LDSP represents a new and highly efficient paradigm for DSP design, which is poised to attract considerable attention across various domains of optical communications.

FPGA implementation of power-lite Volterra-inspired neural network equalizer in 180-Gb/s net bitrate IMDD short-reach optical system.

A white-box power-lite Volterra-inspired neural network (VINN) equalizer is proposed to solve the problem of complexity discontinuity in a Volterra nonlinear equalizer (VNLE). By adjusting the granularity of the solution space, it conserves computational resources while maintaining nonlinear compensation capability. The performance of VINN is verified on a field-programmable gate array (FPGA) in a short-reach intensity modulation and direct detection (IMDD) system, and a 240-Gb/s real-time signal processing rate is achieved. Under the 25% overhead soft-decision forward error correction (SD-FEC) bit error rate (BER) threshold, we realize a record net rate of up to 180 Gb/s based on the FPGA.

miR-429-3p mediates memory decline by targeting MKP-1 to reduce surface GluA1-containing AMPA receptors in a mouse model of Alzheimer's disease.

Alzheimer's disease (AD) is a leading cause of dementia in the elderly. Mitogen-activated protein kinase phosphatase 1 (MKP-1) plays a neuroprotective role in AD. However, the molecular mechanisms underlying the effects of MKP-1 on AD have not been extensively studied. MicroRNAs (miRNAs) regulate gene expression at the post-transcriptional level, thereby repressing mRNA translation. Here, we reported that the microRNA-429-3p (miR-429-3p) was significantly increased in the brain of APP23/PS45 AD model mice and N2AAPP AD model cells. We further found that miR-429-3p could downregulate MKP-1 expression by directly binding to its 3'-untranslated region (3' UTR). Inhibition of miR-429-3p by its antagomir (A-miR-429) restored the expression of MKP-1 to a control level and consequently reduced the amyloidogenic processing of APP and Aß accumulation. More importantly, intranasal administration of A-miR-429 successfully ameliorated the deficits of hippocampal CA1 long-term potentiation and spatial learning and memory in AD model mice by suppressing extracellular signal-regulated kinase (ERK1/2)-mediated GluA1 hyperphosphorylation at Ser831 site, thereby increasing the surface expression of GluA1-containing α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPARs). Together, these results demonstrate that inhibiting miR-429-3p to upregulate MKP-1 effectively improves cognitive and synaptic functions in AD model mice, suggesting that miR-429/MKP-1 pathway may be a novel therapeutic target for AD treatment.

PCSK6 exacerbates Alzheimer's disease pathogenesis by promoting MT5-MMP maturation.

Proprotein convertase subtilisin/kexin type 6 (PCSK6) is a calcium-dependent serine proteinase that regulates the proteolytic activity of various precursor proteins and facilitates protein maturation. Dysregulation of PCSK6 expression or function has been implicated in several pathological processes including nervous system diseases. However, whether and how PCSK6 is involved in the pathogenesis of Alzheimer's disease (AD) remains unclear. In this study, we reported that the expression of PCSK6 was significantly increased in the brain tissues of postmortem AD patients and APP23/PS45 transgenic AD model mice, as well as N2AAPP cells. Genetic knockdown of PCSK6 reduced amyloidogenic processing of APP in N2AAPP cells by suppressing the activation of membrane-type 5-matrix metalloproteinase (MT5-MMP), referred to as η-secretase. We further found that PCSK6 cleaved and activated MT5-MMP by recognizing the RRRNKR sequence in its N-terminal propeptide domain in N2A cells. The mutation or knockout of this cleavage motif prevented PCSK6 from interacting with MT5-MMP and performing cleavage. Importantly, genetic knockdown of PCSK6 with adeno-associated virus (AAV) reduced Aß production and ameliorated hippocampal long-term potentiation (LTP) and long-term spatial learning and memory in APP23/PS45 transgenic mice. Taken together, these results demonstrate that genetic knockdown of PCSK6 effectively alleviate AD-related pathology and cognitive impairments by inactivating MT5-MMP, highlighting its potential as a novel therapeutic target for AD treatment.

Data-driven inverse design of mode-locked fiber lasers.

The diverse applications of mode-locked fiber lasers (MLFLs) raise various demands on the output of the laser, including the pulse duration, energy, and shape. Simulation is an excellent method to guide the design and construction of an MLFL for on-demand laser output. Traditional simulation of an MLFL uses the split-step Fourier method (SSFM) to solve the nonlinear Schrödinger (NLS) equation, which suffers from high computational complexity. As a result, the inverse design of MLFLs via the traditional SSFM-based simulation method relies on the design experience. Here, a completely data-driven approach for the inverse design of MLFLs is proposed, which significantly reduces the computational complexity and achieves a fast automatic inverse design of MLFLs. We utilize a recurrent neural network to realize fast and accurate MLFL modeling, then the desired cavity settings meeting the output demands are searched via a deep-reinforcement learning algorithm. The results prove that the data-driven method enables the accurate inverse design of an MLFL to produce a preset target femtosecond pulse with a certain duration and pulse energy. In addition, the cavity settings generating soliton molecules with different target separations can also be located via the data-driven inverse design. With the GPU acceleration, the time consumption of the data-driven inverse design of an MLFL is less than 1.3 hours. The proposed data-driven approach is applicable to guide the inverse design of an MLFL to meet the different demands of various applications.

Fangchinoline alleviates cognitive impairments through enhancing autophagy and mitigating oxidative stress in Alzheimer's disease models.

Introduction: Alzheimer's disease (AD) is a debilitating, progressive, neurodegenerative disorder characterized by the deposition of amyloid-ß (Aß) peptides and subsequent oxidative stress, resulting in a cascade of cytotoxic effects. Fangchinoline (Fan), a bisbenzylisoquinoline alkaloid isolated from traditional Chinese herb Stephania tetrandra S. Moorec, has been reported to possess multiple potent biological activities, including anti-inflammatory and antioxidant properties. However, the potential neuroprotective efficacy of Fan against AD remains unknown. Methods: N2AAPP cells, the mouse neuroblastoma N2A cells stably transfected with human Swedish mutant APP695, were served as an in vitro AD model. A mouse model of AD was constructed by microinjection of Aß1-42 peptides into lateral ventricle of WT mice. The neuroprotective effects of Fan on AD were investigated through a combination of Western blot analysis, immunoprecipitation and behavioral assessments. Results and discussion: It was found that Fan effectively attenuated the amyloidogenic processing of APP by augmenting autophagy and subsequently fostering lysosomal degradation of BACE1 in N2AAPP cells, as reflected by the decrease in P62 levels, concomitant with the increase in Beclin-1 and LC3-II levels. More importantly, Fan significantly ameliorated cognitive impairment in an Aß1-42-induced mouse model of AD via the induction of autophagy and the inhibition of oxidative stress, as evidenced by an increase in antioxidants including glutathione reductase (GR), total antioxidant capacity (T-AOC), nuclear factor erythroid-2-related factor 2 (Nrf2), heme oxygenase-1 (HO-1), and superoxide dismutase-1 (SOD-1) and a decrease in pro-oxidants including hydrogen peroxide (H2O2) and inducible nitric oxide synthase (i-NOS), coupled with a reduction in apoptosis marker, cleaved caspase-3. Taken together, our study demonstrate that Fan ameliorates cognitive dysfunction through promoting autophagy and mitigating oxidative stress, making it a potential therapeutic agent for AD.

Hypoxia-sensing VGLL4 promotes LDHA-driven lactate production to ameliorate neuronal dysfunction in a cellular model relevant to Alzheimer's disease.

Alzheimer's disease (AD) is a neurodegenerative disease where abnormal amyloidogenic processing of amyloid-ß precursor protein (APP) occurs and has been linked to neuronal dysfunction. Hypometabolism of glucose in the brain can lead to synaptic loss and neuronal death, which in turn exacerbates energy deficiency and amyloid-ß peptide (Aß) accumulation. Lactate produced by anaerobic glycolysis serves as an energy substrate supporting neuronal function and facilitating neuronal repair. Vestigial-like family member 4 (VGLL4) has been recognized as a key regulator of the hypoxia-sensing pathway. However, the role of VGLL4 in AD remains unexplored. Here, we reported that the expression of VGLL4 protein was significantly decreased in the brain tissue of AD model mice and AD model cells. We further found that overexpression of VGLL4 reduced APP amyloidogenic processing and ameliorated neuronal synaptic damage. Notably, we identified a compromised hypoxia-sensitive capability of LDHA regulated by VGLL4 in the context of AD. Upregulation of VGLL4 increased the response of LDHA to hypoxia and enhanced the expression levels of LDHA and lactate by inhibiting the ubiquitination and degradation of LDHA. Furthermore, the inhibition of lactate production by using sodium oxamate, an inhibitor of LDHA, suppressed the neuroprotective function of VGLL4 by increasing APP amyloidogenic processing. Taken together, our findings demonstrate that VGLL4 exerts a neuroprotective effect by upregulating LDHA expression and consequently promoting lactate production. Thus, this study suggests that VGLL4 may be a novel player involved in molecular mechanisms relevant for ameliorating neurodegeneration.

Signal-to-noise ratio degradation analysis for optoelectronic feedback-based chaotic optical communication systems.

Chaotic optical communication encrypts transmitted signals through physical noise; this ensures high security while causing a certain decrease in the signal-to-noise ratio (SNR). Thus, it is necessary to analyze the SNR degradation of decrypted signals after chaotic encryption and the minimum requirements for the SNR of the fiber channel to meet the required bit error rate (BER) performance. Accordingly, an SNR model of decrypted signals for optoelectronic feedback-based chaotic optical communication systems is proposed. Under different channel SNRs, the SNR degradation of 40 Gbit/s phase chaos and intensity chaos models is investigated by simulation and experiment, respectively, with a 15 GHz wideband chaotic carrier. Comparing decrypted signals with original signals, the simulation results show that there is a 2.9 dB SNR degradation for both intensity chaos and phase chaos. Further, in experiments, SNR degradation from 4.5 dB to 5.6 dB, with various channel SNRs for intensity chaos, is analyzed, while there is an SNR degradation from 7.1 dB to 8.3 dB for phase chaos. The simulation and experimental results provide guidance for long-distance transmission chaotic optical communication systems.

Inhibiting NLRP3 inflammasome signaling pathway promotes neurological recovery following hypoxic-ischemic brain damage by increasing p97-mediated surface GluA1-containing AMPA receptors.

BACKGROUND: The nucleotide-binding oligomeric domain (NOD)-like receptor protein 3 (NLRP3) inflammasome is believed to be a key mediator of neuroinflammation and subsequent secondary brain injury induced by ischemic stroke. However, the role and underlying mechanism of the NLRP3 inflammasome in neonates with hypoxic-ischemic encephalopathy (HIE) are still unclear. METHODS: The protein expressions of the NLRP3 inflammasome including NLRP3, cysteinyl aspartate specific proteinase-1 (caspase-1) and interleukin-1ß (IL-1ß), the α-amino-3-hydroxy-5-methyl-4-isoxazole-propionicacid receptor (AMPAR) subunit, and the ATPase valosin-containing protein (VCP/p97), were determined by Western blotting. The interaction between p97 and AMPA glutamate receptor 1 (GluA1) was determined by co-immunoprecipitation. The histopathological level of hypoxic-ischemic brain damage (HIBD) was determined by triphenyltetrazolium chloride (TTC) staining. Polymerase chain reaction (PCR) and Western blotting were used to confirm the genotype of the knockout mice. Motor functions, including myodynamia and coordination, were evaluated by using grasping and rotarod tests. Hippocampus-dependent spatial cognitive function was measured by using the Morris-water maze (MWM). RESULTS: We reported that the NLRP3 inflammasome signaling pathway, such as NLRP3, caspase-1 and IL-1ß, was activated in rats with HIBD and oxygen-glucose deprivation (OGD)-treated cultured primary neurons. Further studies showed that the protein level of the AMPAR GluA1 subunit on the hippocampal postsynaptic membrane was significantly decreased in rats with HIBD, and it could be restored to control levels after treatment with the specific caspase-1 inhibitor AC-YVAD-CMK. Similarly, in vitro studies showed that OGD reduced GluA1 protein levels on the plasma membrane in cultured primary neurons, whereas AC-YVAD-CMK treatment restored this reduction. Importantly, we showed that OGD treatment obviously enhanced the interaction between p97 and GluA1, while AC-YVAD-CMK treatment promoted the dissociation of p97 from the GluA1 complex and consequently facilitated the localization of GluA1 on the plasma membrane of cultured primary neurons. Finally, we reported that the deficits in motor function, learning and memory in animals with HIBD, were ameliorated by pharmacological intervention or genetic ablation of caspase-1. CONCLUSION: Inhibiting the NLRP3 inflammasome signaling pathway promotes neurological recovery in animals with HIBD by increasing p97-mediated surface GluA1 expression, thereby providing new insight into HIE therapy.

Low-complexity amplitude-division irregular QAM formats for short-reach unamplified coherent optical systems.

A low-complexity scheme is proposed to realize irregular uniform quadrature amplitude modulation (QAM) formats with Gray mapping, which are named amplitude-division irregular QAM (AD-Ir-QAM) formats. Compared to conventional probabilistic shaping (PS) with the Maxwell-Boltzmann (MB) distribution (PS-MB), irregular QAM formats show a smaller peak-to-average power (PAPR) and achieve a better performance in systems with the peak-power constraint. Compared with irregular QAM formats realized by PS (PS-Ir-QAM), AD-Ir-QAM formats realize a more flexible rate adaptation and have a lower implementation complexity. Experimental results obtained in an unamplified coherent optical system show that, at a generalized mutual information (GMI) of 4.5 bits/2D-symbol, AD-Ir-100QAM achieves gains of 2.1 and 0.5 dB in the power budget over PS-MB-100QAM and PS-Ir-100QAM, respectively.

Digital pre-distortion using a Gauss-Newton-based direct learning architecture for coherent optical transmitters.

Digital pre-distortion (DPD) is a powerful technique to mitigate transmitter nonlinear distortion in optical transmissions. In this Letter, the identification of DPD coefficients based on the direct learning architecture (DLA) using the Gauss-Newton (GN) method is applied in optical communications for the first time. To the best of our knowledge, this is the first time that the DLA has been realized without training an auxiliary neural network to mitigate optical transmitter nonlinear distortion. We describe the principle of the DLA using the GN method and compare the DLA with the indirect learning architecture (ILA) that uses the least-square (LS) method. Extensive numerical and experimental results indicate that the GN-based DLA is superior to the LS-based ILA, especially in a low signal-to-noise ratio scenario.

Spectral broadening scheme for suppressing SBS effects based on time-domain optimized chirp-like signals.

We propose a novel (to our knowledge) driving scheme to suppress the stimulated Brillouin scattering (SBS) effect in master oscillator power amplification (MOPA) systems based on an external high-order phase modulation. Since seed sources with the linear chirp can uniformly broaden the SBS gain spectrum with a high SBS threshold, a chirp-like signal was designed by applying further editing and processing to the piecewise parabolic signal. Compared with the traditional piecewise parabolic signal, the chirp-like signal has similar linear chirp characteristics and can reduce the driving power and sampling rate requirements, enabling more efficient spectral spreading. The SBS threshold model is constructed theoretically based on the three-wave coupling equation. The spectrum modulated by the chirp-like signal is compared with the flat-top and Gaussian spectra in terms of the SBS threshold and the bandwidth-distribution normalized threshold, and a considerable improvement is demonstrated. Meanwhile, the experimental validation is carried out in a watt-class amplifier based on the MOPA structure. At a 3 dB bandwidth of ∼10 GHz, the SBS threshold of the seed source modulated by the chirp-like signal is improved by 35% compared to the flat-top spectrum and 18% compared to the Gaussian spectrum, respectively, and the normalized threshold is also the highest among them. Our study shows that the SBS suppression effect is not only related to the power distribution of the spectrum but also can be improved by the time domain design, which provides a new idea for analyzing and improving the SBS threshold of narrow-linewidth fiber lasers.

Modeling of a multi-parameter chaotic optoelectronic oscillator based on the Fourier neural operator.

A model construction scheme of chaotic optoelectronic oscillator (OEO) based on the Fourier neural operator (FNO) is proposed. Different from the conventional methods, we learn the nonlinear dynamics of OEO (actual components) in a data-driven way, expecting to obtain a multi-parameter OEO model for generating chaotic carrier with high-efficiency and low-cost. FNO is a deep learning architecture which utilizes neural network as a parameter structure to learn the trajectory of the family of equations from training data. With the assistance of FNO, the nonlinear dynamics of OEO characterized by differential delay equation can be modeled easily. In this work, the maximal Lyapunov exponent is applied to judge whether these time series have chaotic behavior, and the Pearson correlation coefficient (PCC) is introduced to evaluate the modeling performance. Compare with long and short-term memory (LSTM), FNO is not only superior to LSTM in modeling accuracy, but also requires less training data. Subsequently, we analyze the modeling performance of FNO under different feedback gains and time delays. Both numerical and experimental results show that the PCC can be greater than 0.99 in the case of low feedback gain. Next, we further analyze the influence of different system oscillation frequencies, and the generalization ability of FNO is also analyzed.

Low-complexity full-field ultrafast nonlinear dynamics prediction by a convolutional feature separation modeling method.

The modeling and prediction of the ultrafast nonlinear dynamics in the optical fiber are essential for the studies of laser design, experimental optimization, and other fundamental applications. The traditional propagation modeling method based on the nonlinear Schrödinger equation (NLSE) has long been regarded as extremely time-consuming, especially for designing and optimizing experiments. The recurrent neural network (RNN) has been implemented as an accurate intensity prediction tool with reduced complexity and good generalization capability. However, the complexity of long grid input points and the flexibility of neural network structure should be further optimized for broader applications. Here, we propose a convolutional feature separation modeling method to predict full-field ultrafast nonlinear dynamics with low complexity and strong generalization ability with high accuracy, where the linear effects are firstly modeled by NLSE-derived methods, then a convolutional deep learning method is implemented for nonlinearity modeling. With this method, the temporal relevance of nonlinear effects is substantially shortened, and the parameters and scale of neural networks can be greatly reduced. The running time achieves a 94% reduction versus NLSE and an 87% reduction versus RNN without accuracy deterioration. In addition, the input pulse conditions, including grid point numbers, durations, peak powers, and propagation distance, can be generalized accurately during the predicting process. The results represent a remarkable improvement in ultrafast nonlinear dynamics prediction and this work also provides novel perspectives of the feature separation modeling method for quickly and flexibly studying the nonlinear characteristics in other fields.

Coherent digital-analog radio-over-fiber (DA-RoF) system with a CPRI-equivalent data rate beyond 1 Tb/s for fronthaul.

We propose and experimentally demonstrate a coherent digital-analog radio-over-fiber (DA-RoF) system and achieve the transmission of Tb/s common public radio interface (CPRI)-equivalent data rate for fronthaul. The proposed coherent DA-RoF system includes DA-RoF modulation, demodulation and DA-RoF compatible coherent digital signal processing (DSP) blocks. A theoretical analysis of the DA-RoF scheme together with parameter optimization is accomplished as well. In the experiment, a 25 Gbaud DA-RoF signal with 1 Tb/s CPRI-equivalent data rate is transmitted in the system, satisfying the error vector magnitude (EVM) requirement for 256-quadrature amplitude modulation (QAM) signal transmission. With the symbol rate reduced to 10 Gbaud, an EVM below 2.5% is achieved, which meets the requirement for 1024-QAM transmission. The experimental results show that the coherent DA-RoF system is a promising solution for future fronthaul.

Optical filtering impairment monitoring based on model fusion for optical networks.

Reconfigurable optical add-and-drop multiplexer (ROADM) is a key element in optical networks. As several ROADMs are cascaded over long paths, the penalty induced by ROADM has become non-negligible due to the tight optical filtering. In this case, for efficient and reliable network planning and operation, accurate monitoring of optical filtering penalty is very important. In this paper, we propose a real-time optical filtering monitoring scheme based on model fusion. We combine an analytical model based on the digital communications theory of band-limited channels with linear equalization and a data-driven model implemented using artificial neural network (ANN). The scheme can achieve high accuracy and interpretability. Moreover, since the input features are extracted from configuration parameters and receiver digital signal processing (DSP), no additional devices are needed, which is attractive for practical deployment. Extensive simulations and experiments are conducted to investigate the performance of the scheme, and the results show the superior performance.

Asymmetric point-to-multipoint coherent architecture with a frequency aliasing recovery algorithm for cost-constraint short-reach access networks.

An asymmetric point-to-multipoint (PTMP) coherent architecture combined with a frequency aliasing recovery (FAR) algorithm is proposed for cost-constraint short-reach access networks. In this architecture, the uplink transmitters are simplified significantly with the uplink dual-polarization four-level pulse amplitude modulation (DP-PAM4) and downlink DP quadrature phase shift keying (DP-QPSK) asymmetric transmission design. Digital to analog converters (DACs) and radio frequency (RF) drivers are reduced by half, and in-phase and quadrature modulators (IQMs) are replaced by Mach-Zehnder modulators (MZMs), saving four MZ interferometers (MZIs). Furthermore, based on the asymmetric architecture, the FAR algorithm can recover signals from frequency aliasing caused by frequency offset (FO), even when half of the signal spectrum is aliased. This algorithm enables the asymmetric architecture to narrow down guard bands between subcarriers or even overlap the subcarriers, saving the receiver bandwidth at the aggregation/hub side. The performance of the asymmetric uplink DP-PAM4 transmission with the FAR algorithm is evaluated in both simualations and experiments. The effects of laser linewidths and IQ skew on the performance of the FAR algorithm are also analyzed. Simulation results show the algorithm can recover 16 Gbaud and 32 Gbaud signal from 8 GHz and 16 GHz aliasing, respectively. In the experiments with 10 km fiber transmissions, the FAR algorithm can recover 10 Gbaud signals from 5 GHz frequency aliasing, saving about 20.83% of the total receiver bandwidth in a 2-subcarrier system.

Coherent chaotic optical communication of 30 Gb/s over 340-km fiber transmission via deep learning.

Chaotic optical communication has attracted much attention as a hardware encryption method in the physical layer. Limited by the requirements of chaotic hardware synchronization, fiber transmission impairments are restrictedly compensated in the optical domain. There has been little experimental demonstration of high-speed and long-distance chaotic optical communication systems. Here, we propose a method to overcome such limitations. Using a deep-learning model to realize chaotic synchronization in the digital domain, fiber transmission impairments can be compensated by digital-signal processing (DSP) algorithms with coherent detection. A successful transmission of 30 Gb/s quadrature phase-shift keying messages hidden in a 15 GHz wideband chaotic optical carrier was experimentally demonstrated over a 340-km fiber link. Meanwhile, the chaotic receiver can be significantly simplified without compromising security. The proposed method is a possible way to promote the practical application of chaotic optical communications.