arXiv subjects

Lihua Li

Publications and source records attributed to Lihua Li.

15 recordsLinked to original sources

arXiv2025

Two-Timescale Learning for Pilot-Free ISAC Systems

A pilot-free integrated sensing and communication (ISAC) system is investigated, in which phase-modulated continuous wave (PMCW) and non-orthogonal multiple access (NOMA) waveforms are co-designed to achieve simultaneous target sensing and data transmission. To enhance effective data throughput (i.e., Goodput) in PMCW-NOMA ISAC systems, we propose a deep learning-based receiver architecture, termed two-timescale Transformer (T3former), which leverages a Transformer architecture to perform joint channel estimation and multi-user signal detection without the need for dedicated pilot signals. By treating the deterministic structure of the PMCW waveform as an implicit pilot, the proposed T3former eliminates the overhead associated with traditional pilot-based methods. The proposed T3former processes the received PMCW-NOMA signals on two distinct timescales, where a fine-grained attention mechanism captures local features across the fast-time dimension, while a coarse-grained mechanism aggregates global spatio-temporal dependencies of the slow-time dimension. Numerical results demonstrate that the proposed T3former significantly outperforms traditional successive interference cancellation (SIC) receivers, which avoids inherent error propagation in SIC. Specifically, the proposed T3former achieves a substantially lower bit error rate and a higher Goodput, approaching the theoretical maximum capacity of a pilot-free system.

cs.IT↗

arXiv2025

Inductive-Associative Meta-learning Pipeline with Human Cognitive Patterns for Unseen Drug-Target Interaction Prediction

Significant differences in protein structures hinder the generalization of existing drug-target interaction (DTI) models, which often rely heavily on pre-learned binding principles or detailed annotations. In contrast, BioBridge designs an Inductive-Associative pipeline inspired by the workflow of scientists who base their accumulated expertise on drawing insights into novel drug-target pairs from weakly related references. BioBridge predicts novel drug-target interactions using limited sequence data, incorporating multi-level encoders with adversarial training to accumulate transferable binding principles. On these principles basis, BioBridge employs a dynamic prototype meta-learning framework to associate insights from weakly related annotations, enabling robust predictions for previously unseen drug-target pairs. Extensive experiments demonstrate that BioBridge surpasses existing models, especially for unseen proteins. Notably, when only homologous protein binding data is available, BioBridge proves effective for virtual screening of the epidermal growth factor receptor and adenosine receptor, underscoring its potential in drug discovery.

cs.LG↗

arXiv2020

DML-GANR: Deep Metric Learning With Generative Adversarial Network Regularization for High Spatial Resolution Remote Sensing Image Retrieval

With a small number of labeled samples for training, it can save considerable manpower and material resources, especially when the amount of high spatial resolution remote sensing images (HSR-RSIs) increases considerably. However, many deep models face the problem of overfitting when using a small number of labeled samples. This might degrade HSRRSI retrieval accuracy. Aiming at obtaining more accurate HSR-RSI retrieval performance with small training samples, we develop a deep metric learning approach with generative adversarial network regularization (DML-GANR) for HSR-RSI retrieval. The DML-GANR starts from a high-level feature extraction (HFE) to extract high-level features, which includes convolutional layers and fully connected (FC) layers. Each of the FC layers is constructed by deep metric learning (DML) to maximize the interclass variations and minimize the intraclass variations. The generative adversarial network (GAN) is adopted to mitigate the overfitting problem and validate the qualities of extracted high-level features. DML-GANR is optimized through a customized approach, and the optimal parameters are obtained. The experimental results on the three data sets demonstrate the superior performance of DML-GANR over state-of-the-art techniques in HSR-RSI retrieval.

cs.CV↗

arXiv2020

SLCRF: Subspace Learning with Conditional Random Field for Hyperspectral Image Classification

Subspace learning (SL) plays an important role in hyperspectral image (HSI) classification, since it can provide an effective solution to reduce the redundant information in the image pixels of HSIs. Previous works about SL aim to improve the accuracy of HSI recognition. Using a large number of labeled samples, related methods can train the parameters of the proposed solutions to obtain better representations of HSI pixels. However, the data instances may not be sufficient enough to learn a precise model for HSI classification in real applications. Moreover, it is well-known that it takes much time, labor and human expertise to label HSI images. To avoid the aforementioned problems, a novel SL method that includes the probability assumption called subspace learning with conditional random field (SLCRF) is developed. In SLCRF, first, the 3D convolutional autoencoder (3DCAE) is introduced to remove the redundant information in HSI pixels. In addition, the relationships are also constructed using the spectral-spatial information among the adjacent pixels. Then, the conditional random field (CRF) framework can be constructed and further embedded into the HSI SL procedure with the semi-supervised approach. Through the linearized alternating direction method termed LADMAP, the objective function of SLCRF is optimized using a defined iterative algorithm. The proposed method is comprehensively evaluated using the challenging public HSI datasets. We can achieve stateof-the-art performance using these HSI sets.

cs.CV↗

arXiv2020

Secrecy Analysis of Ambient Backscatter NOMA Systems under I/Q Imbalance

We investigate the reliability and security of the ambient backscatter (AmBC) non-orthogonal multiple access (NOMA) systems, where the source aims to communication with two NOMA users in the presence of an eavesdropper. We consider a more practical case that nodes and backscatter device (BD) suffer from in-phase and quadrature-phase imbalance (IQI). More specifically, exact analytical expressions for the outage probability (OP) and the intercept probability (IP) are derived in closedform. Moreover, the asymptotic behaviors and corresponding diversity orders for the OP are discussed. Numerical results show that: 1) Although IQI reduces the reliability, it can enhance the security. 2) Compared with the traditional orthogonal multiple access (OMA) system, the AmBC-NOMA system can obtain better reliability when the signal-to-noise (SNR) ratio is low; 3) There are error floors for the OP because of the reflection coefficient \b{eta} .

cs.IT↗

arXiv2020

Diffusion of Chiral Janus Particles in Convection Rolls

The diffusion of an artificial active particle in a two-dimensional periodic pattern of stationary convection cells is investigated by means of extensive numerical simulations. In the limit of large Péclet numbers, i.e., for self-propulsion speeds below a certain depinning threshold and weak roto-translational fluctuations, the particle undergoes asymptotic normal diffusion with diffusion constant proportional to the square root of its diffusion constant at zero flow. Chirality effects in the propulsion mechanism, modeled here by a tunable applied torque, favors particle's jumping between adjacent convection rolls. Roll jumping is signaled by an excess diffusion peak, which appears to separate two distinct active diffusion regimes for low and high chirality. A qualitative interpretation of our simulation results is proposed as a first step toward a fully analytical study of this phenomenon.

cond-mat.soft↗

arXiv2019

Robust Vector Perturbation Precoding Design for MIMO Broadcast Channel

We consider the vector perturbation (VP) precoder design for multiuser multiple-input single output (MU-MISO) broadcast channel systems which is robust to power scaling factor errors. VP precoding has so far been developed and analyzed under the assumption that receivers could have known the power scaling factor in advance of tranmission perfectly, which is hard to obtain due to the large dynamic range and limited feedforward. However, as demonstrated in our results the performance of VP precoding is quite sensitive to the accuracy of power scaling factor and always encounter an error floor at mid to high signal-to-noise ratio (SNR) regimes. Motivated by such observations, we propose a robust VP precoder based on the minimum mean square error (MMSE) criterion. Simulation results show that, the robust VP precoder outperforms conventional VP precoding designs, as less sensitive to power scaling factor errors.

eess.SP↗

arXiv2018

DLBI: Deep learning guided Bayesian inference for structure reconstruction of super-resolution fluorescence microscopy

Super-resolution fluorescence microscopy, with a resolution beyond the diffraction limit of light, has become an indispensable tool to directly visualize biological structures in living cells at a nanometer-scale resolution. Despite advances in high-density super-resolution fluorescent techniques, existing methods still have bottlenecks, including extremely long execution time, artificial thinning and thickening of structures, and lack of ability to capture latent structures. Here we propose a novel deep learning guided Bayesian inference approach, DLBI, for the time-series analysis of high-density fluorescent images. Our method combines the strength of deep learning and statistical inference, where deep learning captures the underlying distribution of the fluorophores that are consistent with the observed time-series fluorescent images by exploring local features and correlation along time-axis, and statistical inference further refines the ultrastructure extracted by deep learning and endues physical meaning to the final image. Comprehensive experimental results on both real and simulated datasets demonstrate that our method provides more accurate and realistic local patch and large-field reconstruction than the state-of-the-art method, the 3B analysis, while our method is more than two orders of magnitude faster. The main program is available at https://github.com/lykaust15/DLBI

cs.CV↗

arXiv2018

Handover Control in Wireless Systems via Asynchronous Multi-User Deep Reinforcement Learning

In this paper, we propose a two-layer framework to learn the optimal handover (HO) controllers in possibly large-scale wireless systems supporting mobile Internet-of-Things (IoT) users or traditional cellular users, where the user mobility patterns could be heterogeneous. In particular, our proposed framework first partitions the user equipments (UEs) with different mobility patterns into clusters, where the mobility patterns are similar in the same cluster. Then, within each cluster, an asynchronous multi-user deep reinforcement learning scheme is developed to control the HO processes across the UEs in each cluster, in the goal of lowering the HO rate while ensuring certain system throughput. In this scheme, we use a deep neural network (DNN) as an HO controller learned by each UE via reinforcement learning in a collaborative fashion. Moreover, we use supervised learning in initializing the DNN controller before the execution of reinforcement learning to exploit what we already know with traditional HO schemes and to mitigate the negative effects of random exploration at the initial stage. Furthermore, we show that the adopted global-parameter-based asynchronous framework enables us to train faster with more UEs, which could nicely address the scalability issue to support large systems. Finally, simulation results demonstrate that the proposed framework can achieve better performance than the state-of-art on-line schemes, in terms of HO rates.

cs.NI↗

arXiv2017

UAV-assisted Cooperative Communications with Wireless Information and Power Transfer

In this paper, we focus on a typical cooperative communication system with one pair of source and destination, where a unmanned aerial vehicle (UAV) flying from a start location to an end location serves as a mobile relay. To efficiently utilize energy in ambient environment, the UAV's transmission capability is powered exclusively by radio signal transmitted from the source via the power-splitting mechanism. In such a cooperative communication system, we study the end-to-end cooperative throughput maximization problem by optimizing the UAV's power profile, power-splitting ratio profile and trajectory for both amplify-and-forward (AF) and decode-and-forward (DF) protocols. The problem is decomposed into two subproblems: profile optimization given trajectory and trajectory optimization given profile. The former one is solved via the dual decomposition with in-depth analysis and the latter one is solved via successive convex optimization, by which a lower bound is iteratively maximized. Then the end-to-end cooperative throughput maximization problem is solved by alternately solving the two subproblems. The numerical results show that with the proposed optimal solution, choice for the UAV's power profile and power-splitting ratio profile is more long-sighted than the greedy strategy from our previous work and the successive optimization is able to converge in a few rounds of iteration. Moreover, as for the end-to-end cooperative throughput, the proposed optimal solution outperforms both static and greedy strategies, especially for the AF protocol.

cs.NI↗

arXiv2012

Reliable Multicasting for Device-to-Device Radio Underlaying Cellular Networks

This paper proposes Leader in Charge (LiC), a reliable multicast architecture for device-to-device (D2D) radio underlaying cellular networks. The multicast-requesting user equipments (UEs) in close proximity form a D2D cluster to receive the multicast packets through cooperation. In addition to receiving the multicast packets from the eNB, UEs share what they received from the multicast on short-range links among UEs, namely the D2D links, to exploit the wireless resources a more efficient way. Consequently, we show that utilizing the D2D links in cellular networks increases the throughput of a multicast session by means of simulation. We also discuss some practical issues facing the integration of LiC into the current cellular networks. In particular, we propose efficient delay control mechanism to reduce the average and maximum delay experienced by LiC users, which is further confirmed by the simulation results.

cs.MM↗

arXiv2012

Energy-Efficient Transmission Schemes in Cooperative Cellular Systems

Energy-efficient communication is an important requirement for mobile devices, as the battery technology has not kept up with the growing requirements stemming from ubiquitous multimedia applications. This paper considers energy-efficient transmission schemes in cooperative cellular systems with unbalanced traffic between uplink and downlink. Theoretically, we derive the optimal transmission data rate, which minimizes the total energy consumption of battery-powered terminals per information bit. The energy-efficient cooperation regions are then investigated to illustrate the effects of relay locations on the energy-efficiency of the systems, and the optimal relay location is found for maximum energy-efficiency. Finally, numerical results are provided to demonstrate the tradeoff between energy-efficiency and spectral efficiency.

cs.IT↗

arXiv2010

Joint Uplink and Downlink Relay Selection in Cooperative Cellular Networks

We consider relay selection technique in a cooperative cellular network where user terminals act as mobile relays to help the communications between base station (BS) and mobile station (MS). A novel relay selection scheme, called Joint Uplink and Downlink Relay Selection (JUDRS), is proposed in this paper. Specifically, we generalize JUDRS in two key aspects: (i) relay is selected jointly for uplink and downlink, so that the relay selection overhead can be reduced, and (ii) we consider to minimize the weighted total energy consumption of MS, relay and BS by taking into account channel quality and traffic load condition of uplink and downlink. Information theoretic analysis of the diversity-multiplexing tradeoff demonstrates that the proposed scheme achieves full spatial diversity in the quantity of cooperating terminals in this network. And numerical results are provided to further confirm a significant energy efficiency gain of the proposed algorithm comparing to the previous best worse channel selection and best harmonic mean selection algorithms.

cs.IT↗

arXiv2010

Joint Relay Selection and Link Adaptation for Distributed Beamforming in Regenerative Cooperative Networks

Relay selection enhances the performance of the cooperative networks by selecting the links with higher capacity. Meanwhile link adaptation improves the spectral efficiency of wireless data-centric networks through adapting the modulation and coding schemes (MCS) to the current link condition. In this paper, relay selection is combined with link adaptation for distributed beamforming in a two-hop regenerative cooperative system. A novel signaling mechanism and related optimal algorithms are proposed for joint relay selection and link adaptation. In the proposed scheme, there is no need to feedback the relay selection results to each relay. Instead, by broadcasting the link adaptation results from the destination, each relay will automatically understand whether it is selected or not. The lower and upper bounds of the throughput of the proposed scheme are derived. The analysis and simulation results indicate that the proposed scheme provides synergistic gains compared to the pure relay selection and link adaptation schemes.

cs.IT↗

arXiv2010

Energy-Efficient Relay Selection and Optimal Relay Location in Cooperative Cellular Networks with Asymmetric Traffic

Energy-efficient communication is an important requirement for mobile relay networks due to the limited battery power of user terminals. This paper considers energy-efficient relaying schemes through selection of mobile relays in cooperative cellular systems with asymmetric traffic. The total energy consumption per information bit of the battery-powered terminals, i.e., the mobile station (MS) and the relay, is derived in theory. In the Joint Uplink and Downlink Relay Selection (JUDRS) scheme we proposed, the relay which minimizes the total energy consumption is selected. Additionally, the energy-efficient cooperation regions are investigated, and the optimal relay location is found for cooperative cellular systems with asymmetric traffic. The results reveal that the MS-relay and the relay-base station (BS) channels have different influence over relay selection decisions for optimal energy-efficiency. Information theoretic analysis of the diversity-multiplexing tradeoff (DMT) demonstrates that the proposed scheme achieves full spatial diversity in the quantity of cooperating terminals in this network. Finally, numerical results further confirm a significant energy efficiency gain of the proposed algorithm comparing to the previous best worse channel selection and best harmonic mean selection algorithms.

cs.IT↗