To address the dual demands of high-efficiency communication for multiple devices and high-precision positioning for multiple passive radio frequency identification (RFID) tags in the Internet of Things (IoT), this paper investigates an integrated sensing and backscatter communication (ISABC) system based on a uniform planar array (UPA). Focusing on resource-constrained IoT gateways, we aim to minimize the Cramér-Rao lower bound (CRLB) for high-precision two-dimensional (2D) direction-of-arrival (DoA) estimation of passive tags via precoding design, subject to the gateway power budget and communication quality-of-service (QoS) constraints. To balance high-precision positioning with real-time processing in dynamic IoT environments, we propose a configurable framework comprising three schemes: 1) joint sensing and communication precoding (JSCP) as a performance benchmark; 2) layered-precoding communication-optimized precoding (LP-COP) to effectively balance sensing accuracy and algorithmic latency; and 3) a low-complexity fixed-beam with dynamic power allocation precoding (FB-PDPA). Simulation results demonstrate that the proposed LP-COP scheme achieves near-optimal sensing accuracy comparable to the JSCP benchmark while reducing computation time by approximately 96% and maintaining strong robustness under strict signal-to-interference-plus-noise ratio (SINR) constraints. The proposed framework provides a flexibly configurable technical path for IoT ISABC deployments.

Recently, we have been witnessing the scale-up of superconducting quantum computers; however, the noise of quantum bits (qubits) is still an obstacle for real-world applications to leveraging the power of quantum computing. Although there exist error mitigation or error-aware designs for quantum applications, the inherent fluctuation of noise (a.k.a., instability) can easily collapse the performance of error-aware designs. What’s worse, users can even not be aware of the performance degradation caused by the change in noise. To address both issues, in this paper we use Quantum Neural Network (QNN) as a vehicle to present a novel compression-aided framework, namely QuCAD, which will adapt a trained QNN to fluctuating quantum noise. In addition, with the historical calibration (noise) data, our framework will build a model repository offline, which will significantly reduce the optimization time in the online adaption process. Emulation results on an earthquake detection dataset show that QuCAD can achieve 14.91% accuracy gain on average in 146 days over a noise-aware training approach. For the execution on a 7-qubit IBM quantum processor, ibm-jakarta, QuCAD can consistently achieve 12.52% accuracy gain on earthquake detection.

Model compression, such as pruning and quantization, has been widely applied to optimize neural networks on resource-limited classical devices. Recently, there are growing interest in variational quantum circuits (VQC), that is, a type of neural network on quantum computers (a.k.a., quantum neural networks). It is well known that the near-term quantum devices have high noise and limited resources (i.e., quantum bits, qubits); yet, how to compress quantum neural networks has not been thoroughly studied. One might think it is straightforward to apply the classical compression techniques to quantum scenarios. However, this paper reveals that there exist differences between the compression of quantum and classical neural networks. Based on our observations, we claim that the compilation/traspilation has to be involved in the compression process. On top of this, we propose the very first systematical framework, namely CompVQC, to compress quantum neural networks (QNNs). In CompVQC, the key component is a novel compression algorithm, which is based on the alternating direction method of multipliers (ADMM) approach. Experiments demonstrate the advantage of the CompVQC, reducing the circuit depth (almost over 2.5×) with a negligible accuracy drop (<1%), which outperforms other competitors. Another promising truth is our CompVQC can indeed promote the robustness of the QNN on the near-term noisy quantum devices.

The rapid growth in the size of Graph Convolutional Neural Networks (GCNs) encounters both computational- and memory-wall on classical computing platforms (e.g., CPU, GPU, FPGA, etc.). Quantum computing, on the other hand, provides extremely high parallelism for computation. Although quantum neural networks have been recently studied, the research on quantum graph neural networks is still in its infancy. The key challenge here is how to integrate both the graph topology information and the learning ability of GCNs into quantum circuits. In this work, we leverage the Givens rotations and its quantum implementation to encode graph information; in addition, we employ the widely used variational quantum circuit to bring the learnable parameters. On top of these, we present a full-quantum design of Graph Convolutional Neural Networks, namely "QuGCN", for semi-supervised learning on graph-structured data. Experiment results show our design is competitive with classical GCNs in terms of node classification accuracy on Cora sub-dataset. More importantly, we show the potential advantages that can be achieved by the proposed quantum GCN design when the number of features grows.

Abstract This paper focuses on the fairness‐aware nonlinear joint transceiver design for the energy‐harvesting (EH)‐powered coordinated multi‐point (CoMP) systems. In the EH‐powered CoMP systems, each node harvests the energy independently first. Then, these nodes collaborate for joint signal transmission. Since there is no conventional grid to connect the nodes, the energy among nodes cannot be coordinated. Therefore, per‐node power constraints should be satisfied, which however has not been considered in existing schemes. In this paper, a fairness‐aware nonlinear transceiver scheme is developed under the per‐node power constraints. The nonlinear transceiver design is formulated as an optimization problem to maximize the minimum signal‐to‐interference noise ratio of data streams. A two‐step optimization algorithm is proposed to obtain the closed‐form solutions in the cases with perfect channel state information (CSI) and imperfect CSI, respectively. Simulation results verify the fairness of the proposed algorithm, and demonstrate its performance enhancement in sum‐rate and bit ratio error (BER). For sum‐rate, the proposed algorithm can achieve 12% gain over WS‐MSE and 40% gain over BD at SNR = 20 dB. For BER, it can achieve an approximately 4 dB gain over WS‐MSE and 8 dB gain over BD at BER = 10 −3 .

Interpreting the user intent in search queries is a key task in query understanding. Query intent classification has been widely studied. In this article, we go one step further to understand the query from the view of head--modifier analysis. For example, given the query “popular iphone 5 smart cover,” instead of using coarse-grained semantic classes (e.g.,find electronic product), we interpret that “smart cover” is the head or the intent of the query and “iphone 5” is its modifier. Query head--modifier detection can help search engines to obtain particularly relevant content, which is also important for applications such as ads matching and query recommendation. We introduce an unsupervised semantic approach for query head--modifier detection. First, we mine a large number of instance level head--modifier pairs from search log. Then, we develop a conceptualization mechanism to generalize the instance level pairs to concept level. Finally, we derive weighted concept patterns that are concise, accurate, and have strong generalization power in head--modifier detection. The developed mechanism has been used in production for search relevance and ads matching. We use extensive experiment results to demonstrate the effectiveness of our approach.

This paper studies the nonlinear transceiver design in downlink coordinated multi-cell system where base stations (BSs) are powered by hybrid sources, including conventional grid and renewable energy. Renewable energy is collected by energy harvesting equipment deployed at BS. However, the harvested energy from different BSs varies in space and time, and the system gain will be limited by the least energy of the coordinated BSs. One of the methods to settle this drawback is energy cooperation among BSs. The nonlinear joint transceiver is formulated as an optimization problem to maximize the minimum signal-to-interference noise ratio of streams under the transmit power constraint. The closed-form solutions are derived by the proposed two-step algorithm. The proposed algorithm guarantees the performance balancing among all users and all streams of each user. Simulation results illustrate the performance improvement of the proposed algorithm.

In multiuser MIMO systems, the required feedback rate per user increases linearly with the number of transmit antennas in order to achieve full multiplexing gain. When it comes to massive MIMO systems, the feedback overhead grows unacceptable. This motivates us to explore a novel feedback reduction scheme based on principal component analysis (PCA). The proposed PCA based feedback scheme exploits the spatial correlation characteristics of massive MIMO channel model, since transmit antennas are deployed compactly at base station (BS). In the proposed scheme, mobile station (MS) utilizes compression matrix to compress spatially correlated high-dimensional channel state information (CSI) into low-dimensional one. Then the compressed low-dimensional CSI is fed back to BS instantaneously with reduced feedback overhead and codebook search complexity. The compression matrix is attained by operating PCA on CSI which is estimated over a long-term period by MS. In order to recover high-dimensional CSI at BS, compression matrix is refreshed and fed back from MS to BS every long-term period. Numerical results and feedback overhead analysis show that the proposed PCA based feedback scheme can offer a tradeoff between system performance and feedback overhead.

Per-cell codebook based limited feedback technique in the coordinated multi-point transmission (CoMP) system has received more and more attention because of its flexibility and scalability. In practical CoMP systems, quantized channel state information (CSI) is subject to a feedback delay due to signal processing at the receiver, finite bandwidth of feedback links, and finite capacity of backhaul links. In this paper, we study the effect of different codeword selection schemes based on per-cell codebook on data rate in CoMP systems. The impact of CSI feedback delay is considered by using Gauss-Markov temporal correlation channel model. We derive the theoretical upper bounds of data rate for two codeword selection schemes, joint codeword selection (JCS) and independent codeword selection (ICS) respectively. Simulation results show that the practical data rate is close to the theoretical data rate upper bound both for JCS and ICS, which has validated the availability of the theoretical data rate upper bounds.

Multi-cell coordinated system is an attractive way of improving data rate, by serving one user through several base stations (BSs). BSs might share both data and their channel state information (CSI), but the demand on backhaul capacity makes it inflexible. In this paper, we consider the case with BSs sharing data dedicated to users but having only local CSI, and propose a decentralized nonlinear precoding algorithm for multi-cell coordinated system with multi-stream multi-antenna users. The proposed algorithm is designed to eliminate the inter-user interference and the inter-stream interference, and meanwhile to achieve equal performance for streams of each user. In the proposed algorithm, Tomlinson-Harashima precoding is applied to eliminate partial interference, and transmit space matrix is designed to eliminate the other interference. After the interference elimination, closed-form expression of the transmit and receive processing matrix are derived to maximize the minimum signal to interference plus noise ratio for streams of each user. Simulation results show that, compared with the decentralized linear precoding, the proposed algorithm achieves better performance.

Head and modifier detection is an important problem for applications that handle short texts such as search queries, ads keywords, titles, captions, etc. In many cases, short texts such as search queries do not follow grammar rules, and existing approaches for head and modifier detection are coarse-grained, domain specific, and/or require labeling of large amounts of training data. In this paper, we introduce a semantic approach for head and modifier detection. We first obtain a large number of instance level head-modifier pairs from search log. Then, we develop a conceptualization mechanism to generalize the instance level pairs to concept level. Finally, we derive weighted concept patterns that are concise, accurate, and have strong generalization power in head and modifier detection. Furthermore, we identify a subset of modifiers that we call constraints. Constraints are usually specific and not negligible as far as the intent of the short text is concerned, while non-constraint modifiers are more subjective. The mechanism we developed has been used in production for search relevance and ads matching. We use extensive experiment results to demonstrate the effectiveness of our approach.

Per-cell codebook based precoding is an effective transmission method in coordinated multi-point (CoMP) systems with the limited feedback constraint. There are two codeword selection methods for per-cell codebook, one is joint codeword selection (JCS) and another is independent codeword selection (ICS). In this paper, the system throughput obtained by the JCS and the ICS is analyzed firstly, which shows that JCS has a better trade-off between system throughput and feedback overhead. Then a codebook compression scheme for the JCS is proposed, which decreases the selection complexity of JCS by reducing the size of the codebook. Simulation results show that the proposed scheme can decrease the selection complexity greatly with tiny loss on performance and no additional feedback overhead.