Rapid development of web services has led to an explosion of multimodal content, making multimodal recommender systems (MRSs) vital tools for mitigating information overload. Current MRSs have achieved remarkable progress by incorporating advanced technologies such as Graph Neural Networks (GNNs) and Large Language Models (LLMs). However, these studies still suffer from the semantic shift problem. Generally, item's multimodal content usually contain multiple objects, including target object (core content of item) and auxiliary objects (decorations of item). Existing MRSs overlooked this distinction, failing to prevent auxiliary objects from dominating the representation, leading to biased item representation. To address this issue, we propose a model-agnostic framework ''TargetMR''. Concretely, TargetMR comprises two core modules, including Object Disentangler and Object Identifier. The Object Disentangler decouples item text and image into multiple objects via text syntactic parsing and image segmentation. The Object Identifier performs knowledge distillation based on LLMs to efficiently identify the target text object. It then identifies the target image object through cross-modal semantic evaluation. Moreover, this module refines the representation of image target object by optimizing the semantic correlation. Owing to the model-agnostic design of TargetMR, it can be integrated into various backbone MRSs. Extensive experiments on three benchmark datasets show that TargetMR consistently improves the performance of five backbone MRSs, with an average improvement of 12.26%. Our codes are available at https://github.com/gutang-97/TargetMR/.

In this brief, a high-speed [multiway parallel (MWP)] hardware-implemented deflate data compressor (DDC) is proposed for reducing the storage of solid-state drives (SSDs). To minimize the area of the DDC, registers instead of static random access memories (SRAMs) are utilized for building hash tables because multiway data within the DDC are able to access a register-based hash table simultaneously. To further reduce the area of the DDC, the output data of indefinite length are concatenated with a tree-type hardware architecture for reducing the overall concatenation complexity. Moreover, a solid mathematical foundation is established for optimizing the latency values of Lempel–Ziv (LZ)77 circuit, the Huffman encoding circuit, and the output data concatenation circuit within the MWP DDC. The results show that the proposed MWP DDC is capable of achieving a 12.1-Gb/s throughput and a 1.76 compression ratio (CR) with a 1.17-mm2area and 0.103-$\mu $s latency, under the synthesis of SMIC 55-nm process design kits (PDKs). Hence, the proposed DDC satisfies the SSD compression requirement for a universal serial bus (USB) 3.2 connector.

Utilizing large language models (LLMs) for tool planning has emerged as a promising avenue for developing general AI systems, where LLMs automatically schedule external tools (e.g., vision models) to tackle complex tasks based on task descriptions. To push this paradigm toward practical applications, it is crucial for LLMs to consider tool execution costs (e.g., execution time) for tool planning. Unfortunately, prior studies overlook the tool execution costs, leading to the generation of expensive plans whose costs outweigh their benefits in terms of task performance. To fill this gap, we propose the Cost-Aware Tool Planning with LLMs (CATP-LLM) framework, which for the first time provides a coherent design to empower LLMs for cost-aware tool planning. Specifically, To facilitate efficient concurrent tool execution and cost reduction, we design a tool planning language to enhance the LLM for creating multi-branch non-sequential plans. Moreover, we propose a cost-aware offline reinforcement learning algorithm to fine-tune the LLM to optimize the performance-cost trade-off in tool planning. In the lack of public cost-related datasets, we further present OpenCATP, the first dataset for cost-aware planning, which comprises 11,100 evaluation samples from diverse tasks. Extensive experiments show that CATP-LLM outperforms GPT-4 even when using Llama2-7B as its backbone, with the average improvement of 1.5%-93.9% in terms of plan quality. Codes and dataset are available at: https://github.com/duowuyms/OpenCATP-LLM.

With the explosive growth of online multimodal content, multimodal recommender systems(MRSs) have brought significant benefits to multimedia platforms. As MRSs evolve, many studies incorporate advanced technologies like graph neural networks(GNNs) and self-supervised learning(SSL), achieving remarkable results. However, these efforts still suffer from the quality disparity problem. It refers to the mixture of high and low quality across items' multiple modalities, owing to disparities in construction costs or design levels. These low-quality modalities often lack crucial details or introduce noise to the depiction of item, leading to insufficient or polluted item representation. Therefore, we propose a novel framework R2MR: Review and Rewrite Modality for Recommendation to tackle this issue. Specifically, R2MR is composed of two key components: Modality Reviewer and Modality Rewriter. The Modality Reviewer introduces a Consensus Review Mechanism. It performs perspective decomposition based on user representations and learns the consensus quality scores for modalities from diverse perspectives of multiple users. The Modality Rewriter proposes a Latent Mapping Model, which improves the quality of inferior modalities by learning various mapping patterns from high-quality modalities. Comprehensive experiments across three benchmark datasets reveal that R2MR substantially outperforms state-of-the-art methods, achieving an average improvement of 9.20%. The implementations are available at https://github.com/gutang-97/R2MR.

With the enrichment of user-item interactions, Graph Neural Networks (GNNs) are widely used in recommender systems to alleviate information overload. Nevertheless, they still suffer from the cold-start issue. Knowledge Graphs (KGs), providing external information, have been extensively applied in GNN-based methods to mitigate this issue. However, current KG-aware recommendation methods suffer from the knowledge imbalance problem caused by incompleteness of existing KGs. This imbalance is reflected by the long-tail phenomenon of item attributes, i.e., unpopular items usually lack more attributes compared to popular items. To tackle this problem, we propose a novel framework called EditKG: Editing Knowledge Graph for Recommendation, to balance attribute distribution of items via editing KGs. EditKG consists of two key designs: Knowledge Generator and Knowledge Deleter. Knowledge Generator generates attributes for items by exploring their mutual information correlations and semantic correlations. Knowledge Deleter removes the task-irrelevant item attributes according to the parameterized task relevance score, while dropping the spurious item attributes through aligning the attribute scores. Extensive experiments on three benchmark datasets demonstrate that EditKG significantly outperforms state-of-the-art methods, and achieves 8.98% average improvement. The implementations are available at https://github.com/gutang-97/2024SIGIR-EditKG.

The reconfigurable intelligent surface (RIS) has received a lot of interest because of its capacity to reconfigure the wireless communication environment in a cost- and energy-efficient way. However, the realistic power consumption modeling and measurement validation of RIS has received far too little attention. Therefore, in this work, we model the power consumption of RIS and conduct measurement validations using various RISs to fill this vacancy. Firstly, we propose a practical power consumption model of RIS. The RIS hardware is divided into three basic parts: the FPGA control board, the drive circuits, and the RIS unit cells. The power consumption of the first two parts is modeled asPstaticand that of the last part is modeled asPunits. Expressions ofPstaticandPunitsvary amongst different types of RISs. Secondly, we conduct measurements on various RISs to validate the proposed model. Five different RISs including the PIN diode, varactor diode, and RF switch types are measured, and measurement results validate the generality and applicability of the proposed power consumption model of RIS. Finally, we summarize the measurement results and discuss the approaches to achieve the low-power-consumption design of RIS-assisted wireless communication systems.

Reconfigurable intelligent surface (RIS) is widely accepted as a potential technology to assist in communication between base stations (BSs) and users in edge areas. We study the energy efficiency of a RIS-assisted multi-cell communication system with a realistic RIS power consumption model. With the goal of maximizing the energy efficiency of the system, we optimize the transmit beamforming vectors at the BS and the RIS phase shift matrix by a proposed alternative optimization algorithm. First, the transmit beamforming vector is optimized by solving the transformed weighted minimum mean square error (WMMSE) problem. Subsequently, to solve the inconvenience incurred by the discrete relationship between the RIS reflecting unit power consumption and its discrete phase shift, we use a continuous function to approximate their relationship. With this approximation, we can use the majorization minimization (MM) technique to optimize the continuous RIS phase shifts, and then quantize the obtained phase shifts to discrete ones. Simulation results demonstrate that the energy efficiency of the system is effectively optimized by the proposed algorithm.

Reconfigurable intelligent surface (RIS) has emerged as a competitive solution to the blocking problem in millimeter wave (mmWave) communications. However, due to the passive nature of the RIS, obtaining channel state information (CSI) for RIS-assisted mmWave communication systems is rather difficult. Considering that the currently available RIS hardware cannot arbitrarily switch between the active (reflection with configurable phase response) and deactivate (absorption) modes, we suggest a new beam training method for RIS-assisted mmWave communication systems in this study. First, a predefined hierarchical codebook is created using the pattern synthesis method. Then, we provide a novel hierarchical beam training method using two multi-mainlobe codewords in each layer of the hierarchical codebook for beam sweeping. Combining the results of the beam identification in all the layers will yield the ultimate ideal beam direction. Theoretical analyses demonstrate that the suggested approach can effectively reduce training overhead while ensuring successful beam alignment. Simulation results show that the practical codebook can be created successfully, and the suggested method can achieve accurate beam alignment with reduced training overhead.

Future wireless communication networks are expected to fulfill the unprecedented performance requirements to support our highly digitized and globally data-driven society. Various technological challenges must be overcome to achieve our goal. Among many potential technologies, reconfigurable intelligent surface (RIS) and artificial intelligence (AI) have attracted extensive attention, thereby leading to a proliferation of studies for utilizing them in wireless communication systems. The RIS-based wireless communication frameworks and AI-enabled technologies, two of the promising technologies for the sixth-generation networks, interact and promote with each other, striving to collaboratively create a controllable, intelligent, reconfigurable, and programmable wireless propagation environment. This paper explores the road to implementing the combination of RIS and AI, more specifically, integrating AI-enabled technologies into RIS-based frameworks for maximizing the practicality of RIS to facilitate the realization of smart radio propagation environments, elaborated from shallow to deep insights. We begin with the basic concept and fundamental characteristics of RIS, followed by the overview of the research status of RIS. Then, we analyze the inevitable trend of RIS to be combined with AI. In particular, we focus on recent research about RIS-based architectures embedded with AI, elucidating from the intelligent structures and systems of metamaterials to the AI-embedded RIS-assisted wireless communication systems. Finally, the challenges and potential of the topic are discussed.

Existing dynamic programming based track-before-detect (DP-TBD) strategies suffer from merit function expansion phenomenon (MFEP), which aggravated the burden of designing the detection threshold. The traditional constant false alarm rate (CFAR) detection is ineffective because the noise energy can not be exactly estimated from the area of merit function expansion. the threshold setting of existing DP-TBD strategies usually resort to the traditional Monte-Carlo counting, the extreme-value theory or its generalized version. For the nonhomogeneous clutter background and the fluctuating target, all of these constant threshold setting strategies inevitably exist the target losing or higher false alarm rate. In addition, for the multi-target scenes, in order to avoid solving high-dimensional optimization problems, existing the most effective DP-TBD methods all use the additional heuristic procedures to extract target trajectories one-by-one from the merit function expansion area by assuming target tracks are always independent. To overcome the aforementioned challenges, a novel one-step greedy optimization TBD algorithm (OSP-TBD) is proposed in this paper. By constraining the physically admissible trajectories, such that the different targets do not occupy the same resolution cell during the same stage and the trajectory with higher merit function (MF) is estimated ahead of others, OSP-TBD can eliminate the MFEP intrinsically and traditional CFAR procedure can be used to detect target adaptively. Besides, the proposed OSP-TBD algorithm can be used to process multi-target situation directly and declare all of the target trajectories corresponding to the states whose MF at the final frame exceed the given detection threshold without any additional heuristic procedure. Numerical simulations are used to assess the performance of the proposed strategies.

Multi-frame detection (MFD) and multi-sensor fusion are two popular methods of target detection and estimation which can improve the performance by increasing the number of measurement samples. In this paper, we combine these two methods together, proposing a novel multi-sensor multi-frame detection (MS-MFD) method. On the one hand, MS-MFD can make use of the target information as much as possible through the multi-frame integration. On the other hand, it can acquire the target space-diversity gain by jointly processing the measurement samples on different observation orientations, providing more accurate estimates. In particular, the proposed method consists of two steps. First, it conducts the MFD processing in each sensor node, computing the local multi-frame jointly posterior probability density. Then, it transmits the local densities to the fusion center for further processing, calculating the global target estimates. Furthermore, in order to improve the implementation efficiency of MS-MFD, a Gaussian Mixture model based method is proposed to approximate the distribution of local posterior probability density, so that the transmission costs of local posterior probability density can be significantly reduced. It is demonstrated by simulations that the proposed methods show superior performance.

In this paper, we address the target detection problem using multi-sensor dynamic programming based track before detect (DP-TBD) methods. First, we give two implementation methods of multi-sensor DP-TBD under the centralized processing and the distributed processing, respectively. Then, in order to improve the implementation efficiency of the multi-sensor DP-TBD, we further propose an improved DP-TBD method based on the approximation of local likelihood. Particularly, the proposed method first calculates the likelihood locally in the sensor nodes, then approximates the likelihood with a weighted sum of a number of basis functions, and finally transmits the weighted coefficients rather than all likelihood to the fusion center for further processing with DP-TBD. By this means, the proposed method can reduce the communication requirements of the system. In addition, since the likelihood are calculated locally, the computational burden of the fusion center can also be alleviated. The analysis and simulation results demonstrate that the proposed method can improve the implementation efficiency significantly with limited performance loss in comparison with the centralized/distributed processing DP-TBD.

In this paper, we consider the fluctuating targets detection problem in a distributed multi-sensor network. A multi-sensor multi-frame track-before-detect (MS-MF-TBD) procedure is proposed to sufficiently make use of the target energy diversity in space and time dimensions (space-time diversity). Two MS-MF-TBD methods, the multi-sensor maximum likelihood-probabilistic data association (MS-ML-PDA) and the multi-sensor dynamic programming based TBD (MS-DP-TBD), are derived, and a number of simulation experiments under different target fluctuation models are performed. Through these simulations, we demonstrate that by using the space-time diversity, MS-MF-TBD methods can efficiently detect the fluctuating targets, achieving significant detection performance gains in comparison to either the single sensor TBD methods or the conventional multi-sensor detection methods.