Long-form books are among the most information-rich and structurally complex forms of written content, often exceeding 100,000 words. While recent methods have enabled basic long-text generation, they remain limited in two key aspects: the inability to generate ultra-long content at book scale, and the lack of mechanisms for integrating rich factual information. To address these limitations, we propose DeepWriter, a multi-agent collaborative framework that follows a structured planning-then-generation paradigm. It first constructs a detailed book outline with narrative arcs and chapter semantics, then incrementally generates content conditioned on retrieved knowledge and contextual signals. DeepWriter supports controllable generation of full-length books exceeding 100,000 words, enriched with citations, trivia and images. To support evaluation beyond surface-level fluency, we introduce DeepWriter-Bench, a bilingual benchmark of 18 annotated books designed to assess book-scale coherence, richness, and factual grounding. Additionally, we propose BookScore, a unified 100-point metric for quantifying book maturity. Experimental results show that DeepWriter achieves a state-of-the-art BookScore of 80.92, consistently outperforming strong baselines.

The current Retrieval-Augmented Generation(RAG) pipeline incorporates retrieved contexts as external knowledge to reduce hallucinations, but the large and redundant context makes it difficult for the model to identify key information. In Question-Answering (QA) tasks, large language models (LLMs) can identify query-focused information through extractive compression, simplify redundant long texts and improve the accuracy, but they are resource-intensive and less efficient. To ensure the accuracy of extractive compression and improve its efficiency, we propose a framework for query-focused extractive compression (QEC-LLM) that leverages the LLM’s knowledge and a queryguided mechanism to construct sample sets, effectively training a smaller student model. The student model sorts and scales paragraph sentences based on similarity scores for compression and performance optimization. In addition, since LLM may generate some information that is missing from the student model during the extraction process, our framework will add this supplementary information to the QA prompt and combine it with the extracted information to achieve generative augmentation. We conduct extensive experiments on QA datasets with different compression ratios. The results demonstrate that QECLLM significantly improves end-to-end QA performance and inference efficiency, while demonstrating strong generalization across different retrieval strategies.

Multigrid (MG) methods are among the best choice for stable and efficient three-dimensional (3D) forward modeling of electromagnetic (EM) fields over large area due to their linear dependence of computational time on grid size. However, as the frequency decreases to near zero and/or the grid is increasingly stretched, MG solvers for EM modeling based on the curl-curl equations converge slowly or even diverge. In this letter, we develop an efficient MG algorithm combined with a two-color plane Gauss-Seidel (GS) smoother for finite difference forward modeling of EM fields particularly at low frequencies. In this algorithm, we group different planes of the grid nodes into two colors. In each color, the components attached to different planes are totally decoupled and can be solved simultaneously, which can be distributed to different processors. The Dublin Test Model 1 is used to verify the accuracy of our algorithm and examine the numerical performance of our method against MG algorithms based on a four-color cell-block GS smoother and the Bi-Conjugate Gradient stabilized (BICGstab) smoother (as four-color cell-block GS MG and BICGstab-MG, respectively), and BICGstab and Quasi-Minimal Residual (QMR) both preconditioned with block incomplete lower-upper (blockILU) decomposition (as blockILU-BICGstab and blockILU-QMR, respectively). The numerical test based on OpenMP shows the good parallelization of our algorithm. Grids stretched to different degrees are designed to examine its ability to handle grid-stretching. The numerical performance comparison indicates its remarkable dominance in efficiency and stability.

It is difficult to annotate all instances for object detection tasks in practice, especially facing the crowned scenes. The model performance could be impaired when training on sparsely annotated datasets, compared to fully annotated ones. To alleviate this problem, we propose a novel robust training framework for sparsely annotated object detection. We take the sparse annotations as a special case of noisy annotations, where the instances without annotations are regarded as background. First, we propose an Ensemble Prediction (EP) module from different temporal models to avoid the model fitting to the background (noisy labels). Further, an Ensemble Models Non-Maximum Suppression (EM-NMS) module is proposed based on the EP module to generate more robust bounding boxes. Finally, the Pseudo Labels Generating (PLG) module is used to produce pseudo labels for the possible missing-annotation instances. Extensive validation experiments are conducted on PASCAL VOC and MS COCO datasets with sparse annotations. The results show that our proposed robust training framework can further improve model performance when compared to related studies.

MOTIVATION: Cancer somatic driver mutations associated with genes within a pathway often show a mutually exclusive pattern across a cohort of patients. This mutually exclusive mutational signal has been frequently used to distinguish driver from passenger mutations and to investigate relationships among driver mutations. Current methods for de novo discovery of mutually exclusive mutational patterns are limited because the heterogeneity in background mutation rate can confound mutational patterns, and the presence of highly mutated genes can lead to spurious patterns. In addition, most methods only focus on a limited number of pre-selected genes and are unable to perform genome-wide analysis due to computational inefficiency. RESULTS: We introduce a statistical framework, MEScan, for accurate and efficient mutual exclusivity analysis at the genomic scale. Our framework contains a fast and powerful statistical test for mutual exclusivity with adjustment of the background mutation rate and impact of highly mutated genes, and a multi-step procedure for genome-wide screening with the control of false discovery rate. We demonstrate that MEScan more accurately identifies mutually exclusive gene sets than existing methods and is at least two orders of magnitude faster than most methods. By applying MEScan to data from four different cancer types and pan-cancer, we have identified several biologically meaningful mutually exclusive gene sets. AVAILABILITY AND IMPLEMENTATION: MEScan is available as an R package at https://github.com/MarkeyBBSRF/MEScan. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Cancer arises through accumulation of somatically acquired genetic mutations. An important question is to delineate the temporal order of somatic mutations during carcinogenesis, which contributes to better understanding of cancer biology and facilitates identification of new therapeutic targets. Although a number of statistical and computational methods have been proposed to estimate the temporal order of mutations, they do not account for the differences in the functional impacts of mutations and thus are likely to be obscured by the presence of passenger mutations that do not contribute to cancer progression. In addition, many methods infer the order of mutations at the gene level, which have limited power due to the low mutation rate in most genes. RESULTS: In this paper, we develop a Probabilistic Approach for estimating the Temporal Order of Pathway mutations by leveraging functional Annotations of mutations (PATOPA). PATOPA infers the order of mutations at the pathway level, wherein it uses a probabilistic method to characterize the likelihood of mutational events from different pathways occurring in a certain order. The functional impact of each mutation is incorporated to weigh more on a mutation that is more integral to tumor development. A maximum likelihood method is used to estimate parameters and infer the probability of one pathway being mutated prior to another. Simulation studies and analysis of whole exome sequencing data from The Cancer Genome Atlas (TCGA) demonstrate that PATOPA is able to accurately estimate the temporal order of pathway mutations and provides new biological insights on carcinogenesis of colorectal and lung cancers. CONCLUSIONS: PATOPA provides a useful tool to estimate temporal order of mutations at the pathway level while leveraging functional annotations of mutations.

This paper presents a novel low cost and double node upset (DNU) online self-recoverable latch design using radiation hardening by design (RHBD) technology. The latch mainly consists of 8 interlocked input-split inverters. Since all internal nodes are interlocked, if any of the possible node pairs occurs a DNU, the latch can restore back. Simulation results have demonstrated the DNU online self-recoverability and also demonstrated that the proposed latch design saves 71.68% transmission delay, 72.92% power dissipation and 93.69% comprehensive delay-power-area product (DPAP) on average, compared with the up-to-date DNU online self-recoverable latch designs.

Reinforcement learning (RL) is a powerful paradigm for sequential decision-making under uncertainties, and most RL algorithms aim to maximize some numerical value which represents only one long-term objective. However, multiple long-term objectives are exhibited in many real-world decision and control systems, so recently there has been growing interest in solving multiobjective reinforcement learning (MORL) problems where there are multiple conflicting objectives. The aim of this paper is to present a comprehensive overview of MORL. The basic architecture, research topics, and naïve solutions of MORL are introduced at first. Then, several representative MORL approaches and some important directions of recent research are comprehensively reviewed. The relationships between MORL and other related research are also discussed, which include multiobjective optimization, hierarchical RL, and multiagent RL. Moreover, research challenges and open problems of MORL techniques are suggested.

Mining imbalanced data has recently received increasing attention due to its challenge and wide applications in the real world. Most of the existing work focuses on numerical data by manipulating the data structure which essentially changes the data characteristics or developing new distance or similarity measures which are designed for data with the so-called IID assumption, namely data is independent and identically distributed. This is not consistent with the real-life data and business needs, which request to fully respect the data structure and coupling relationships embedded in data objects, features and feature values. In this paper, we propose a novel coupled fuzzy similarity-based classification approach to cater for the difference between classes by a fuzzy membership and the couplings by coupled object similarity, and incorporate them into the most popular classifier: kNN to form a coupled fuzzy kNN (ie. CF-kNN). We test the approach on 14 categorical data sets compared to several kNN variants and classic classifiers including C4.5 and NaiveBayes. The experimental results show that CF-kNN outperforms the baselines, and those classifiers incorporated with the proposed coupled fuzzy similarity perform better than their original editions.

The state-of-the-art classification algorithms rarely consider the relationship between the attributes in the data sets and assume the attributes are independently to each other (IID). However, in real-world data, these attributes are more or less interacted via explicit or implicit relationships. Although the classifiers for class-balanced data are relatively well developed, the classification of class-imbalanced data is not straightforward, especially for mixed type data which has both categorical and numerical features. Limited research has been conducted on the class-imbalanced data. Some algorithms mainly synthesize or remove instances to force the sizes of each class comparable, which may change the inherent data structure or introduces noise to the source data. While for the distance or similarity based algorithms, they ignored the relationship between features when computing the similarity. This paper proposes a hybrid coupled k-nearest neighbor classification algorithm (HC-kNN) for mixed type data, by doing discretization on numerical features to adapt the inter coupling similarity as we do on categorical features, then combing this coupled similarity to the original similarity or distance, to overcome the shortcoming of the previous algorithms. The experiment results demonstrate that our proposed algorithm can get a higher average performance than that of the relevant algorithms (e.g. the variants of kNN, Decision Tree, SMOTE and NaiveBayes).

The robotic urban search and rescue (USAR) solution still has a lot of challenges, but it also has huge potentials to reduce the robot operator's in disaster scenes which are usually cluttered, dusty, and time-critical. In this paper, a Kinect-based 3D sensing and human action recognition solution is proposed to fuse 3D mapping, fine object reconstruction, and human identification/tracking into a Rich Information Map for evaluation. Preliminary experiments show the effectiveness of our approach's individual modules and are ready to integrate and perform robotic USAR tasks.

Approximate policy iteration (API) has been shown to be a class of reinforcement learning methods with stability and sample efficiency. However, sample collection is still an open problem which is critical to the performance of API methods. In this paper, a novel adaptive sample collection strategy using active learning-based exploration is proposed to enhance the performance of kernel-based API. In this strategy, an online kernel-based least squares policy iteration (KLSPI) method is adopted to construct nonlinear features and approximate the Q-function simultaneously. Therefore, more representative samples can be obtained for value function approximation. Simulation results on typical learning control problems illustrate that by using the proposed strategy, the performance of KLSPI can be improved remarkably.

In recent years, approximate policy iteration (API) has attracted increasing attention in reinforcement learning (RL), e.g., least-squares policy iteration (LSPI) and its kernelized version, the kernel-based LSPI algorithm. However, it remains difficult for API algorithms to obtain near-optimal policies for Markov decision processes (MDPs) with large or continuous state spaces. To address this problem, this paper presents a hierarchical API (HAPI) method with binary-tree state space decomposition for RL in a class of absorbing MDPs, which can be formulated as time-optimal learning control tasks. In the proposed method, after collecting samples adaptively in the state space of the original MDP, a learning-based decomposition strategy of sample sets was designed to implement the binary-tree state space decomposition process. Then, API algorithms were used on the sample subsets to approximate local optimal policies of sub-MDPs. The original MDP was decomposed into a binary-tree structure of absorbing sub-MDPs, constructed during the learning process, thus, local near-optimal policies were approximated by API algorithms with reduced complexity and higher precision. Furthermore, because of the improved quality of local policies, the combined global policy performed better than the near-optimal policy obtained by a single API algorithm in the original MDP. Three learning control problems, including path-tracking control of a real mobile robot, were studied to evaluate the performance of the HAPI method. With the same setting for basis function selection and sample collection, the proposed HAPI obtained better near-optimal policies than previous API methods such as LSPI and KLSPI.

The next-generation (4G) wireless communication systems focus on the seamless integration of various existing wireless communication networks, such as integrating the third-generation (3G) cellular networks and wireless local area network (WLAN). Resource management (RM) is one of the essential techniques to improve both overall resource utilization and quality-of-service (QoS) provisioning. This paper proposed a novel joint resource management scheme with QoS provisioning. A joint call admission policy is derived to support heterogeneous network architecture, service types, user mobility characteristics, and QoS levels. To relieve traffic congestion, an optimal channel searching and replacement algorithm (oCSR) and passive handoff techniques are further developed to balance total system traffic between WLAN and 3G cellular network as well as to reduce average system QoS cost, such as weighted blocking probability. Theoretical analysis and simulation results show that this scheme outperforms both traditional disjoint RM and joint RM without optimization.

This paper aims to provide Quality of Service (QoS) for IEEE 802.11 Wireless Local Area Network (WLAN). Specifically, we develop a new scheduling algorithm in link layer to support multimedia services with guaranteed QoS in WLAN. The proposed scheduling algorithm makes use of the idle system time to reduce average packet loss of real-time (RT) variable-bit-rate (VBR) services, while satisfying required transmission delay and bandwidth requirements of all RT services. Detailed analysis is given by taking into account several key statistical factors, such as service traffic model, the ratio between the polling period and the contention period (CP), burst generation probabilities. Performance is evaluated through OPNET simulation. Simulation results demonstrate that our proposed joint scheme is effective in terms of both system resource utilization and QoS provisioning.

This paper aims to provide quality of service (QoS) for IEEE 802.11 wireless local area network (WLAN) system integrated with UMTS cellular network. Due to heterogeneous QoS metrics in WLAN and cellular network, our scheme jointly considers the characteristics of the link and network layers. We adopt a scheduling algorithm called IIT (inserting idle time) in the link layer to reduce the packet loss ratio for variable-bit-rate services (VBR). Specifically, the IIT scheduling algorithm allocates the idle system time to VBR services to accommodate the bursts, while satisfying the required transmission delay and bandwidth of all RT services. Considering future WLAN coupled with UMTS cellular network, we propose a policy-based call admission control (CAC) scheme at the network layer. Based on different interworking policies, the proposed CAC provides the long-term QoS by limiting the number of admitted connections, while differentiating the services according to their priorities. Detailed analysis is given by taking into account several key statistical factors, such as traffic model, the ratio between the contention free period (CFP) and the contention period (CP), handoff rate. Performance is evaluated through simulation. The simulation results demonstrate that our proposed joint scheme is effective in terms of both system resource utilization and QoS provisioning

The seamless internetworking among various communication networks is in demand to provide anywhere, anytime connectivity with high data rate and enhanced service quality. In this paper, we propose a novel internetworking architecture between WLAN and UMTS cellular networks, hybrid coupling with radio access system (HCRAS), based on IPv6, mobile IP, and fast hand off techniques. Inheriting loose coupling internetworking techniques, HCRAS utilizes IEEE 802.16 standard air interface to set up direct wireless communication between base stations in UMTS cellular networks and local WLAN. In particular, we propose two new algorithms for intersystem communications and vertical handoff management in HCRAS system. The HCRAS can dynamically distribute traffic among the internetworking networks and reduce signaling cost and hand off latency significantly. Through mathematical analysis and numerical results, it is shown that HCRAS outperforms the tight-coupling and loose-coupling in various scenarios in terms of system routing efficiency, signaling cost, and handoff cost.

The paper proposes an interworking architecture, TCWA (tight coupling with wireless access), for integrating Universal Mobile Telecommunications System (UMTS) and wireless local area networks (WLAN). Based on a tight coupling interworking architecture, TCWA creates an additional wireless link between the base station of a UMTS cell and the gateway router of a WLAN located within this UMTS cell. In TCWA, signaling is routed according to the original tight coupling to preserve network security. However, the data traffic in the inter-network communications can be dynamically distributed as the added wireless link provides alternative routes. The issue of seamless vertical handoff in the TCWA architecture is also addressed in detail. To reduce handoff latency, a fast handoff algorithm is adopted. Based on detailed analysis and numerical results, we show that TCWA can reduce the traffic cost, relieve the burden of the core UMTS network, and enhance fault tolerance.

One of the critical and costly components in digital cellular communication systems is the RF power amplifier. Theoretically, one of the main concerns in an RF power amplifier design is the nonlinear effect of the amplifier. Quantitatively, no explicit relationship or expression currently exists between the out-of-band emission level and the nonlinearity description related to the third-order intercept point (IP/sub 3/). Further, in experiments and analysis, it was discovered that, in some situations, using IP/sub 3/ only is not accurate enough to describe the spectrum regrowth, especially when the fifth-order intercept point (IP/sub 5/) is relatively high compared to the third-order intermodulation. In this article, we analyze the nonlinear effect of an RF power amplifier in CDMA (IS-95 Standard), TDMA (IS-54 Standard) and MIRS M-16-QAM (Motorola MIRS standard) system, give expressions of the estimation of the out-of-band emission levels for each system expressed by power spectrum respectively, in terms of the IP/sub 3/ and the IP/sub 5/ as well as the power level of the signal. This result will be useful in the design of RF power amplifier for these wireless communication systems.

One of the critical and costly components in digital cellular communication systems is the RF power amplifier. Theoretically, one of the main concerns in an RF power amplifier design is the nonlinear effect of the amplifier. Quantitatively, so far, no explicit relationship or expression currently exists between the out-of-band emission level and the nonlinearity description related to the third-order intercept point (IP3). Further, in experiments and analysis, it was discovered that, in some situations, using IP3 only is not accurate enough to describe the spectrum regrowth, especially when the fifth-order intercept point (IP5) is relatively significant compared to the third-order intermodulation. In this article, we analyze the nonlinear effect of an RF power amplifier in the MIRS M-16-QAM system, give an expression of the estimation of the out-of-band emission levels for an MIRS signal power spectrum in terms of the IP3 and the IP5, as well as the power level of the signal. This result will be useful in the design of RF power amplifier for the MIRS M-16-QAM system and other wireless communication systems.

One of the critical and costly components in digital cellular communication systems is the RF power amplifier. Theoretically, one of the main concerns in an RF power amplifier design is the nonlinear effect of the amplifier. Quantitatively, so far, no explicit relationship or expression currently exists between the out-of-band emission level and the nonlinearity description related to the third-order intercept point ( IP/sub 3/). Further, in experiments and analysis, it was discovered that, in some situations, using IP/sub 3/ only is not accurate enough to describe the spectrum regrowth, especially when the fifth-order intercept point ( IP/sub 5/) is relatively significant compared to the third-order intermodulation. We analyze the nonlinear effect of an RE power amplifier in a TDMA (IS-54 standard) system, give an expression of the estimation of the out-of-band emission levels for a TDMA power spectrum in terms of the IP/sub 3/ and the IP/sub 5/, as well as the power level of the signal. This result will be useful in the design of an RF power amplifier for a TDMA wireless system.