Search

Vertical federated learning is a natural and elegant approach to integrate multi-view data vertically partitioned across devices (clients) while preserving their privacies. Apart from the model training, existing methods requires the collaboration of all clients in the model inference. However, the model inference is probably maintained for service in a long time, while the collaboration, especially when the clients belong to different organizations, is unpredictable in real-world scenarios, such as concellation of contract, network unavailablity, etc., resulting in the failure of them. To address this issue, we, at the first attempt, propose a flexible Active-Passive Federated learning (APFed) framework. Specifically, the active client is the initiator of a learning task and responsible to build the complete model, while the passive clients only serve as assistants. Once the model built, the active client can make inference independently. In addition, we instance the APFed framework into two classification methods with employing the reconstruction loss and the contrastive loss on passive clients, respectively. Meanwhile, the two methods are tested in a set of experiments and achieves desired results, validating their effectiveness.

Evolutionary Multi-task Optimization (EMTO) is a paradigm that leverages knowledge transfer across simultaneously optimized tasks for enhanced search performance. To facilitate EMTO's performance, various knowledge transfer models have been developed for specific optimization tasks. However, designing these models often requires substantial expert knowledge. Recently, large language models (LLMs) have achieved remarkable success in autonomous programming, aiming to produce effective solvers for specific problems. In this work, a LLM-based optimization paradigm is introduced to establish an autonomous model factory for generating knowledge transfer models, ensuring effective and efficient knowledge transfer across various optimization tasks. To evaluate the performance of the proposed method, we conducted comprehensive empirical studies comparing the knowledge transfer model generated by the LLM with existing state-of-the-art knowledge transfer methods. The results demonstrate that the generated model is able to achieve superior or competitive performance against hand-crafted knowledge transfer models in terms of both efficiency and effectiveness., Comment: 10 pages, 11 pages

Though current long-context large language models (LLMs) have demonstrated impressive capacities in answering user questions based on extensive text, the lack of citations in their responses makes user verification difficult, leading to concerns about their trustworthiness due to their potential hallucinations. In this work, we aim to enable long-context LLMs to generate responses with fine-grained sentence-level citations, improving their faithfulness and verifiability. We first introduce LongBench-Cite, an automated benchmark for assessing current LLMs' performance in Long-Context Question Answering with Citations (LQAC), revealing considerable room for improvement. To this end, we propose CoF (Coarse to Fine), a novel pipeline that utilizes off-the-shelf LLMs to automatically generate long-context QA instances with precise sentence-level citations, and leverage this pipeline to construct LongCite-45k, a large-scale SFT dataset for LQAC. Finally, we train LongCite-8B and LongCite-9B using the LongCite-45k dataset, successfully enabling their generation of accurate responses and fine-grained sentence-level citations in a single output. The evaluation results on LongBench-Cite show that our trained models achieve state-of-the-art citation quality, surpassing advanced proprietary models including GPT-4o.

Vision-language navigation (VLN) is a critical domain within embedded intelligence, requiring agents to navigate 3D environments based on natural language instructions. Traditional VLN research has focused on improving environmental understanding and decision accuracy. However, these approaches often exhibit a significant performance gap when agents are deployed in novel environments, mainly due to the limited diversity of training data. Expanding datasets to cover a broader range of environments is impractical and costly. We propose the Vision-Language Navigation with Continual Learning (VLNCL) paradigm to address this challenge. In this paradigm, agents incrementally learn new environments while retaining previously acquired knowledge. VLNCL enables agents to maintain an environmental memory and extract relevant knowledge, allowing rapid adaptation to new environments while preserving existing information. We introduce a novel dual-loop scenario replay method (Dual-SR) inspired by brain memory replay mechanisms integrated with VLN agents. This method facilitates consolidating past experiences and enhances generalization across new tasks. By utilizing a multi-scenario memory buffer, the agent efficiently organizes and replays task memories, thereby bolstering its ability to adapt quickly to new environments and mitigating catastrophic forgetting. Our work pioneers continual learning in VLN agents, introducing a novel experimental setup and evaluation metrics. We demonstrate the effectiveness of our approach through extensive evaluations and establish a benchmark for the VLNCL paradigm. Comparative experiments with existing continual learning and VLN methods show significant improvements, achieving state-of-the-art performance in continual learning ability and highlighting the potential of our approach in enabling rapid adaptation while preserving prior knowledge.

Recent advancements in large language models (LLMs) with extended context windows have significantly improved tasks such as information extraction, question answering, and complex planning scenarios. In order to achieve success in long context tasks, a large amount of work has been done to enhance the long context capabilities of the model through synthetic data. Existing methods typically utilize the Self-Instruct framework to generate instruction tuning data for better long context capability improvement. However, our preliminary experiments indicate that less than 35% of generated samples are multi-hop, and more than 40% exhibit poor quality, limiting comprehensive understanding and further research. To improve the quality of synthetic data, we propose the Multi-agent Interactive Multi-hop Generation (MIMG) framework, incorporating a Quality Verification Agent, a Single-hop Question Generation Agent, a Multiple Question Sampling Strategy, and a Multi-hop Question Merger Agent. This framework improves the data quality, with the proportion of high-quality, multi-hop, and diverse data exceeding 85%. Furthermore, we systematically investigate strategies for document selection, question merging, and validation techniques through extensive experiments across various models. Our findings show that our synthetic high-quality long-context instruction data significantly enhances model performance, even surpassing models trained on larger amounts of human-annotated data. Our code is available at: https://github.com/WowCZ/LongMIT., Comment: Work in progress

Despite recent advances in stereo matching, the extension to intricate underwater settings remains unexplored, primarily owing to: 1) the reduced visibility, low contrast, and other adverse effects of underwater images; 2) the difficulty in obtaining ground truth data for training deep learning models, i.e. simultaneously capturing an image and estimating its corresponding pixel-wise depth information in underwater environments. To enable further advance in underwater stereo matching, we introduce a large synthetic dataset called UWStereo. Our dataset includes 29,568 synthetic stereo image pairs with dense and accurate disparity annotations for left view. We design four distinct underwater scenes filled with diverse objects such as corals, ships and robots. We also induce additional variations in camera model, lighting, and environmental effects. In comparison with existing underwater datasets, UWStereo is superior in terms of scale, variation, annotation, and photo-realistic image quality. To substantiate the efficacy of the UWStereo dataset, we undertake a comprehensive evaluation compared with nine state-of-the-art algorithms as benchmarks. The results indicate that current models still struggle to generalize to new domains. Hence, we design a new strategy that learns to reconstruct cross domain masked images before stereo matching training and integrate a cross view attention enhancement module that aggregates long-range content information to enhance the generalization ability., Comment: 12pages

Inspired by the complexity of certain real-world datasets, this article introduces a novel flexible linear spline index regression model. The model posits piecewise linear effects of an index on the response, with continuous changes occurring at knots. Significantly, it possesses the interpretability of linear models, captures nonlinear effects similar to nonparametric models, and achieves dimension reduction like single-index models. In addition, the locations and number of knots remain unknown, which further enhances the adaptability of the model in practical applications. We propose a new method that combines penalized approaches and convolution techniques to simultaneously estimate the unknown parameters and determine the number of knots. Noteworthy is that the proposed method allows the number of knots to diverge with the sample size. We demonstrate that the proposed estimators can identify the number of knots with a probability approaching one and estimate the coefficients as efficiently as if the number of knots is known in advance. We also introduce a procedure to test the presence of knots. Simulation studies and two real datasets are employed to assess the finite sample performance of the proposed method., Comment: 84 pages, 4 figures

Currently, most applications in the Industrial Internet of Things (IIoT) still rely on CNN-based neural networks. Although Transformer-based large models (LMs), including language, vision, and multimodal models, have demonstrated impressive capabilities in AI-generated content (AIGC), their application in industrial domains, such as detection, planning, and control, remains relatively limited. Deploying pre-trained LMs in industrial environments often encounters the challenge of stability and plasticity due to the complexity of tasks, the diversity of data, and the dynamic nature of user demands. To address these challenges, the pre-training and fine-tuning strategy, coupled with continual learning, has proven to be an effective solution, enabling models to adapt to dynamic demands while continuously optimizing their inference and decision-making capabilities. This paper surveys the integration of LMs into IIoT-enhanced General Industrial Intelligence (GII), focusing on two key areas: LMs for GII and LMs on GII. The former focuses on leveraging LMs to provide optimized solutions for industrial application challenges, while the latter investigates continuous optimization of LMs learning and inference capabilities in collaborative scenarios involving industrial devices, edge computing, and cloud computing. This paper provides insights into the future development of GII, aiming to establish a comprehensive theoretical framework and research direction for GII, thereby advancing GII towards a more general and adaptive future.

In this study, we explore the application of Physics-Informed Neural Networks (PINNs) to the analysis of bifurcation phenomena in ecological migration models. By integrating the fundamental principles of diffusion-advection-reaction equations with deep learning techniques, we address the complexities of species migration dynamics, particularly focusing on the detection and analysis of Hopf bifurcations. Traditional numerical methods for solving partial differential equations (PDEs) often involve intricate calculations and extensive computational resources, which can be restrictive in high-dimensional problems. In contrast, PINNs offer a more flexible and efficient alternative, bypassing the need for grid discretization and allowing for mesh-free solutions. Our approach leverages the DeepXDE framework, which enhances the computational efficiency and applicability of PINNs in solving high-dimensional PDEs. We validate our results against conventional methods and demonstrate that PINNs not only provide accurate bifurcation predictions but also offer deeper insights into the underlying dynamics of diffusion processes. Despite these advantages, the study also identifies challenges such as the high computational costs and the sensitivity of PINN performance to network architecture and hyperparameter settings. Future work will focus on optimizing these algorithms and expanding their application to other complex systems involving bifurcations. The findings from this research have significant implications for the modeling and analysis of ecological systems, providing a powerful tool for predicting and understanding complex dynamical behaviors.

Background and Objective: Diabetes presents a significant challenge to healthcare due to the negative impact of poor blood sugar control on health and associated complications. Computer simulation platforms, notably exemplified by the UVA/Padova Type 1 Diabetes simulator, has emerged as a promising tool for advancing diabetes treatments by simulating patient responses in a virtual environment. The UVA Virtual Lab (UVLab) is a new simulation platform to mimic the metabolic behavior of people with Type 2 diabetes (T2D) with a large population of 6062 virtual subjects. Methods: The work introduces the Distribution-Based Population Selection (DSPS) method, a systematic approach to identifying virtual subsets that mimic the clinical behavior observed in real trials. The method transforms the sub-population selection task into a Linear Programing problem, enabling the identification of the largest representative virtual cohort. This selection process centers on key clinical outcomes in diabetes research, such as HbA1c and Fasting plasma Glucose (FPG), ensuring that the statistical properties (moments) of the selected virtual sub-population closely resemble those observed in real-word clinical trial. Results: DSPS method was applied to the insulin degludec (IDeg) arm of a phase 3 clinical trial. This method was used to select a sub-population of virtual subjects that closely mirrored the clinical trial data across multiple key metrics, including glycemic efficacy, insulin dosages, and cumulative hypoglycemia events over a 26-week period. Conclusion: The DSPS algorithm is able to select virtual sub-population within UVLab to reproduce and predict the outcomes of a clinical trial. This statistical method can bridge the gap between large population simulation platforms and previously conducted clinical trials.

Egocentric human pose estimation (HPE) using wearable sensors is essential for VR/AR applications. Most methods rely solely on either egocentric-view images or sparse Inertial Measurement Unit (IMU) signals, leading to inaccuracies due to self-occlusion in images or the sparseness and drift of inertial sensors. Most importantly, the lack of real-world datasets containing both modalities is a major obstacle to progress in this field. To overcome the barrier, we propose EMHI, a multimodal \textbf{E}gocentric human \textbf{M}otion dataset with \textbf{H}ead-Mounted Display (HMD) and body-worn \textbf{I}MUs, with all data collected under the real VR product suite. Specifically, EMHI provides synchronized stereo images from downward-sloping cameras on the headset and IMU data from body-worn sensors, along with pose annotations in SMPL format. This dataset consists of 885 sequences captured by 58 subjects performing 39 actions, totaling about 28.5 hours of recording. We evaluate the annotations by comparing them with optical marker-based SMPL fitting results. To substantiate the reliability of our dataset, we introduce MEPoser, a new baseline method for multimodal egocentric HPE, which employs a multimodal fusion encoder, temporal feature encoder, and MLP-based regression heads. The experiments on EMHI show that MEPoser outperforms existing single-modal methods and demonstrates the value of our dataset in solving the problem of egocentric HPE. We believe the release of EMHI and the method could advance the research of egocentric HPE and expedite the practical implementation of this technology in VR/AR products.

Learned image compression have attracted considerable interests in recent years. It typically comprises an analysis transform, a synthesis transform, quantization and an entropy coding model. The analysis transform and synthesis transform are used to encode an image to latent feature and decode the quantized feature to reconstruct the image, and can be regarded as coupled transforms. However, the analysis transform and synthesis transform are designed independently in the existing methods, making them unreliable in high-quality image compression. Inspired by the invertible neural networks in generative modeling, invertible modules are used to construct the coupled analysis and synthesis transforms. Considering the noise introduced in the feature quantization invalidates the invertible process, this paper proposes an Approximately Invertible Neural Network (A-INN) framework for learned image compression. It formulates the rate-distortion optimization in lossy image compression when using INN with quantization, which differentiates from using INN for generative modelling. Generally speaking, A-INN can be used as the theoretical foundation for any INN based lossy compression method. Based on this formulation, A-INN with a progressive denoising module (PDM) is developed to effectively reduce the quantization noise in the decoding. Moreover, a Cascaded Feature Recovery Module (CFRM) is designed to learn high-dimensional feature recovery from low-dimensional ones to further reduce the noise in feature channel compression. In addition, a Frequency-enhanced Decomposition and Synthesis Module (FDSM) is developed by explicitly enhancing the high-frequency components in an image to address the loss of high-frequency information inherent in neural network based image compression. Extensive experiments demonstrate that the proposed A-INN outperforms the existing learned image compression methods.

We develop CNIMA (Chinese Non-Native Interactivity Measurement and Automation), a Chinese-as-a-second-language labelled dataset with 10K dialogues. We annotate CNIMA using an evaluation framework -- originally introduced for English-as-a-second-language dialogues -- that assesses micro-level features (e.g.\ backchannels) and macro-level interactivity labels (e.g.\ topic management) and test the framework's transferability from English to Chinese. We found the framework robust across languages and revealed universal and language-specific relationships between micro-level and macro-level features. Next, we propose an approach to automate the evaluation and find strong performance, creating a new tool for automated second language assessment. Our system can be adapted to other languages easily as it uses large language models and as such does not require large-scale annotated training data.

Beginning with VisualGLM and CogVLM, we are continuously exploring VLMs in pursuit of enhanced vision-language fusion, efficient higher-resolution architecture, and broader modalities and applications. Here we propose the CogVLM2 family, a new generation of visual language models for image and video understanding including CogVLM2, CogVLM2-Video and GLM-4V. As an image understanding model, CogVLM2 inherits the visual expert architecture with improved training recipes in both pre-training and post-training stages, supporting input resolution up to $1344 \times 1344$ pixels. As a video understanding model, CogVLM2-Video integrates multi-frame input with timestamps and proposes automated temporal grounding data construction. Notably, CogVLM2 family has achieved state-of-the-art results on benchmarks like MMBench, MM-Vet, TextVQA, MVBench and VCGBench. All models are open-sourced in https://github.com/THUDM/CogVLM2 and https://github.com/THUDM/GLM-4, contributing to the advancement of the field.

Detecting locally, non-overlapping, near-clique densest subgraphs is a crucial problem for community search in social networks. As a vertex may be involved in multiple overlapped local cliques, detecting locally densest sub-structures considering h-clique density, i.e., locally h-clique densest subgraph (LhCDS) attracts great interests. This paper investigates the LhCDS detection problem and proposes an efficient and exact algorithm to list the top-k non-overlapping, locally h-clique dense, and compact subgraphs. We in particular jointly consider h-clique compact number and LhCDS and design a new "Iterative Propose-Prune-and-Verify" pipeline (IPPV) for top-k LhCDS detection. (1) In the proposal part, we derive initial bounds for h-clique compact numbers; prove the validity, and extend a convex programming method to tighten the bounds for proposing LhCDS candidates without missing any. (2) Then a tentative graph decomposition method is proposed to solve the challenging case where a clique spans multiple subgraphs in graph decomposition. (3) To deal with the verification difficulty, both a basic and a fast verification method are proposed, where the fast method constructs a smaller-scale flow network to improve efficiency while preserving the verification correctness. The verified LhCDSes are returned, while the candidates that remained unsure reenter the IPPV pipeline. (4) We further extend the proposed methods to locally more general pattern densest subgraph detection problems. We prove the exactness and low complexity of the proposed algorithm. Extensive experiments on real datasets show the effectiveness and high efficiency of IPPV., Comment: This paper has been accepted by SIGMOD 2025

Model-based Offline Reinforcement Learning trains policies based on offline datasets and model dynamics, without direct real-world environment interactions. However, this method is inherently challenged by distribution shift. Previous approaches have primarily focused on tackling this issue directly leveraging off-policy mechanisms and heuristic uncertainty in model dynamics, but they resulted in inconsistent objectives and lacked a unified theoretical foundation. This paper offers a comprehensive analysis that disentangles the problem into two key components: model bias and policy shift. We provide both theoretical insights and empirical evidence to demonstrate how these factors lead to inaccuracies in value function estimation and impose implicit restrictions on policy learning. To address these challenges, we derive adjustment terms for model bias and policy shift within a unified probabilistic inference framework. These adjustments are seamlessly integrated into the vanilla reward function to create a novel Shifts-aware Reward (SAR), aiming at refining value learning and facilitating policy training. Furthermore, we introduce Shifts-aware Model-based Offline Reinforcement Learning (SAMBO-RL), a practical framework that efficiently trains classifiers to approximate the SAR for policy optimization. Empirically, we show that SAR effectively mitigates distribution shift, and SAMBO-RL demonstrates superior performance across various benchmarks, underscoring its practical effectiveness and validating our theoretical analysis.

Learning meaningful local and global information remains a challenge in point cloud segmentation tasks. When utilizing local information, prior studies indiscriminately aggregates neighbor information from different classes to update query points, potentially compromising the distinctive feature of query points. In parallel, inaccurate modeling of long-distance contextual dependencies when utilizing global information can also impact model performance. To address these issues, we propose GSTran, a novel transformer network tailored for the segmentation task. The proposed network mainly consists of two principal components: a local geometric transformer and a global semantic transformer. In the local geometric transformer module, we explicitly calculate the geometric disparity within the local region. This enables amplifying the affinity with geometrically similar neighbor points while suppressing the association with other neighbors. In the global semantic transformer module, we design a multi-head voting strategy. This strategy evaluates semantic similarity across the entire spatial range, facilitating the precise capture of contextual dependencies. Experiments on ShapeNetPart and S3DIS benchmarks demonstrate the effectiveness of the proposed method, showing its superiority over other algorithms. The code is available at https://github.com/LAB123-tech/GSTran., Comment: ICPR 2024

Occupancy prediction has attracted intensive attention and shown great superiority in the development of autonomous driving systems. The fine-grained environmental representation brought by occupancy prediction in terms of both geometry and semantic information has facilitated the general perception and safe planning under open scenarios. However, it also brings high computation costs and heavy parameters in existing works that utilize voxel-based 3d dense representation and Transformer-based quadratic attention. To address these challenges, in this paper, we propose a Mamba-based occupancy prediction method (MambaOcc) adopting BEV features to ease the burden of 3D scenario representation, and linear Mamba-style attention to achieve efficient long-range perception. Besides, to address the sensitivity of Mamba to sequence order, we propose a local adaptive reordering (LAR) mechanism with deformable convolution and design a hybrid BEV encoder comprised of convolution layers and Mamba. Extensive experiments on the Occ3D-nuScenes dataset demonstrate that MambaOcc achieves state-of-the-art performance in terms of both accuracy and computational efficiency. For example, compared to FlashOcc, MambaOcc delivers superior results while reducing the number of parameters by 42\% and computational costs by 39\%. Code will be available at https://github.com/Hub-Tian/MambaOcc.

In order to explore how the ubiquitous stochastic low-frequency (SLF) variability of O-type stars is related to various stellar characteristics, we compiled a sample of 150 O-type stars observed via ground-based spectroscopic surveys, alongside photometric data obtained from the Transiting Exoplanet Survey Satellite (TESS). We analyzed 298 light curves obtained from TESS sectors 1-65 for the stars in our sample. Leveraging the spectroscopic parameters, we used Bonnsai to determine masses, radii, fractional main sequence ages, and mass-loss rates for stars of our sample. Subsequently, we identified possible correlations between the fitted parameters of SLF variability and stellar properties. Our analysis unveiled four significant correlations between the amplitude and stellar parameters, including mass, radius, fractional main sequence ages, and mass-loss rate. For stars with $M \gtrsim 30 M_{\odot}$, we observed a decrease in characteristic frequency and steepness with increasing radius. Finally, we compared various physical processes that may account for the SLF variability with our results. The observed SLF variability may arise from the combined effects of FeCZ and IGWs, with IGWs potentially more dominant in the early stages of stellar evolution, and the contribution of FeCZ becoming more significant as stars evolve. Meanwhile, our results indicate that the SLF variability of O-type stars bears certain signatures of the line-driven wind instability and granulation., Comment: 40 pages, 12 figures,accepted by Astrophysical Journal Supplement Series

Generative adversarial networks (GANs) have made impressive advances in image generation, but they often require large-scale training data to avoid degradation caused by discriminator overfitting. To tackle this issue, we investigate the challenge of training GANs with limited data, and propose a novel regularization method based on the idea of renormalization group (RG) in physics.We observe that in the limited data setting, the gradient pattern that the generator obtains from the discriminator becomes more aggregated over time. In RG context, this aggregated pattern exhibits a high discrepancy from its coarse-grained versions, which implies a high-capacity and sensitive system, prone to overfitting and collapse. To address this problem, we introduce a \textbf{m}ulti-\textbf{s}cale \textbf{s}tructural \textbf{s}elf-\textbf{d}issimilarity (MS$^3$D) regularization, which constrains the gradient field to have a consistent pattern across different scales, thereby fostering a more redundant and robust system. We show that our method can effectively enhance the performance and stability of GANs under limited data scenarios, and even allow them to generate high-quality images with very few data.

Molecular property prediction is a crucial foundation for drug discovery. In recent years, pre-trained deep learning models have been widely applied to this task. Some approaches that incorporate prior biological domain knowledge into the pre-training framework have achieved impressive results. However, these methods heavily rely on biochemical experts, and retrieving and summarizing vast amounts of domain knowledge literature is both time-consuming and expensive. Large Language Models (LLMs) have demonstrated remarkable performance in understanding and efficiently providing general knowledge. Nevertheless, they occasionally exhibit hallucinations and lack precision in generating domain-specific knowledge. Conversely, Domain-specific Small Models (DSMs) possess rich domain knowledge and can accurately calculate molecular domain-related metrics. However, due to their limited model size and singular functionality, they lack the breadth of knowledge necessary for comprehensive representation learning. To leverage the advantages of both approaches in molecular property prediction, we propose a novel Molecular Graph representation learning framework that integrates Large language models and Domain-specific small models (MolGraph-LarDo). Technically, we design a two-stage prompt strategy where DSMs are introduced to calibrate the knowledge provided by LLMs, enhancing the accuracy of domain-specific information and thus enabling LLMs to generate more precise textual descriptions for molecular samples. Subsequently, we employ a multi-modal alignment method to coordinate various modalities, including molecular graphs and their corresponding descriptive texts, to guide the pre-training of molecular representations. Extensive experiments demonstrate the effectiveness of the proposed method.

Integrating large language models (LLMs) into autonomous driving enhances personalization and adaptability in open-world scenarios. However, traditional edge computing models still face significant challenges in processing complex driving data, particularly regarding real-time performance and system efficiency. To address these challenges, this study introduces EC-Drive, a novel edge-cloud collaborative autonomous driving system with data drift detection capabilities. EC-Drive utilizes drift detection algorithms to selectively upload critical data, including new obstacles and traffic pattern changes, to the cloud for processing by GPT-4, while routine data is efficiently managed by smaller LLMs on edge devices. This approach not only reduces inference latency but also improves system efficiency by optimizing communication resource use. Experimental validation confirms the system's robust processing capabilities and practical applicability in real-world driving conditions, demonstrating the effectiveness of this edge-cloud collaboration framework. Our data and system demonstration will be released at https://sites.google.com/view/ec-drive.

Prostate cancer is the second most common cancer in males worldwide, and mpMRI is commonly used for diagnosis. However, interpreting mpMRI is challenging and requires expertise from radiologists. This highlights the urgent need for automated grading in mpMRI. Existing studies lack integration of clinical prior information and suffer from uneven training sample distribution due to prevalence. Therefore, we propose a solution that incorporates prior knowledge, addresses the issue of uneven medical sample distribution, and maintains high interpretability in mpMRI. Firstly, we introduce Prior Knowledge-Based Feature Extraction, which mathematically models the PI-RADS criteria for prostate cancer as diagnostic information into model training. Secondly, we propose Adaptive Recall Feedback Loss to address the extremely imbalanced data problem. This method adjusts the training dynamically based on accuracy and recall in the validation set, resulting in high accuracy and recall simultaneously in the testing set.Thirdly, we design an Enhanced Cascade Prostate Cancer Classifier that classifies prostate cancer into different levels in an interpretable way, which refines the classification results and helps with clinical intervention. Our method is validated through experiments on the PI-CAI dataset and outperforms other methods with a more balanced result in both accuracy and recall rate.

A notable problem within permutation patterns that has attracted considerable attention in literature since 1973 is the search for a bijective proof demonstrating that 123-avoiding and 132-avoiding permutations are equinumerous, both counted by the Catalan numbers. Despite this equivalence, the distributions of occurrences of the patterns 123 and 132 are distinct. When considering 123 and 132 as mesh patterns and selectively shading boxes, similar scenarios arise, even when avoidance is defined by the Bell numbers or other sequences, rather than the Catalan numbers. However, computer experiments suggest that mesh patterns 123 and 132 may indeed be equidistributed. Furthermore, by considering symmetric shadings relative to the anti-diagonal, a maximum of 93 such equidistributed pairs can potentially exist. This paper establishes 75 such equidistributions, leaving the justification of the remaining cases as open problems. As a by-product, we also prove 36 relevant non-symmetric equidistributions. All our proofs are bijective and involve swapping occurrences of the patterns in question, thereby demonstrating their joint equidistribution. Our findings are a continuation of the systematic study of distributions of short-length mesh patterns initiated by Kitaev and Zhang in 2019.

In a multi-agent system (MAS), action semantics indicates the different influences of agents' actions toward other entities, and can be used to divide agents into groups in a physically heterogeneous MAS. Previous multi-agent reinforcement learning (MARL) algorithms apply global parameter-sharing across different types of heterogeneous agents without careful discrimination of different action semantics. This common implementation decreases the cooperation and coordination between agents in complex situations. However, fully independent agent parameters dramatically increase the computational cost and training difficulty. In order to benefit from the usage of different action semantics while also maintaining a proper parameter-sharing structure, we introduce the Unified Action Space (UAS) to fulfill the requirement. The UAS is the union set of all agent actions with different semantics. All agents first calculate their unified representation in the UAS, and then generate their heterogeneous action policies using different available-action-masks. To further improve the training of extra UAS parameters, we introduce a Cross-Group Inverse (CGI) loss to predict other groups' agent policies with the trajectory information. As a universal method for solving the physically heterogeneous MARL problem, we implement the UAS adding to both value-based and policy-based MARL algorithms, and propose two practical algorithms: U-QMIX and U-MAPPO. Experimental results in the SMAC environment prove the effectiveness of both U-QMIX and U-MAPPO compared with several state-of-the-art MARL methods.

Open-source software (OSS) has experienced a surge in popularity, attributed to its collaborative development model and cost-effective nature. However, the adoption of specific software versions in development projects may introduce security risks when these versions bring along vulnerabilities. Current methods of identifying vulnerable versions typically analyze and trace the code involved in vulnerability patches using static analysis with pre-defined rules. They then use syntactic-level code clone detection to identify the vulnerable versions. These methods are hindered by imprecisions due to (1) the inclusion of vulnerability-irrelevant code in the analysis and (2) the inadequacy of syntactic-level code clone detection. This paper presents Vercation, an approach designed to identify vulnerable versions of OSS written in C/C++. Vercation combines program slicing with a Large Language Model (LLM) to identify vulnerability-relevant code from vulnerability patches. It then backtraces historical commits to gather previous modifications of identified vulnerability-relevant code. We propose semantic-level code clone detection to compare the differences between pre-modification and post-modification code, thereby locating the vulnerability-introducing commit (vic) and enabling to identify the vulnerable versions between the patch commit and the vic. We curate a dataset linking 74 OSS vulnerabilities and 1013 versions to evaluate Vercation. On this dataset, our approach achieves the F1 score of 92.4%, outperforming current state-of-the-art methods. More importantly, Vercation detected 134 incorrect vulnerable OSS versions in NVD reports.

Speculative decoding (SD), where an extra draft model is employed to provide multiple \textit{draft} tokens first and then the original target model verifies these tokens in parallel, has shown great power for LLM inference acceleration. However, existing SD methods suffer from the mutual waiting problem, i.e., the target model gets stuck when the draft model is \textit{guessing} tokens, and vice versa. This problem is directly incurred by the asynchronous execution of the draft model and the target model, and is exacerbated due to the fixed draft length in speculative decoding. To address these challenges, we propose a conceptually simple, flexible, and general framework to boost speculative decoding, namely \textbf{P}arallel sp\textbf{E}culative decoding with \textbf{A}daptive d\textbf{R}aft \textbf{L}ength (PEARL). Specifically, PEARL proposes \textit{pre-verify} to verify the first draft token in advance during the drafting phase, and \textit{post-verify} to generate more draft tokens during the verification phase. PEARL parallels the drafting phase and the verification phase via applying the two strategies, and achieves adaptive draft length for different scenarios, which effectively alleviates the mutual waiting problem. Moreover, we theoretically demonstrate that the mean accepted tokens of PEARL is more than existing \textit{draft-then-verify} works. Experiments on various text generation benchmarks demonstrate the effectiveness of our \name, leading to a superior speedup performance up to \textbf{3.79$\times$} and \textbf{1.52$\times$}, compared to auto-regressive decoding and vanilla speculative decoding, respectively.

This paper proposes a novel consensus-on-only-measurement distributed filter over directed graphs under the collectively observability condition. First, the distributed filter structure is designed with an augmented leader-following measurement fusion strategy. Subsequently, two parameter design methods are presented, and the consensus gain parameter is devised utilizing local information exclusively rather than global information. Additionally, the lower bound of the fusion step is derived to guarantee a uniformly upper bound of the estimation error covariance. Moreover, the lower bounds of the convergence rates of the steady-state performance gap between the proposed algorithm and the centralized filter are provided with the fusion step approaching infinity. The analysis demonstrates that the convergence rate is, at a minimum, as rapid as exponential convergence under the spectral norm condition of the communication graph. The transient performance is also analyzed with the fusion step tending to infinity. The inherent trade-off between the communication cost and the filtering performance is revealed from the analysis of the steady-state performance and the transient performance. Finally, the theoretical results are substantiated through the validation of two simulation examples.

Current long context large language models (LLMs) can process inputs up to 100,000 tokens, yet struggle to generate outputs exceeding even a modest length of 2,000 words. Through controlled experiments, we find that the model's effective generation length is inherently bounded by the sample it has seen during supervised fine-tuning (SFT). In other words, their output limitation is due to the scarcity of long-output examples in existing SFT datasets. To address this, we introduce AgentWrite, an agent-based pipeline that decomposes ultra-long generation tasks into subtasks, enabling off-the-shelf LLMs to generate coherent outputs exceeding 20,000 words. Leveraging AgentWrite, we construct LongWriter-6k, a dataset containing 6,000 SFT data with output lengths ranging from 2k to 32k words. By incorporating this dataset into model training, we successfully scale the output length of existing models to over 10,000 words while maintaining output quality. We also develop LongBench-Write, a comprehensive benchmark for evaluating ultra-long generation capabilities. Our 9B parameter model, further improved through DPO, achieves state-of-the-art performance on this benchmark, surpassing even much larger proprietary models. In general, our work demonstrates that existing long context LLM already possesses the potential for a larger output window--all you need is data with extended output during model alignment to unlock this capability. Our code & models are at: https://github.com/THUDM/LongWriter.