Although Large language Model (LLM)-powered information extraction (IE) systems have shown impressive capabilities, current fine-tuning paradigms face two major limitations: high training costs and difficulties in aligning with LLM preferences. To address these issues, we propose a novel universal IE paradigm—the Self-Correcting Iterative Refinement (SCIR) framework—along with a Multi-task Bilingual (Chinese-English) Self-Correcting (MBSC) dataset containing over 100,000 entries. The SCIR framework achieves plug-and-play compatibility with existing LLMs and IE systems through its Dual-Path Self-Correcting module and feedback-driven optimization, thereby significantly reducing training costs. Concurrently, the MBSC dataset tackles the challenge of preference alignment by indirectly distilling GPT-4's capabilities into IE result detection models. Experimental results demonstrate that SCIR outperforms state-of-the-art IE methods across three key tasks— named entity recognition, relation extraction, and event extraction—achieving a 5.27 percent average improvement in span-based Micro-F1 while reducing training costs by 87 percent compared to baseline approaches. These advancements not only enhance the flexibility and accuracy of IE systems but also pave the way for lightweight and efficient IE paradigms.

Large Language Models (LLMs) have shown strong potential for text-attributed graph (TAG) learning, yet effectively integrating LLM semantics with graph structural information remains challenging.Embeddings obtained from frozen LLMs lack topology awareness, while fine-tuning LLMs is often computationally expensive.Moreover, LLM embeddings are highdimensional, and naively reducing dimensionality tends to destroy semantics.To address these issues, we propose GASE, a framework for learning Graph-Aware Semantic Embeddings using frozen LLMs.GASE consists of two key stages: First, we introduce a Training-Free Structure-Aware Semantic Extraction (TSSE) module.Through inter-layer semantic feedback and progressive masked attention, it efficiently compresses and propagates semantic context from neighboring nodes without updating LLM parameters.Second, we propose a Subspace Decomposition and Structural Injection (SDSI) strategy.Embeddings obtained from TSSE are decomposed into a semantic-rich subspace and a structural injection subspace, and structural signals are injected into the latter, which preserves original semantics while integrating graph information.Experiments demonstrate that GASE outperforms state-of-the-art baselines on node classification and achieves a 5× speedup over fine-tuning-based methods.

Multi-step reasoning remains challenging for language models with limited capacity.While recent reasoning distillation approaches transfer chain-of-thought supervision from large teacher models, they typically apply uniform supervision across all Transformer components, overlooking the fact that different modules contribute unequally to reasoning.We propose Module-Aware Reasoning Distillation, a parameter-efficient framework that explicitly targets key Transformer components for effective reasoning transfer.Through systematic analysis, we identify the feed-forward network projections and the output projection of self-attention as primary bottlenecks for reasoning.Based on these findings, we introduce lightweight adapter modules at these components while freezing the backbone parameters, enabling focused and efficient distillation.Our approach adopts an offline distillation setting, where a strong teacher model provides reasoning trajectories in advance, and incorporates an adaptive supervision strategy that adjusts the strength of reasoning-related losses according to problem difficulty.Experiments on mathematical reasoning benchmarks demonstrate consistent improvements over strong baselines, and ablation studies confirm the importance of both module-aware placement and adaptive supervision.

Clustering has attracted more and more attention as one of the most fundamental techniques in the field of unsupervised learning. To deal with nonlinear problems, clustering methods have been extended to the kernel version. As a traditional kernel clustering algorithm, multiple kernel k-means (MKKM) aims to learn clustering results from a consensus kernel obtained by combining a set of predefined kernels optimally. However, we observe that the existing MKKM algorithm and its variants insufficiently consider the noise that existed in kernel space and the underlying structure of kernelized data points. To this end, we propose a novel smooth MKKM via underlying graph filtering (SMKKM-UGF) to learn the smooth representations of kernelized data points through their nearby nodes in the underlying graph. In particular, different from the common graph filter, we jointly update the graph filter while learning the smooth kernel, so that the graph filter can be guaranteed to adapt to the updating kernel space constantly. Besides, an iterative algorithm with proven convergence is designed to solve the resultant optimization problem. Extensive experiments have been performed on numerous benchmark datasets, whose results prove the superiority of the proposed SMKKM-UGF compared to the other state-of-the-art clustering methods. The demo code of this work is publicly available at https://github.com/wqyang23/SMKKM-UGF.git.

Multiple kernel clustering (MKC) aims to extract and integrate the clustering information from a set of pre-defined kernels for handling data which cannot be linearly separated well. More precisely, existing MKC methods generally devote to learn the complementary information from a set of kernel partitions, whose feature dimensions are commonly fixed as the upper bound$n$or lower bound$c$, where$n$and$c$represents the number of samples and clusters, respectively. However, the adopting of the lower bound or upper bound generally leads to poor clustering performance caused by the lack or redundancy of clustering information carried by kernel partitions. To tackle this issue, we propose a novel late fusion multiple kernel clustering method, termed as Eigenvalue Ratio Inspired Partition Learning and Fusion for Multiple Kernel Clustering (ERMKC), in this paper. Specifically, we propose an eigenvalue ratio based criterion to guide the kernel partition learning for each single kernel matrix, which ensures more suitable feature dimensions for the learnt kernel partitions. In addition, we also propose a novel late fusion model for fusing the learnt kernel partitions optimally. Furthermore, we conduct extensive experiments on numerous benchmark datasets to evaluate the proposed ERMKC method, whose results verify the effectiveness and advantage of the proposed method compared to the other state-of-the-art methods.

This paper addresses the problem of face video inpainting. Existing video inpainting methods target primarily at natural scenes with repetitive patterns. They do not make use of any prior knowledge of the face to help retrieve correspondences for the corrupted face. They therefore only achieve sub-optimal results, particularly for faces under large pose and expression variations where face components appear very differently across frames. In this paper, we propose a two-stage deep learning method for face video inpainting. We employ 3DMM as our 3D face prior to transform a face between the image space and the UV (texture) space. In Stage I, we perform face inpainting in the UV space. This helps to largely remove the influence of face poses and expressions and makes the learning task much easier with well aligned face features. We introduce a frame-wise attention module to fully exploit correspondences in neighboring frames to assist the inpainting task. In Stage II, we transform the inpainted face regions back to the image space and perform face video refinement that inpaints any background regions not covered in Stage I and also refines the inpainted face regions. Extensive experiments have been carried out which show our method can significantly outperform methods based merely on 2D information, especially for faces under large pose and expression variations. Project page: https://ywq.github.io/FVIP.

This paper addresses the problem of single view 3D human reconstruction. Recent implicit function based methods have shown impressive results, but they fail to recover fine face details in their reconstructions. This largely degrades user experience in applications like 3D telepresence. In this paper, we focus on improving the quality of face in the reconstruction and propose a novel Jointly-aligned Implicit Face Function (JIFF) that combines the merits of the implicit function based approach and model based approach. We employ a 3D morphable face model as our shape prior and compute space-aligned 3D features that capture detailed face geometry information. Such space-aligned 3D features are combined with pixel-aligned 2D features to jointly predict an implicit face function for high quality face reconstruction. We further extend our pipeline and introduce a coarse-to-fine architecture to predict high quality texture for our detailed face model. Extensive evaluations have been carried out on public datasets and our proposed JIFF has demonstrates superior performance (both quantitatively and qualitatively) over existing state-of-the-arts.

In this paper, we address the "dual problem" of multi-view scene reconstruction in which we utilize single-view images captured under different point lights to learn a neural scene representation. Different from existing single-view methods which can only recover a 2.5D scene representation (i.e., a normal / depth map for the visible surface), our method learns a neural reflectance field to represent the 3D geometry and BRDFs of a scene. Instead of relying on multi-view photo-consistency, our method exploits two information-rich monocular cues, namely shading and shadow, to infer scene geometry. Experiments on multiple challenging datasets show that our method is capable of recovering 3D geometry, including both visible and invisible parts, of a scene from single-view images. Thanks to the neural reflectance field representation, our method is robust to depth discontinuities. It supports applications like novel-view synthesis and relighting. Our code and model can be found at https://ywq.github.io/s3nerf.