arXiv subjects

Yijiashun Qi

Publications and source records attributed to Yijiashun Qi.

12 recordsLinked to original sources

arXiv2026

When Not to Write Memory: Governing False Promotion from Correlated Agent Traces

Long-lived language agents increasingly write reusable memories from their own execution traces. The key safety question is not only what agents should remember, but when they should refuse to write memory at all. Repeated observations across agents are not necessarily independent evidence: the same claim may be copied from a shared source, induced by a shared prompt, stale under a new environment, or valid only in a narrower scope. We study this failure mode as a memory writepath governance problem. We introduce GovMem as a conservative diagnostic reference policy that estimates dependency-aware support, retrieves counterevidence, assigns scope, and outputs one of three decisions: promote, reject, or needs-review. In controlled synthetic stress tests, GovMem reduces false promotion from 0.597 to 0.040 in the default setting while preserving 0.960 recall, at an explicit review burden. In a project-internal 120candidate human-labeled real-trace subset spanning 79 recorded traces and project reports, dependency-aware promotion reduces false promotion from 0.371 for source+scope to 0.032 overall, but held-out false promotion remains 0.111 and the method is highly conservative, with 0.692 review burden and 0.448 direct recall. A final human adjudication of 133 high-impact external codingagent candidates is more severe: none are safe for automatic promotion, and all 11 verification-gate positives are rejected as boilerplate, shared-tool artifacts, file dumps, or non-reusable debugging traces. These results support GovMem primarily as a diagnostic governance design point, not as a generally validated or efficient automatic memory writer: agent memory write paths should be evaluated as risk-controlled evidence-governance systems, while broader external coverage and downstream harm evidence are still needed before stronger claims.

cs.SE↗

arXiv2026

Detecting High-Potential SMEs with Heterogeneous Graph Neural Networks

Small and Medium Enterprises (SMEs) constitute 99.9% of U.S. businesses and generate 44% of economic activity, yet systematically identifying high-potential SMEs remains an open challenge. We introduce SME-HGT, a Heterogeneous Graph Transformer framework that predicts which SBIR Phase I awardees will advance to Phase II funding using exclusively public data. We construct a heterogeneous graph with 32,268 company nodes, 124 research topic nodes, and 13 government agency nodes connected by approximately 99,000 edges across three semantic relation types. SME-HGT achieves an AUPRC of 0.621 0.003 on a temporally-split test set, outperforming an MLP baseline (0.590 0.002) and R-GCN (0.608 0.013) across five random seeds. At a screening depth of 100 companies, SME-HGT attains 89.6% precision with a 2.14 lift over random selection. Our temporal evaluation protocol prevents information leakage, and our reliance on public data ensures reproducibility. These results demonstrate that relational structure among firms, research topics, and funding agencies provides meaningful signal for SME potential assessment, with implications for policymakers and early-stage investors.

cs.LG↗

arXiv2026

Coverage-Aware Web Crawling for Domain-Specific Supplier Discovery via a Web--Knowledge--Web Pipeline

Identifying the full landscape of small and medium-sized enterprises (SMEs) in specialized industry sectors is critical for supply-chain resilience, yet existing business databases suffer from substantial coverage gaps -- particularly for sub-tier suppliers and firms in emerging niche markets. We propose a \textbf{Web--Knowledge--Web (W$\to$K$\to$W)} pipeline that iteratively (1)~crawls domain-specific web sources to discover candidate supplier entities, (2)~extracts and consolidates structured knowledge into a heterogeneous knowledge graph using domain-adapted few-shot LLM prompting, and (3)~uses the knowledge graph's topology and coverage signals to guide subsequent crawling toward under-represented regions of the supplier space. To quantify discovery completeness, we introduce a \textbf{coverage estimation framework} inspired by ecological species-richness estimators (Chao1, ACE) adapted for web-entity populations. Experiments on the semiconductor equipment manufacturing sector (NAICS 333242) demonstrate that the W$\to$K$\to$W pipeline achieves the highest precision (0.165) and F1 (0.123) among all methods while using only 144 pages -- 32\% fewer than the 213-page baseline budget -- building a knowledge graph of 664 entities and 542 relations with 100\% relation type-consistency.

cs.LG↗

arXiv2026

When Corrective Hints Hurt: Prompt Design in Reasoner-Guided Repair of LLM Overcaution on Entailed Negations under OWL~2~DL

We report a reproducible error pattern in GPT-5.4 on OWL~2~DL compliance queries: the model frequently answers ``unknown'' when the reasoner-entailed answer is ``no'' under \emph{FunctionalProperty} closure or class \emph{disjointness}. Using 180 reasoner-audited queries from a procedural expansion of the observed pattern plus 18 hand-authored held-out queries in two unrelated domains (insurance and clinical), we compare four interaction modes under matched query budget: single-shot, three rounds of generic ``you-are-wrong'' retry, three rounds of reasoner-verdict repair with an open-world-assumption (OWA) hint, and the same repair without the hint. Direct faithfulness is 43.9\,\% (Wilson 95\,\% CI $[36.8,51.2]$); generic retry reaches 81.7\,\% ($[75.4,86.6]$); the verdict-with-hint variant is \emph{worse} at 67.2\,\% ($[60.1,73.7]$); the verdict-only variant reaches 97.8\,\% ($[94.4,99.1]$). All pairwise comparisons remain significant under McNemar's exact test with Bonferroni correction ($α= 0.01$; all $p < 10^{-5}$). The same fingerprint accounts for 4/4 errors on the held-out queries. Our interpretation is bounded: prompt framing can matter more than corrective content, and reasoner-guided wrappers should be ablated explicitly.

cs.AI↗

arXiv2025

Modality-Aware Infrared and Visible Image Fusion with Target-Aware Supervision

Infrared and visible image fusion (IVIF) is a fundamental task in multi-modal perception that aims to integrate complementary structural and textural cues from different spectral domains. In this paper, we propose FusionNet, a novel end-to-end fusion framework that explicitly models inter-modality interaction and enhances task-critical regions. FusionNet introduces a modality-aware attention mechanism that dynamically adjusts the contribution of infrared and visible features based on their discriminative capacity. To achieve fine-grained, interpretable fusion, we further incorporate a pixel-wise alpha blending module, which learns spatially-varying fusion weights in an adaptive and content-aware manner. Moreover, we formulate a target-aware loss that leverages weak ROI supervision to preserve semantic consistency in regions containing important objects (e.g., pedestrians, vehicles). Experiments on the public M3FD dataset demonstrate that FusionNet generates fused images with enhanced semantic preservation, high perceptual quality, and clear interpretability. Our framework provides a general and extensible solution for semantic-aware multi-modal image fusion, with benefits for downstream tasks such as object detection and scene understanding.

cs.CV↗

arXiv2025

Predicting Asphalt Pavement Friction Using Texture-Based Image Indicator

Pavement skid resistance is of vital importance for road safety. The objective of this study is to propose and validate a texture-based image indicator to predict pavement friction. This index enables pavement friction to be measured easily and inexpensively using digital images. Three different types of asphalt surfaces (dense-graded asphalt mix, open-grade friction course, and chip seal) were evaluated subject to various tire polishing cycles. Images were taken with corresponding friction measured using Dynamic Friction Tester (DFT) in the laboratory. The aggregate protrusion area is proposed as the indicator. Statistical models are established for each asphalt surface type to correlate the proposed indicator with friction coefficients. The results show that the adjusted R-square values of all relationships are above 0.90. Compared to other image-based indicators in the literature, the proposed image indicator more accurately reflects the changes in pavement friction with the number of polishing cycles, proving its cost-effective use for considering pavement friction in mix design stage.

cs.CV↗

arXiv2025

Structured Memory Mechanisms for Stable Context Representation in Large Language Models

This paper addresses the limitations of large language models in understanding long-term context. It proposes a model architecture equipped with a long-term memory mechanism to improve the retention and retrieval of semantic information across paragraphs and dialogue turns. The model integrates explicit memory units, gated writing mechanisms, and attention-based reading modules. A forgetting function is introduced to enable dynamic updates of memory content, enhancing the model's ability to manage historical information. To further improve the effectiveness of memory operations, the study designs a joint training objective. This combines the main task loss with constraints on memory writing and forgetting. It guides the model to learn better memory strategies during task execution. Systematic evaluation across multiple subtasks shows that the model achieves clear advantages in text generation consistency, stability in multi-turn question answering, and accuracy in cross-context reasoning. In particular, the model demonstrates strong semantic retention and contextual coherence in long-text tasks and complex question answering scenarios. It effectively mitigates the context loss and semantic drift problems commonly faced by traditional language models when handling long-term dependencies. The experiments also include analysis of different memory structures, capacity sizes, and control strategies. These results further confirm the critical role of memory mechanisms in language understanding. They demonstrate the feasibility and effectiveness of the proposed approach in both architectural design and performance outcomes.

cs.CL↗

arXiv2025

NeRF-Based defect detection

The rapid growth of industrial automation has highlighted the need for precise and efficient defect detection in large-scale machinery. Traditional inspection techniques, involving manual procedures such as scaling tall structures for visual evaluation, are labor-intensive, subjective, and often hazardous. To overcome these challenges, this paper introduces an automated defect detection framework built on Neural Radiance Fields (NeRF) and the concept of digital twins. The system utilizes UAVs to capture images and reconstruct 3D models of machinery, producing both a standard reference model and a current-state model for comparison. Alignment of the models is achieved through the Iterative Closest Point (ICP) algorithm, enabling precise point cloud analysis to detect deviations that signify potential defects. By eliminating manual inspection, this method improves accuracy, enhances operational safety, and offers a scalable solution for defect detection. The proposed approach demonstrates great promise for reliable and efficient industrial applications.

cs.CV↗

arXiv2025

Graph Neural Network-Driven Hierarchical Mining for Complex Imbalanced Data

This study presents a hierarchical mining framework for high-dimensional imbalanced data, leveraging a depth graph model to address the inherent performance limitations of conventional approaches in handling complex, high-dimensional data distributions with imbalanced sample representations. By constructing a structured graph representation of the dataset and integrating graph neural network (GNN) embeddings, the proposed method effectively captures global interdependencies among samples. Furthermore, a hierarchical strategy is employed to enhance the characterization and extraction of minority class feature patterns, thereby facilitating precise and robust imbalanced data mining. Empirical evaluations across multiple experimental scenarios validate the efficacy of the proposed approach, demonstrating substantial improvements over traditional methods in key performance metrics, including pattern discovery count, average support, and minority class coverage. Notably, the method exhibits superior capabilities in minority-class feature extraction and pattern correlation analysis. These findings underscore the potential of depth graph models, in conjunction with hierarchical mining strategies, to significantly enhance the efficiency and accuracy of imbalanced data analysis. This research contributes a novel computational framework for high-dimensional complex data processing and lays the foundation for future extensions to dynamically evolving imbalanced data and multi-modal data applications, thereby expanding the applicability of advanced data mining methodologies to more intricate analytical domains.

cs.LG↗

arXiv2025

Optimized Unet with Attention Mechanism for Multi-Scale Semantic Segmentation

Semantic segmentation is one of the core tasks in the field of computer vision, and its goal is to accurately classify each pixel in an image. The traditional Unet model achieves efficient feature extraction and fusion through an encoder-decoder structure, but it still has certain limitations when dealing with complex backgrounds, long-distance dependencies, and multi-scale targets. To this end, this paper proposes an improved Unet model combined with an attention mechanism, introduces channel attention and spatial attention modules, enhances the model's ability to focus on important features, and optimizes skip connections through a multi-scale feature fusion strategy, thereby improving the combination of global semantic information and fine-grained features. The experiment is based on the Cityscapes dataset and compared with classic models such as FCN, SegNet, DeepLabv3+, and PSPNet. The improved model performs well in terms of mIoU and pixel accuracy (PA), reaching 76.5% and 95.3% respectively. The experimental results verify the superiority of this method in dealing with complex scenes and blurred target boundaries. In addition, this paper discusses the potential of the improved model in practical applications and future expansion directions, indicating that it has broad application value in fields such as autonomous driving, remote sensing image analysis, and medical image processing.

cs.CV↗

arXiv2024

Collaborative Optimization in Financial Data Mining Through Deep Learning and ResNeXt

This study proposes a multi-task learning framework based on ResNeXt, aiming to solve the problem of feature extraction and task collaborative optimization in financial data mining. Financial data usually has the complex characteristics of high dimensionality, nonlinearity, and time series, and is accompanied by potential correlations between multiple tasks, making it difficult for traditional methods to meet the needs of data mining. This study introduces the ResNeXt model into the multi-task learning framework and makes full use of its group convolution mechanism to achieve efficient extraction of local patterns and global features of financial data. At the same time, through the design of task sharing layers and dedicated layers, it is established between multiple related tasks. Deep collaborative optimization relationships. Through flexible multi-task loss weight design, the model can effectively balance the learning needs of different tasks and improve overall performance. Experiments are conducted on a real S&P 500 financial data set, verifying the significant advantages of the proposed framework in classification and regression tasks. The results indicate that, when compared to other conventional deep learning models, the proposed method delivers superior performance in terms of accuracy, F1 score, root mean square error, and other metrics, highlighting its outstanding effectiveness and robustness in handling complex financial data. This research provides an efficient and adaptable solution for financial data mining, and at the same time opens up a new research direction for the combination of multi-task learning and deep learning, which has important theoretical significance and practical application value.

cs.LG↗

arXiv2024

Benchmarking Large Language Models for Image Classification of Marine Mammals

As Artificial Intelligence (AI) has developed rapidly over the past few decades, the new generation of AI, Large Language Models (LLMs) trained on massive datasets, has achieved ground-breaking performance in many applications. Further progress has been made in multimodal LLMs, with many datasets created to evaluate LLMs with vision abilities. However, none of those datasets focuses solely on marine mammals, which are indispensable for ecological equilibrium. In this work, we build a benchmark dataset with 1,423 images of 65 kinds of marine mammals, where each animal is uniquely classified into different levels of class, ranging from species-level to medium-level to group-level. Moreover, we evaluate several approaches for classifying these marine mammals: (1) machine learning (ML) algorithms using embeddings provided by neural networks, (2) influential pre-trained neural networks, (3) zero-shot models: CLIP and LLMs, and (4) a novel LLM-based multi-agent system (MAS). The results demonstrate the strengths of traditional models and LLMs in different aspects, and the MAS can further improve the classification performance. The dataset is available on GitHub: https://github.com/yeyimilk/LLM-Vision-Marine-Animals.git.

cs.CV↗