The paper introduces FlightLLM, a prior-guided semantic approach that uses large language models to explain flight safety events. It tackles challenges such as modal inconsistency, limited classification ability, and scarce domain data by combining feature engineering, semantic discretization, a CatBoost statistical expert, contrastive few-shot learning, and structured prompts. Evaluated on 704 real‑world A320 flights, FlightLLM achieves competitive classification and produces clear, aviation‑specific explanations for hard landing events.
By Lu Xu, Xu Li, Linjiang Zheng, Fan Li, Riquan Zhang, Jiaxing Shang
The paper introduces OVRSISBench, a unified benchmark for open‑vocabulary remote sensing image segmentation, and evaluates existing OVS/OVRSIS models, uncovering their shortcomings in remote sensing contexts. Leveraging insights from this evaluation, the authors propose RSKT‑Seg, a new framework featuring a Multi‑Directional Cost Map Aggregation module, an Efficient Cost Map Fusion transformer, and a Remote Sensing Knowledge Transfer module. Experiments on the benchmark demonstrate that RSKT‑Seg outperforms strong baselines by +3.8 mIoU and +5.9 mACC while achieving twice the inference speed.
By Bingyu Li, Haocheng Dong, Da Zhang, Zhiyuan Zhao, Junyu Gao, Xuelong Li
G-ReAct is a reasoning framework that frames deep search as state evolution over a fixed-topology query graph, enabling explicit tracking of search progress and constraint preservation. It generates high-quality trajectories for fine-tuning and provides structured guidance during inference without extra fine-tuning. Experiments show that with only 1.9K generated trajectories, a Qwen3 model achieves strong accuracy on BrowseComp-ZH and XBench, outperforming larger open-source baselines, and consistently improves existing LLMs on deep-search tasks.
By Shaoxiong Yang, Mengyuan Zhang, Shaojun Lin, Chao Li, Wei Liu, Kun Shao, Jian Luan
The paper explores a synthetic-first training approach for detecting drones in medium- and long-wave infrared imagery, combining synthetic scene generation with fine-tuning on real data. It demonstrates that synthetic data can establish initial object representations, but real infrared data is crucial to close domain gaps and improve reliability. The study finds that aligning datasets has a greater impact on performance than increasing model size, and that semantic alignment in feature space is the strongest predictor of success, with radiometric factors like entropy and dynamic range also contributing.
By Tanel Liiv, Sander Soodla, Nzamba Bignoumba, Alma M. Liezenga, Toomas Pruuden
The paper introduces SEE, a two-stage framework for generating long-horizon GUI agent trajectories. First, an exploration stage builds an explicit UI transition graph over screens and elements. Second, a graph-based synthesis stage composes diverse multi-step trajectories through planning and controlled sampling, preventing spurious cycles and enabling long-horizon composition. Across real-world apps, SEE produces trajectories averaging 14.8 steps and improves agent task success and generalization to unseen screens.
By Zhuohang Fan, Beichen Zhang, Yuanfa Li, Changqiao Wu, Wei Liu, Jian Luan, Weigang Zhang
Wuying-Browser-Agent is a unified framework designed to improve long-horizon browser agents by aligning execution, supervision, optimization, and evaluation. It introduces a structured browser harness, reflection and UI-specialized Curriculum SFT (RUIC‑SFT) for recovery and complex UI interactions, and Divergence‑Aware Online GRPO (DAO‑GRPO) for better credit assignment. The framework is evaluated on BrowserBench—a bilingual real‑web benchmark of 350 tasks—and achieves state‑of‑the‑art results on multiple browser‑use benchmarks, while also transferring well to other agentic tasks.
By AIMAE Team, Tianxiang Chen, Yan Cheng, Zhangye Han, Xiaowei Li, Chang Liu, Cheng Liu, Zhongqiang Ma, Long Peng, Xiaobing Tu, Yinggui Wang, Hongliang Wei, Chen Wu, Daiping Xin, Kunyu Zhou, Pengyang Zhou, Peiyuan Chen, Ziyuan Chen, Yutao Deng, Chunyu Dong, Xiangyu Fu, Yicheng Feng, Ruian He, Haochen Li, Miancan Liu, Zhengqin Liu, Wei Peng, Jinkui Ren, Haoyu Tan, Dong Xiao, Rongkun Xue, Shujian Yang, Xianhang Ye, Ziqi Yuan, Ziyang Yu, Linghan Zhang, Xiantao Zhang, Xuanpu Zhao, Yinan Zhao, Zhenghui Zhao, Bin Zhu, Likai Zou
J-Miner extracts executable decision knowledge from fine‑tuned language‑model classifiers by aggregating internal signals aligned with named concepts and learning rules over them. The mined rules reproduce up to 98.3% of the original classifier’s decisions and outperform compact rules based on input words by 6.0–29.5 percentage points in behavioral fidelity. Moreover, these rules can be transferred to lightweight student models that use only about 1/24 of the parameters yet retain 99.8% of the source classifier’s accuracy.
By Yunfan Gao, Xinyi Huang, Tao Sheng, Haorui Song, Yun Xiong, Haofen Wang
The paper introduces SW-ProxyCE, a zero-query adversarial attack that exploits publicly released EEG foundation encoders to generate transferable adversarial examples for private downstream models. By using a small labeled reference set and shrinkage-whitened class prototypes, the method recovers task-level decision geometry without training a surrogate classifier. Experiments across three EEG tasks and multiple encoders show that SW-ProxyCE consistently outperforms task-agnostic attacks, demonstrating that the strong transferability of EEG foundation models does not guarantee adversarial robustness.
By Linhua Cong, Dingkun Liu, Dongrui Wu
The paper investigates using a small language model (SBERT) for invoice categorisation, showing that fine‑tuning on a single GPU yields 0.96 accuracy and 0.9 F1 with about 100 client‑specific invoices. It analyses the embedding geometry, finding that the sentence‑embedding space is globally anisotropic but locally isotropic, with clusters strongly linked to vendor identity. The study demonstrates that an in‑house SLM can outperform zero‑shot LLMs and vendor baselines while improving cost, security, and interpretability.
By Emma Ceccherini, Daniel Lawson, Anjulika Salhan
SignalReasoner evaluates reinforcement fine‑tuning methods for the Qwen2.5‑3B‑Base language model on graduate‑level signal processing math problems from the WirelessMATHBench‑XL benchmark. Two training paradigms are compared: direct reinforcement learning with verifiable rewards, and supervised chain‑of‑thought fine‑tuning followed by domain‑specific reinforcement learning. The study benchmarks three policy optimization algorithms—GRPO, GSPO, and GMPO—to determine if domain‑aware chain‑of‑thought initialization improves performance, achieving a best accuracy of 39.12% versus 12.37% for the untrained base model.
By Guozheng Sun
The paper investigates how different reward specifications affect the reliability of unlearning in large language models using a LoRA-GRPO framework. It compares four reward designs—lexical suppression, anti-refusal shaping, rubric-based broad answering, and explicit refusal contrast—both with and without a supervised fine-tuning warm-up. The results reveal that successful optimization does not guarantee behavioral unlearning, as various evaluation metrics can yield conflicting conclusions due to reward-hacking, policy-support limits, and benchmark probe limitations.
By Rub\'en Balbastre, Juan Manuel Ordu\~na, Mariano P\'erez
The study investigates whether Large Language Models (LLMs) can translate technical explanations from credit risk models into stakeholder-friendly narratives. Using Freddie Mac loan data, the authors compare standard tabular models (XGBoost + SHAP) with alternative data pipelines (GNN + GNNExplainer and a bimodal mix) and generate explanations with three LLM configurations: a small fine‑tuned Gemma 3 4B, a large fine‑tuned DeepSeek R1 70B, and a zero‑shot Gemini 2.5. Findings show that the quality of explanations is more dependent on the evidence representation than on the LLM, that narratives reliably identify influential factors but are less consistent about the direction of influence, and that credit professionals demand higher evidentiary standards than non‑professionals.
By Sahab Zandi, Noah Kostesku, Christophe Mues, Mar\'ia \'Oskarsd\'ottir, Cristi\'an Bravo
The paper proposes Entanglement-Weighted Pruning (EWP), a method for unlearning a client’s contribution from a federated quantum classifier without retraining from scratch. EWP scores each trainable circuit parameter by combining a Fisher‑information estimate on the target client’s data with a structural entanglement weight, pruning the lowest‑scoring parameters and optionally fine‑tuning the remaining ones. Experiments on a four‑qubit data‑re‑uploading ansatz trained with FedAvg across five simulated supply‑chain‑risk clients show that EWP achieves accuracy comparable to full retraining while reducing forgetting and wall‑clock time by about sixteenfold, outperforming random, Fisher‑only, or entanglement‑only pruning.
By Aditya Kumar, Sumit Chongder
The paper proposes Self‑Distillation Fine‑Tuning (SDFT) as a method to recover performance in Large Language Models that has been degraded by catastrophic forgetting, quantization, or pruning. It shows that SDFT restores model capabilities by aligning the high‑dimensional manifold of the student model’s hidden layers with that of a teacher model, as measured by Centered Kernel Alignment (CKA). The authors provide both empirical evidence of strong correlation between manifold alignment and performance recovery and a theoretical explanation linking generative capability to the structure of these manifolds.
By Chi Liu, Xin Chen, Xu Zhou, Fangbo Tu, Srinivasan Manoharan
Data-DPO is a target model‑oriented supervised fine‑tuning data selection method that uses one‑step probing of the target model to generate pairwise data preferences, trains a lightweight reward model to capture these preferences, and then selects a training subset by combining target‑model preference, external quality scores, and marginal diversity. Experiments on Vision‑Flan and LLaVA‑CoT demonstrate that Data‑DPO consistently outperforms existing data selection baselines across multiple data budgets and even surpasses full data training performance.
By Peng Sun, Yi Yang, Antong Zhang, Chunxiao Li, Yanbo Wang, Dianbo Liu, xin chen, Kai Yu, Lu Chen, Tianfan Fu
The paper introduces MASS, a hierarchical data selection method that first groups data using low‑dimensional manifold coordinates learned by a dense autoencoder, then applies a TopK sparse autoencoder for quality‑aware feature coverage within each group. This approach addresses the shortcomings of traditional diversity metrics that mix semantic, supervisory, and noise signals. Experiments on Vision Flan and LLaVA‑CoT demonstrate that MASS outperforms existing baselines across various budgets and can match or exceed full‑data training with only a small subset.
By Peng Sun, Yi Yang, Antong Zhang, Chunxiao Li, Yanbo Wang, Dianbo Liu, xin chen, Kai Yu, Lu Chen, Tianfan Fu
PTXBench is a benchmark designed to evaluate and adapt large language models (LLMs) for GPU kernel optimization using architecture‑specific PTX code. It assesses functional correctness, runtime execution of target instructions, and speedup over state‑of‑the‑art libraries on GEMM and attention workloads on H100 and B200 GPUs. The study finds uneven success rates, especially on complex attention backward tasks, and shows that executing target instructions does not guarantee competitive performance, with no model consistently outperforming frontier libraries. The authors also fine‑tune Qwen3.6‑27B, noting that repair‑conditioned training improves some tasks but generalization remains inconsistent, highlighting the importance of data coverage, balance, and teacher quality.
By Genghan Zhang, Yixin Dong, Chengze Fan, Zhichen Zeng, Yueming Yuan, Shaowei Zhu, Kunle Olukotun
The paper explores using a small language model (SBERT) for invoice categorisation, a task that requires nuanced accounting judgement. By analysing the embedding geometry of SBERT and DeBERTa, the authors find that the sentence‑embedding space is globally anisotropic but contains locally isotropic clusters tied to vendor identity. Fine‑tuned SBERT achieves 0.96 accuracy and 0.9 F1 with only about 100 client‑specific invoices, outperforming zero‑shot LLMs and vendor baselines, and demonstrates that in‑house SLMs can reduce cost, enhance security, and improve interpretability.
The paper introduces TSN4PI, a unified framework for tracking the evolution of political ideologies on social media. It combines a PIDN module that uses large language models, style transfer, and unsupervised domain adaptation to detect ideologies and filter noise, with a PIPN module that employs temporal graph neural networks to predict future ideological shifts. The authors release two large-scale datasets and validate the approach on platforms such as X and Truth Social, offering empirical insights into political polarization and ideology evolution.
Authorship verification (AV) assumes that an author's writing style remains sufficiently stable to distinguish it from that of other writers. In practice, however, this assumption is challenged by dis...