The paper introduces CARTS, a steganographic method that uses autoregressive language models to encode a payload text into a stegotext of identical token length by preserving per‑position rank information across contexts. It provides a formal security analysis, proving exact correctness under deterministic model assumptions, and defines key security notions such as context search, key collisions, message equivocation, and non‑commutativity of encoding maps. Empirical tests on Llama 3 8B confirm perfect payload recovery, no random key collisions, and no commuting key pairs, indicating resistance to the studied attack vectors.
By Wissam Ghantous, Alexander V. Mantzaris
The paper presents a deep‑learning approach for detecting cyberattacks in Low‑Earth Orbit satellite systems, leveraging the UNSW‑IoTSAT dataset. It explores structured architectures that preserve hardware, orbital, and radio‑frequency data, including a Subsystem‑Fusion MLP and a hierarchical multimodal Transformer that captures cross‑subsystem interactions and temporal dynamics. Experiments show that the hierarchical Transformer achieves up to 91.66% accuracy and 85.63% macro F1 under a leakage‑resistant evaluation protocol, highlighting the importance of multimodal modeling and rigorous testing.
By Kyle Stein, Guillermo Francia III, Eman El-Sheikh, Hossain Shahriar
The paper introduces ORCH, a method that applies human organizational theory to create task‑specific hierarchical structures for large, heterogeneous embodied AI teams. Using ORCH, teams of up to 50 agents across 25 wildfire‑response missions outperformed four existing multi‑agent frameworks, achieving higher mission scores and greater execution efficiency. Both human‑designed and language‑model‑generated ORCH organizations improved performance, with hierarchical organization preserving concurrent activity while coordinating ordered transitions between mission phases.
By Zhengran Ji, Jonathan Hyun, Boyuan Chen
MOSAIC is an open‑source platform that allows agents from different decision‑making paradigms—such as reinforcement learning policies, large language models, vision‑language models, and human operators—to operate together in shared reinforcement learning environments. It achieves this through an IPC‑based worker protocol that isolates each agent’s training and inference logic, an operator abstraction that maps any agent to a minimal universal interface, and a deterministic evaluation framework offering both manual lock‑step and automated script modes for reproducible experiments.
By Abdulhamid M. Mousa, Jinhui Pang, Rakhmonberdi Khajiev, Jalaledin M. Azzabi, Abdulkarim M. Mousa, Peng Yong, Yunusa Haruna, Ming Liu
The paper introduces Perturbation, a method that treats representations in language models as learning conduits rather than activation patterns. By fine‑tuning a model on a single adversarial example and observing how this perturbation spreads to other inputs, the approach avoids geometric assumptions and does not identify representations in untrained models. In trained models, Perturbation uncovers structured transfer across multiple linguistic scales, indicating that language models generalize along representational lines and acquire linguistic abstractions through experience.
By Joshua Rozner, Cory Shain
XAI-Arena proposes using large language models (LLMs) as judges to evaluate the quality of explainable AI (XAI) explanations, aiming for reproducibility, scalability, and multidimensional assessment. The framework assesses dimensions such as simplicity, clarity, task adequacy, trust calibration, actionability, transparency, faithfulness, and overall interpretability across different datasets, models, and stakeholder personas. Human validation shows a strong positive correlation between LLM-generated and human ratings (Spearman's rho = .693, p < .001), supporting the viability of LLM-based evaluations.
By Yanfei Hu Fleischhauer, Alona Zharova, Nadja Klein, Stefan Feuerriegel
The paper proposes a new technique that averages the logits of a frozen reference policy (such as one obtained via supervised fine‑tuning) and a trainable policy, integrating this approach into Group Relative Policy Optimization (GRPO). Unlike methods that use Kullback–Leibler regularization or a critic, the proposed method couples the policies solely through logit averaging, aiming to combine the reasoning strengths of the trainable policy with the formatting benefits of the reference policy. Experiments on MATH, cn‑k12, and MMLU demonstrate that this approach achieves higher or comparable accuracy to the standard KL‑regularized GRPO.
By Xingwei Gan, Ying Zhu
M3-Former is a multimodal transformer framework that uses large language models to encode vessel static attributes and navigational intent as semantic priors for long‑term trajectory prediction. It builds a unified multimodal representation space, aligns static semantic information with dynamic trajectory features via self‑attention, and employs a dual‑granularity Mixture‑of‑Experts architecture to capture both global route planning and fine‑grained maneuvering behaviors. A Steering‑Weighted Cross‑Entropy loss further improves accuracy on sparse turning events, and experiments on a Danish AIS dataset show consistent improvements over state‑of‑the‑art baselines, reducing ADE and FDE by up to 5.1% in 4‑hour predictions.
By Wenzhe Jin, Haina Tang
Halo is a modification to existing deep forecasters that adds a second output for estimating the scale of the predicted distribution, trained with a matching negative log likelihood. Experiments on five electricity price markets show that Halo improves mean squared error and mean absolute error in 28 of 30 model‑market‑metric comparisons, with average MSE reductions of 2.6% to 16.5% and MAE reductions of 1.7% to 11.0%. The study finds that the source of the scale estimate is less important than the fact that the network estimates scale, and that the improvement persists without retuning hyperparameters.
By Adam Cataldo
The paper introduces SIRGE, a sequence-informed geometric evaluator for RNA 3D structures that integrates nucleotide embeddings from a pretrained RNA language model into structural representations. SIRGE demonstrates superior performance over existing evaluators in Kendall–τ alignment, Top‑1 selection, and Top‑3 ranking. Controlled experiments reveal that sequence conditioning corrects errors of a purely geometric model and enhances target‑level ranking, suggesting that pretrained sequence representations provide complementary ranking information to geometric reasoning.
By Andrea Zerio, Yighua Yao, Alessandro Micheli, Roland G. Huber, Mile Sikic, Samir Bhatt, Andres R. Masegosa, Yuangang Pan
The paper explores soft prompting for few‑shot object detection with vision‑language models, showing that optimizing a small number of continuous prompt tokens—especially when placed at the cross‑modal boundary and initialized from an empty space token—can match LoRA performance while training far fewer parameters. Soft prompting also avoids catastrophic forgetting, transfers to newer models, and can be verbalized into readable prompts. The study extends these findings to manipulation tasks, indicating that VLMs already contain much of the necessary knowledge for specialized domains, and the main challenge is learning how to ask for it.
By Gautam Rajendrakumar Gare, Siyi Li, Hewei Wang, Cesar Daniel Hernandez, Wei Zhao, Wolfgang M. Pauli, John Galeotti, Deva Ramanan
The paper introduces BodyCam-VQA, an adaptive visual question answering framework designed to improve captioning of police body‑worn camera footage. By employing structured multimodal reasoning and probe question generation, the system extracts fine‑grained visual evidence that conventional captioning models miss. Experiments with various question generation models show that this VQA‑driven approach yields more reliable, objective, and detailed records of enforcement events.
By Karish Gupta, Matthew Alex, Alex Li, Yang Wu, Yun-Wei Chu, Kashif Munir, Xiaotian Zhou, Zhengping Ji, Xiaozhong Liu
The paper investigates whether large language models (LLMs) are narrowing the methodological diversity of archaeology. By analysing 119,000 abstracts from 2010‑2025 and running controlled experiments, the authors find only a modest shift in method use after 2023, with overall diversity actually increasing. However, LLMs tend to recommend a narrower, less diverse set of methods, especially without guidance, suggesting a potential convergence in methodological choice.
By Lorenzo Cardarelli, Roberto Ragno
The paper introduces DF-LLM, a Dynamic Fusion Large Language Model designed for traffic flow prediction. It combines a spatiotemporal embedding module, a fusion module that uses graph convolution to capture spatial topology and dynamic dependencies, and an LLM backbone with differentiated parameter adaptation and context aggregation attention. Experiments on four datasets demonstrate that DF-LLM outperforms existing methods in predictive accuracy.
By Xue Qiu, Jianli Xiao
CEM‑TUDASR is a lightweight, unsupervised Transformer‑based super‑resolution framework designed for Wireless Capsule Endoscopy (WCE) images. It uses a domain‑adaptive degradation network to synthesize realistic low‑resolution WCE images from high‑resolution conventional endoscopy data, enabling unpaired training. The SR generator incorporates Deep Attention Blocks and a Fusion Attention Block to preserve both global context and fine local structures, achieving superior no‑reference quality metrics and improved restoration of mucosal textures, vascular patterns, and anatomical details while remaining computationally efficient.
By Anjali Sarvaiya, Jay Kadel, Kishor Upla, Kiran Raja
The paper tackles two main issues in multi‑subject video generation—uncontrollable fidelity strength and semantic drift—by exploiting intrinsic attention patterns in Diffusion Transformers. It introduces an Intrinsic Spatial Grounding Map (ISGM) that accurately locates reference subjects and a Dual‑phase Intrinsic Attention Leveraging (DIAL) framework that uses ISGM during both training and inference. DIAL guides attention in low‑noise stages for precise fidelity control and builds preference pairs in high‑noise stages for reinforcement learning, resulting in superior identity consistency and controllable fidelity on the OpenS2V‑Eval benchmark.
By Niange Yu, Ye Tian, Biaolong Chen, Miao Lu, Aixi Zhang, Hao Jiang, Yunhai Tong, Pipei Huang
The paper examines whether adding syntactic and rhetorical structure to text can improve the prediction of incoherence in large language model outputs. Experiments show that plain text actually yields higher accuracy, as the added structural information conflicts with the models’ architectures. The authors also demonstrate that coherence assessment can help detect misleading content by applying zero‑shot experiments to a Brazilian disinformation dataset.
By Victor Mazzotti, Luiz Pereira, Marina Bitencourt dos Santos, Helena Maia, Carlos Caetano, N\'adia Felix, Sandra Avila
The study investigates whether cross‑lingual clinical annotation projection can be treated as a constrained text‑generation task that preserves the original text while inserting entity tags. Using a workflow that embeds tags directly into immutable target‑language text and then validates them deterministically, the authors evaluated this approach against supervised candidate‑span projection and hybrid ML‑LLM refinement across six languages. Results show that direct LLM projection, particularly with GLM 5.2 and Gemma4:31B, achieves the highest strict F1 scores (up to 0.9201) and outperforms previous methods by 0.0564–0.1512, producing over 55,000 grounded mentions with accurate offsets.
By \'Alvaro Rey-Blanes, Francisco J. Moreno-Barea, Francisco J. Veredas
The paper introduces a controlled inversion test to evaluate whether large language models can reverse known framing transformations in news articles while preserving facts. Using 60 articles and three framing types—evaluative lexis, agency realization, and information salience—the study generates 540 paired variants. Results show high factual preservation (~0.84) but low reversal success (0.044–0.068), indicating that recognizing a framing does not guarantee its undoing.
By Yi Liu
arXiv:2609. 11786v1 Announce Type: new Abstract: Automatic speech recognition (ASR) systems and audio language models (audio LMs) now report low error rates on monolingual benchmarks, but their behavior on code switched speech in low resource, diacritic rich languages remains poorly characterized.
By Chibuzor Okocha, Christan Earl Grant