Retrieval-augmented generation (RAG) has shown great potential for knowledge-intensive tasks, but its traditional architectures rely on static retrieval, limiting their effectiveness for complex questions that require sequential information-seeking. While agentic reasoning and search offer a more adaptive approach, most existing methods depend heavily on prompt engineering. In this work, we introduce RAG-Gym, a unified optimization framework that enhances information-seeking agents through fine-grained process supervision at each search step. We also propose ReSearch, a novel agent architecture that synergizes answer reasoning and search query generation within the RAG-Gym framework. Experiments on four challenging datasets show that RAG-Gym improves performance by up to 25.6\% across various agent architectures, with ReSearch consistently outperforming existing baselines. Further analysis highlights the effectiveness of advanced LLMs as process reward judges and the transferability of trained reward models as verifiers for different LLMs. Additionally, we examine the scaling properties of training and inference in agentic RAG. The project homepage is available at https://rag-gym.github.io/.

Current spectrum-sharing frameworks struggle with adaptability, often being either static or insufficiently dynamic. They primarily emphasize temporal sharing while overlooking spatial and spectral dimensions. We propose an adaptive, AI-driven spectrum-sharing framework within the O-RAN architecture, integrating discriminative and generative AI (GenAI) to forecast spectrum needs across multiple timescales and spatial granularities. A marketplace model, managed by an authorized spectrum broker, enables operators to trade spectrum dynamically, balancing static assignments with real-time trading. GenAI enhances traffic prediction, spectrum estimation, and allocation, optimizing utilization while reducing costs. This modular, flexible approach fosters operator collaboration, maximizing efficiency and revenue. A key research challenge is refining allocation granularity and spatio-temporal dynamics beyond existing models.

--The Navigation and Instrumentation (NavINST) Laboratory has developed a comprehensive multisensory dataset from various road-test trajectories in urban environments, featuring diverse lighting conditions, including indoor garage scenarios with dense 3D maps. This dataset includes multiple commercial-grade IMUs and a high-end tactical-grade IMU. Additionally, it contains a wide array of perception-based sensors, such as a solid-state LiDAR--making it one of the first datasets to do so--a mechanical LiDAR, four electronically scanning RADARs, a monocular camera, and two stereo cameras. The dataset also includes forward speed measurements derived from the vehicle's odometer, along with accurately post-processed high-end GNSS/IMU data, providing precise ground truth positioning and navigation information. The NavINST dataset is designed to support advanced research in high-precision positioning, navigation, mapping, computer vision, and multisensory fusion. It offers rich, multi-sensor data ideal for developing and validating robust algorithms for autonomous vehicles. Finally, it is fully integrated with the robot operating system (ROS), ensuring ease of use and accessibility for the research community. The complete dataset and development tools are available at navinst.github.io. The last decade has witnessed significant advancements in autonomous driving, robotics, and computer vision, transforming these fields with innovative applications. In particular, autonomous vehicles (A Vs) technology has ushered in a new era of transportation, promising increased safety, efficiency, and convenience [1]. These advancements in A Vs are fundamentally reliant on robust navigation systems capable of achieving higher levels of autonomy by operating seamlessly in diverse and dynamic environments while ensuring accuracy, reliability, and adaptability [2]. A major enabler of these research advances has been the publication of diverse datasets by research groups that provide high-quality standardized data that supports the development, testing, and benchmarking of innovative algorithms.

Current recommendation systems powered by large language models (LLMs) often underutilize their reasoning capabilities due to a lack of explicit logical structuring. To address this limitation, we introduce CoT-Rec, a framework that integrates Chain-of-Thought (CoT) reasoning into LLM-driven recommendations by incorporating two crucial processes: user preference analysis and item perception evaluation. CoT-Rec operates in two key phases: (1) personalized data extraction, where user preferences and item perceptions are identified, and (2) personalized data application, where this information is leveraged to refine recommendations. Our experimental analysis demonstrates that CoT-Rec improves recommendation accuracy by making better use of LLMs' reasoning potential. The implementation is publicly available at https://anonymous.4open.science/r/CoT-Rec.

Large Vision-Language Models (LVLMs) have recently gained attention due to their distinctive performance and broad applicability. While it has been previously shown that their efficacy in usage scenarios involving non-Western contexts falls short, existing studies are limited in scope, covering just a narrow range of cultures, focusing exclusively on a small number of cultural aspects, or evaluating a limited selection of models on a single task only. Towards globally inclusive LVLM research, we introduce GIMMICK, an extensive multimodal benchmark designed to assess a broad spectrum of cultural knowledge across 144 countries representing six global macro-regions. GIMMICK comprises six tasks built upon three new datasets that span 728 unique cultural events or facets on which we evaluated 20 LVLMs and 11 LLMs, including five proprietary and 26 open-weight models of all sizes. We systematically examine (1) regional cultural biases, (2) the influence of model size, (3) input modalities, and (4) external cues. Our analyses reveal strong biases toward Western cultures across models and tasks and highlight strong correlations between model size and performance, as well as the effectiveness of multimodal input and external geographic cues. We further find that models have more knowledge of tangible than intangible aspects (e.g., food vs. rituals) and that they excel in recognizing broad cultural origins but struggle with a more nuanced understanding.

--Performing accurate confidence quantification and assessment in pixel-wise regression tasks, which are downstream applications of AI Foundation Models for Earth Observation (EO), is important for deep neural networks to predict their failures, improve their performance and enhance their capabilities in real-world applications, for their practical deployment. For pixel-wise regression tasks, specifically utilizing remote sensing data from satellite imagery in EO Foundation Models, confidence quantification is a critical challenge. The focus of this research is on developing a Foundation Model using EO satellite data that computes and assigns a confidence metric alongside regression outputs to improve the reliability and interpretability of predictions generated by deep neural networks. T o this end, we develop, train and evaluate the proposed Confidence-A ware Regression Estimation (CARE) Foundation Model. Our model CARE computes and assigns confidence to regression results as downstream tasks of a Foundation Model for EO data, and performs a confidence-aware self-corrective learning method for the low-confidence regions. We evaluate the model CARE, and experimental results on multi-spectral data from the Copernicus Sentinel-2 constellation to estimate the building density (i.e. We also show that our model CARE outperforms other methods. The significance of confidence quantification and assessment in deep learning, specifically in AI Foundation Models in Earth Observation (EO) that use satellite data, for regression applications is critical. The utility of satellite data seems inexhaustible, and thanks to developments in AI, applications emerge at an accelerated pace in EO Foundation Models using remote sensing data.

Large Language Models (LLMs), when used for conditional text generation, often produce hallucinations, i.e., information that is unfaithful or not grounded in the input context. This issue arises in typical conditional text generation tasks, such as text summarization and data-to-text generation, where the goal is to produce fluent text based on contextual input. When fine-tuned on specific domains, LLMs struggle to provide faithful answers to a given context, often adding information or generating errors. One underlying cause of this issue is that LLMs rely on statistical patterns learned from their training data. This reliance can interfere with the model's ability to stay faithful to a provided context, leading to the generation of ungrounded information. We build upon this observation and introduce a novel self-supervised method for generating a training set of unfaithful samples. We then refine the model using a training process that encourages the generation of grounded outputs over unfaithful ones, drawing on preference-based training. Our approach leads to significantly more grounded text generation, outperforming existing self-supervised techniques in faithfulness, as evaluated through automatic metrics, LLM-based assessments, and human evaluations.

We present PeerQA, a real-world, scientific, document-level Question Answering (QA) dataset. PeerQA questions have been sourced from peer reviews, which contain questions that reviewers raised while thoroughly examining the scientific article. Answers have been annotated by the original authors of each paper. The dataset contains 579 QA pairs from 208 academic articles, with a majority from ML and NLP, as well as a subset of other scientific communities like Geoscience and Public Health. PeerQA supports three critical tasks for developing practical QA systems: Evidence retrieval, unanswerable question classification, and answer generation. We provide a detailed analysis of the collected dataset and conduct experiments establishing baseline systems for all three tasks. Our experiments and analyses reveal the need for decontextualization in document-level retrieval, where we find that even simple decontextualization approaches consistently improve retrieval performance across architectures. On answer generation, PeerQA serves as a challenging benchmark for long-context modeling, as the papers have an average size of 12k tokens. Our code and data is available at https://github.com/UKPLab/peerqa.

Ontology, and more broadly, Knowledge Graph Matching is a challenging task in which expressiveness has not been fully addressed. Despite the increasing use of embeddings and language models for this task, approaches for generating expressive correspondences still do not take full advantage of these models, in particular, large language models (LLMs). This paper proposes to integrate LLMs into an approach for generating expressive correspondences based on alignment need and ABox-based relation discovery. The generation of correspondences is performed by matching similar surroundings of instance sub-graphs. The integration of LLMs results in different architectural modifications, including label similarity, sub-graph matching, and entity matching. The performance word embeddings, sentence embeddings, and LLM-based embeddings, was compared. The results demonstrate that integrating LLMs surpasses all other models, enhancing the baseline version of the approach with a 45\% increase in F-measure.

Machine Learning Potentials (MLPs) can enable simulations of ab initio accuracy at orders of magnitude lower computational cost. However, their effectiveness hinges on the availability of considerable datasets to ensure robust generalization across chemical space and thermodynamic conditions. The generation of such datasets can be labor-intensive, highlighting the need for innovative methods to train MLPs in data-scarce scenarios. Here, we introduce transfer learning of potential energy surfaces between chemically similar elements. Specifically, we leverage the trained MLP for silicon to initialize and expedite the training of an MLP for germanium. Utilizing classical force field and ab initio datasets, we demonstrate that transfer learning surpasses traditional training from scratch in force prediction, leading to more stable simulations and improved temperature transferability. These advantages become even more pronounced as the training dataset size decreases. The out-of-target property analysis shows that transfer learning leads to beneficial but sometimes adversarial effects. Our findings demonstrate that transfer learning across chemical elements is a promising technique for developing accurate and numerically stable MLPs, particularly in a data-scarce regime.