High-quality wind power forecasting is crucial for the operation of modern power grids. However, prevailing data-driven paradigms either train a site-specific model which cannot generalize to other locations or rely on fine-tuning of general-purpose time series foundation models which are difficult to incorporate domain-specific data in the energy sector. This paper introduces WindFM, a lightweight and generative Foundation Model designed specifically for probabilistic wind power forecasting. WindFM employs a discretize-and-generate framework. A specialized time-series tokenizer first converts continuous multivariate observations into discrete, hierarchical tokens. Subsequently, a decoder-only Transformer learns a universal representation of wind generation dynamics by autoregressively pre-training on these token sequences. Using the comprehensive WIND Toolkit dataset comprising approximately 150 billion time steps from more than 126,000 sites, WindFM develops a foundational understanding of the complex interplay between atmospheric conditions and power output. Extensive experiments demonstrate that our compact 8.1M parameter model achieves state-of-the-art zero-shot performance on both deterministic and probabilistic tasks, outperforming specialized models and larger foundation models without any fine-tuning. In particular, WindFM exhibits strong adaptiveness under out-of-distribution data from a different continent, demonstrating the robustness and transferability of its learned representations. Our pre-trained model is publicly available at https://github.com/shiyu-coder/WindFM.

Large language models (LLMs) have advanced general-purpose reasoning, showing strong performance across diverse tasks. However, existing methods often rely on implicit exploration, where the model follows stochastic and unguided reasoning paths-like walking without a map. This leads to unstable reasoning paths, lack of error correction, and limited learning from past experience. To address these issues, we propose a framework that shifts from implicit exploration to structured reasoning through guideline and refinement. First, we extract structured reasoning patterns from successful trajectories and reflective signals from failures. During inference, the model follows these guidelines step-by-step, with refinement applied after each step to correct errors and stabilize the reasoning process. Experiments on BBH and four additional benchmarks (GSM8K, MATH-500, MBPP, HumanEval) show that our method consistently outperforms strong baselines across diverse reasoning tasks. Structured reasoning with stepwise execution and refinement improves stability and generalization, while guidelines transfer well across domains and flexibly support cross-model collaboration, matching or surpassing supervised fine-tuning in effectiveness and scalability.

Sixth generation (6G) systems require environment-aware communication, driven by native artificial intelligence (AI) and integrated sensing and communication (ISAC). Radio maps (RMs), providing spatially continuous channel information, are key enablers. However, generating high-fidelity RM ground truth via electromagnetic (EM) simulations is computationally intensive, motivating machine learning (ML)-based RM construction. The effectiveness of these data-driven methods depends on large-scale, high-quality training data. Current public datasets often focus on single-input single-output (SISO) and limited information, such as path loss, which is insufficient for advanced multi-input multi-output (MIMO) systems requiring detailed channel state information (CSI). To address this gap, this paper presents UrbanMIMOMap, a novel large-scale urban MIMO CSI dataset generated using high-precision ray tracing. UrbanMIMOMap offers comprehensive complex CSI matrices across a dense spatial grid, going beyond traditional path loss data. This rich CSI is vital for constructing high-fidelity RMs and serves as a fundamental resource for data-driven RM generation, including deep learning. We demonstrate the dataset's utility through baseline performance evaluations of representative ML methods for RM construction. This work provides a crucial dataset and reference for research in high-precision RM generation, MIMO spatial performance, and ML for 6G environment awareness. The code and data for this work are available at: https://github.com/UNIC-Lab/UrbanMIMOMap.

Differentially Private Stochastic Gradient Descent (DP-SGD) is a standard method for enforcing privacy in deep learning, typically using the Gaussian mechanism to perturb gradient updates. However, conventional mechanisms such as Gaussian and Laplacian noise are parameterized only by variance or scale. This single degree of freedom ties the magnitude of noise directly to both privacy loss and utility degradation, preventing independent control of these two factors. The problem becomes more pronounced when the number of composition rounds T and batch size B vary across tasks, as these variations induce task-dependent shifts in the privacy-utility trade-off, where small changes in noise parameters can disproportionately affect model accuracy. To address this limitation, we introduce PLRV-O, a framework that defines a broad search space of parameterized DP-SGD noise distributions, where privacy loss moments are tightly characterized yet can be optimized more independently with respect to utility loss. This formulation enables systematic adaptation of noise to task-specific requirements, including (i) model size, (ii) training duration, (iii) batch sampling strategies, and (iv) clipping thresholds under both training and fine-tuning settings. Empirical results demonstrate that PLRV-O substantially improves utility under strict privacy constraints. On CIFAR-10, a fine-tuned ViT achieves 94.03% accuracy at epsilon approximately 0.5, compared to 83.93% with Gaussian noise. On SST-2, RoBERTa-large reaches 92.20% accuracy at epsilon approximately 0.2, versus 50.25% with Gaussian.

Abstract--LLM-based agents have shown promise in various cooperative and strategic reasoning tasks, but their effectiveness in competitive multi-agent environments remains underexplored. T o address this gap, we introduce PillagerBench, a novel framework for evaluating multi-agent systems in real-time competitive team-vs-team scenarios in Minecraft. It provides an extensible API, multi-round testing, and rule-based built-in opponents for fair, reproducible comparisons. We also propose T actiCrafter, an LLM-based multi-agent system that facilitates teamwork through human-readable tactics, learns causal dependencies, and adapts to opponent strategies. Our evaluation demonstrates that T actiCrafter outperforms baseline approaches and showcases adaptive learning through self-play. Additionally, we analyze its learning process and strategic evolution over multiple game episodes. T o encourage further research, we have open-sourced PillagerBench, fostering advancements in multi-agent AI for competitive environments. Witnessing rapid advancements, Large Language Models (LLMs) have emerged as powerful tools for complex reasoning, decision-making, and facilitating multi-agent collaboration [27, 20, 22]. This has driven increasing interest in developing cooperative multi-agent systems [3, 7], leading to the creation of benchmarks based on diverse cooperative games such as Minecraft [4] and Overcooked [1]. Minecraft, in particular, has become an important platform due to its open-ended environment and rich state and action spaces [19, 25]. However, current Minecraft-based benchmarks mainly address cooperative tasks characterized by stationary dynamics and fixed objectives, making them insufficient for evaluating adaptability and strategic decision-making in competitive, dynamic environments. Traditional reinforcement learning benchmarks like StarCraft Multi-Agent Challenge (SMAC) [16] and Lux AI Challenge [18] introduce instability and nonstationarity through competitive adversaries but lack the rich, open-ended interactions found in Minecraft. Bridging this gap by integrating both cooperative and competitive elements within a single dynamic environment is essential to rigorously assess the adaptability and generalizability of advanced multi-agent systems.

Grounding object affordance is fundamental to robotic manipulation as it establishes the critical link between perception and action among interacting objects. However, prior works predominantly focus on predicting single-object affordance, overlooking the fact that most real-world interactions involve relationships between pairs of objects. In this work, we address the challenge of object-to-object affordance grounding under limited data contraints. Inspired by recent advances in few-shot learning with 2D vision foundation models, we propose a novel one-shot 3D object-to-object affordance learning approach for robotic manipulation. Semantic features from vision foundation models combined with point cloud representation for geometric understanding enable our one-shot learning pipeline to generalize effectively to novel objects and categories. We further integrate our 3D affordance representation with large language models (LLMs) for robotics manipulation, significantly enhancing LLMs' capability to comprehend and reason about object interactions when generating task-specific constraint functions. Our experiments on 3D object-to-object affordance grounding and robotic manipulation demonstrate that our O$^3$Afford significantly outperforms existing baselines in terms of both accuracy and generalization capability.

ABSTRACT: The accurate long - term forecasting of the El Ni n o Southern Oscillation (ENSO) is still one of the biggest challenges in climate science . While it is true that short - to medium - range performance has been improved significantly using the advances in deep learning, statistical dynamical hybrids, most operational systems still use the simple mean of all ensemble members, implicitly assuming equal skill across members . In this study, w e demonstrate, through a strictly a - posteriori evaluation, for any large enough ensemble of ENSO forecasts, there is a subset of members whose skill is substantially higher than that of the ensemble mean. Using a s tate - of - the - art ENSO forecast system cross - validated against the 1986 - 2017 observed Ni no 3.4 index, we identify two Top - 5 subsets one ranked on lowest Root Mean Square Error (RMSE) and another on highest Pearson correlation. Generally across all leads, these outstanding members show higher correlation and lower RMSE, with the advantage rising enormously with lead time. Whereas at sho rt leads (1 month) raises the mean correlation by about +0.02 (+1.7%) and lowers the RMSE by around 0.14 C or by 23.3% compared to the All - 40 mean, at extreme leads (23 months) the correlation is raised by +0.43 (+172%) and RMSE by 0.18 C or by 22.5% de crease. The enhancements are largest during crucial ENSO transition periods such as SON and DJF, when accurate amplitude and phase forecasting is of greatest socio - economic benefit, and furthermore season - dependent e.g., mid - year months such as JJA and MJJ have incredibly large RMSE reductions. This study provides a solid foundation for further investigations to identify reliable clues for detecting high - quality ensemble members, thereby enhancing forecasting skill. Introduction Long - lead prediction of the El Niño Southern Oscillation (ENSO) is among the most significant and scientifically challenging problems of climate research. ENSO is a coupled ocean atmosphere phenomenon comprising quasi - periodic variations of sea surface temperature (SST) anomalies in the equatorial Pacific with widespread impacts on global weather patterns, hydrology, agriculture, ecosystems, and socio - economic activities [21,23] . Successful prediction at lead times exceeding one year has particular significance for water resources management planning, disaster preparedness, agricultural planning, and climate - sensitive economic practice [24,25] . Howe ver, the inherent nonlinearity of ocean atmosphere interaction, the sensitivity to initial conditions, and the complex web of teleconnections controlling ENSO variability make the forecast skill decline very quickly with lead time.

We present Beamforming-LLM, a system that enables users to semantically recall conversations they may have missed in multi-speaker environments. The system combines spatial audio capture using a microphone array with retrieval-augmented generation (RAG) to support natural language queries such as, "What did I miss when I was following the conversation on dogs?" Directional audio streams are separated using beamforming, transcribed with Whisper, and embedded into a vector database using sentence encoders. Upon receiving a user query, semantically relevant segments are retrieved, temporally aligned with non-attended segments, and summarized using a lightweight large language model (GPT-4o-mini). The result is a user-friendly interface that provides contrastive summaries, spatial context, and timestamped audio playback. This work lays the foundation for intelligent auditory memory systems and has broad applications in assistive technology, meeting summarization, and context-aware personal spatial computing.

We present a deep learning approach to classify fast radio bursts (FRBs) based purely on morphology as encoded on recorded dynamic spectrum from CHIME/FRB Catalog 2. We implemented transfer learning with a pretrained ConvNext architecture, exploiting its powerful feature extraction ability. ConvNext was adapted to classify dedispersed dynamic spectra (which we treat as images) of the FRBs into one of the two sub-classes, i.e., repeater and non-repeater, based on their various temporal and spectral properties and relation between the sub-pulse structures. Additionally, we also used mathematical model representation of the total intensity data to interpret the deep learning model. Upon fine-tuning the pretrained ConvNext on the FRB spectrograms, we were able to achieve high classification metrics while substantially reducing training time and computing power as compared to training a deep learning model from scratch with random weights and biases without any feature extraction ability. Importantly, our results suggest that the morphological differences between CHIME repeating and non-repeating events persist in Catalog 2 and the deep learning model leveraged these differences for classification. The fine-tuned deep learning model can be used for inference, which enables us to predict whether an FRB's morphology resembles that of repeaters or non-repeaters. Such inferences may become increasingly significant when trained on larger data sets that will exist in the near future.

Abstract--This paper presents MSLEF, a multi-segment ensemble framework that employs LLM fine-tuning to enhance resume parsing in recruitment automation. It integrates fine-tuned Large Language Models (LLMs) using weighted voting, with each model specializing in a specific resume segment to boost accuracy. Building on MLAR [1], MSLEF introduces a segment-aware architecture that leverages field-specific weighting tailored to each resume part, effectively overcoming the limitations of single-model systems by adapting to diverse formats and structures. MSLEF achieves significant improvements in Exact Match (EM), F1 score, BLEU, ROUGE, and Recruitment Similarity (RS) metrics, outperforming the best single model by up to +7% in RS. Its segment-aware design enhances generalization across varied resume layouts, making it highly adaptable to real-world hiring scenarios while ensuring precise and reliable candidate representation. Recruitment automation has transformed hiring by enabling efficient processing of numerous applications, with resume parsing extracting structured data like experience, education, skills, and contact information.