arXiv:2603.06604v1 Announce Type: new Abstract: As large language models (LLMs) are increasingly deployed in critical decision-making systems, the lack of reliable methods to measure their uncertainty presents a fundamental trustworthiness risk. We introduce a normalized confidence score based on output anchor token probabilities: classification labels for structured tasks and self-evaluation responses (Yes/No) for open-ended generation. This enables direct detection of errors and hallucinations with minimal overhead and without external validation. We make three key contributions. First, we propose […]

Ver mais

We study offline learning in KL-regularized two-player zero-sum games, where policies are optimized under a KL constraint to a fixed reference policy. Prior work relies on pessimistic value estimation to handle distribution shift, yielding only $widetilde{mathcal{O}}(1/sqrt n)$ statistical rates. We develop a new pessimism-free algorithm and analytical framework for KL-regularized games, built on the smoothness of KL-regularized best responses and a stability property of the Nash equilibrium induced by skew symmetry. This yields the first $widetilde{mathcal{O}}(1/n)$ sample complexity […]

Ver mais

Anyone who works with PyTorch model code starts asking the same questions: Why is this taking so long? How do I make my training loop faster? Whether you’re an ML engineer, a researcher or just decided to play around with a random ML repository over the weekend, you will eventually try to understand how to speed your code up. However, before we can do that, we need to learn how to measure performance correctly. And then draw the […]

Ver mais

Generating realistic graph-structured data is challenging due to discrete structures, variable sizes, and class-specific connectivity patterns that resist conventional generative modelling. While recent graph generation methods employ generative adversarial network (GAN) frameworks to handle permutation invariance and irregular topologies, they typically rely on random edge sampling with fixed probabilities, limiting their capacity to capture complex structural dependencies between nodes. We propose a density-aware conditional graph generation framework using Wasserstein GANs (WGAN) that replaces random sampling with a learnable […]

Ver mais

Constrained autoencoders (cAE) provide a successful path towards interpretable dimensionality reduction by enforcing geometric structure on latent spaces. However, standard cAEs cannot adapt to varying physical parameters or environmental conditions without conflating these contextual shifts with the primary input. To address this, we integrated a neuromodulatory mechanism into the cAE framework to allow for context-dependent manifold learning. This paper introduces the Neuromodulated Constrained Autoencoder (NcAE), which adaptively parameterizes geometric constraints via gain and bias tuning conditioned on static […]

Ver mais

Vehicular Ad-hoc Networks (VANETs) face critical challenges regarding intermittent connectivity and latency due to high node mobility, often resulting in a performance trade-off between reactive and proactive routing paradigms. This study aims to resolve these inherent limitations and ensure reliable communication in volatile environments. We propose a novel context-aware framework, the Dynamic Hybrid Routing Protocol (DHRP), which integrates Ad hoc On-Demand Distance Vector (AODV) and Optimized Link State Routing (OLSR). Distinguished by a predictive multi-criteria switching logic and […]

Ver mais

arXiv:2506.05059v3 Announce Type: replace-cross Abstract: Deep learning has achieved remarkable success across many domains, but it has also created a growing demand for interpretability in model predictions. Although many explainable machine learning methods have been proposed, post-hoc explanations lack guaranteed fidelity and are sensitive to hyperparameter choices, highlighting the appeal of inherently interpretable models. For example, linear regression provides clear feature effects through its coefficients. However, such models are often outperformed by more complex neural networks (NNs) that […]

Ver mais

Deep Learning (DL) has undergone widespread adoption for medical image analysis and diagnosis. Numerous studies have explored mammographic image analysis for breast cancer screening. For this study, we assessed the hypothesis that stratifying mammography images based on the presence or absence of a corresponding region of interest (ROI) improves classification accuracy for both normal–abnormal and benign–malignant classifications. Our methodology involves independently training models and performing predictions on each subgroup with subsequent integration of the results. We used several […]

Ver mais

arXiv:2604.02343v1 Announce Type: new Abstract: We study the compression of LLM-generated text across lossless and lossy regimes, characterizing a compression-compute frontier where more compression is possible at the cost of more compute. For lossless compression, domain-adapted LoRA adapters can improve LLM-based arithmetic coding by 2x over compression with the base LLM alone. For lossy compression, prompting a model for a succinct rewrite then applying arithmetic coding can achieve compression ratios of approximately 0.03, a 2x improvement over compressing […]

Ver mais