A lot of people start algotrading by training an LSTM to predict the next bar’s close. I did too, until I realized trading is a control problem, not a prediction problem. A supervised model predicting a price move with 53% accuracy can still lose money once you factor in fees, slippage, and path-dependent equity. I recently finished a deep-dive on my autonomous trading architecture, which runs a single Recurrent Soft Actor-Critic (SAC) agent managing a portfolio of six […]

Ver mais

arXiv:2603.23686v1 Announce Type: new Abstract: 3D Gaussian Splatting (3DGS) is increasingly recognized as a powerful paradigm for real-time, high-fidelity 3D reconstruction. However, its per-scene optimization pipeline limits scalability and generalization, and prevents efficient inference. Recently emerged feed-forward 3DGS models address these limitations by enabling fast reconstruction from a few input views after large-scale pretraining, without scene-specific optimization. Despite their advantages and strong potential for commercial deployment, the use of neural networks as the backbone also amplifies the risk […]

Ver mais

arXiv:2603.24648v1 Announce Type: new Abstract: Anomaly detection is a core service in the Internet of Underwater Things, yet training accurate distributed models underwater is difficult because acoustic links are low-bandwidth, energy-intensive, and often unable to support direct sensor-to-surface communication. Standard flat federated learning therefore faces two coupled limitations in underwater deployments: expensive long-range transmissions and reduced participation when only a subset of sensors can reach the gateway. This paper proposes an energy-efficient hierarchical federated learning framework for underwater […]

Ver mais

Over the last 12 months, I’ve built a model to predict the next token and to suggest synonyms or related queries to a user prompt, with 100% correct predictions on the training set in one shot, without training or deep neural networks (DNNs). The same model is now integrated in some of the most recent LLM architectures, albeit with costly training via DNNs. My version does not need DNNs or training. The purpose of this article is to […]

Ver mais

Learning from electronic health records (EHRs) time series is challenging due to irregular sam- pling, heterogeneous missingness, and the resulting sparsity of observations. Prior self-supervised meth- ods either impute before learning, represent missingness through a dedicated input signal, or optimize solely for imputation, reducing their capacity to efficiently learn representations that support clinical downstream tasks. We propose the Augmented-Intrinsic Dual-Masked Autoencoder (AID-MAE), which learns directly from incomplete time series by applying an intrinsic missing mask to represent naturally […]

Ver mais

arXiv:2602.13202v1 Announce Type: new Abstract: Non-Orthogonal Multiple Access (NOMA) is an information-theoretical approach used in 5G networks to improve spectral efficiency, but it is prone to interference during handovers. In this work, we propose a hybrid method that combines Gold-Walsh modulated sequences with Deep Q-Networks (DQN) to intelligently manage interference during NOMA handovers. This method optimizes sequence selection and power allocation dynamically. As a result, it achieves a 95.2% handover success rate, which is an improvement of up […]

Ver mais

arXiv:2603.12439v1 Announce Type: new Abstract: This paper presents a Lyapunov-Halanay method to study global asymptotic stabilization (GAS) of nonlinear retarded systems subject to large constant delays in input/output – a challenging problem due to their inherent destabilizing effects. Under the conditions of global Lipschitz continuity (GLC) and global exponential stabilizability (GES) of the retarded system without input delay, a state feedback controller is designed based on sequential predictors to make the closed-loop retarded system GAS. Moreover, if the […]

Ver mais

7B Models: Cheap, Fast… and Brutally Honest About Your Prompting If you’ve deployed a 7B model locally (or on a modest GPU), you already know the trade: Pros low cost low latency easy to self-host Cons patchy world knowledge weaker long-chain reasoning worse instruction-following unstable formatting (“JSON… but not really”) The biggest mistake is expecting 7B models to behave like frontier models. They won’t. But you can get surprisingly high-quality output if you treat prompting like systems design, […]

Ver mais

arXiv:2605.20271v1 Announce Type: new Abstract: We develop a rigorous statistical theory of multi-head attention (MHA) as an ensemble of Nadaraya-Watson (NW) kernel regression estimators. Building on the algebraic identity between single-head softmax attention and the NW estimator, we prove that MHA is a structured ensemble of H NW estimators, each operating in a distinct learned projection subspace of the key space. We derive an explicit Bias-Variance-Covariance decomposition of the MHA mean squared error, showing that variance reduction depends […]

Ver mais

arXiv:2604.08618v1 Announce Type: new Abstract: Deploying LLM-powered agents in enterprise scenarios such as cloud technical support demands high-quality, domain-specific skills. However, existing skill creators lack domain grounding, producing skills poorly aligned with real-world task requirements. Moreover, once deployed, there is no systematic mechanism to trace execution failures back to skill deficiencies and drive targeted refinements, leaving skill quality stagnant despite accumulating operational evidence. We introduce SkillForge, a self-evolving framework that closes an end-to-end creation-evaluation-refinement loop. To produce well-aligned […]

Ver mais