February 2026

This paper introduces a novel end-to-end framework leveraging Conformer architectures to unify the traditionally fragmented pipeline of hearing-to-speech-to-writing language processing. Unlike conventional automatic speech recognition (ASR) systems that cascade separate acoustic, phonetic, and linguistic models prone to cascading errors our approach employs stacked Conformer encoders, which integrate convolution-augmented transformers to capture both local spectral nuances and long-range contextual dependencies in raw audio inputs. The model processes mel-spectrograms directly into intermediate speech representations and final textual outputs via a […]

Ver mais

We study wide Bayesian neural networks focusing on the rare but statistically dominant fluctuations that govern posterior concentration, beyond Gaussian-process limits. Large-deviation theory provides explicit variational objectives-rate functions-on predictors, providing an emerging notion of complexity and feature learning directly at the functional level. We show that the posterior output rate function is obtained by a joint optimization over predictors and internal kernels, in contrast with fixed-kernel (NNGP) theory. Numerical experiments demonstrate that the resulting predictions accurately describe finite-width […]

Ver mais

This quiz sharpens your intuition for Python’s asyncio module. You’ll decide when async is the right tool, see how the event loop schedules work, and understand how coroutines pause and resume around I/O. Along the way, you’ll revisit async and await, coroutine creation, async generators, asyncio.run(), and concurrent execution with asyncio.gather(). For a quick refresher before you start, check out Hands-On Python 3 Concurrency With the asyncio Module. [ Improve Your Python With 🐍 Python Tricks 💌 – […]

Ver mais

Implementations of symmetric positive definite (SPD) matrix-based neural networks for neural decoding remain fragmented across research codebases and Python packages. Existing implementations often employ ad hoc handling of manifold constraints and non-unified training setups, which hinders reproducibility and integration into modern deep-learning workflows. To address this gap, we introduce SPD Learn, a unified and modular Python package for geometric deep learning with SPD matrices. SPD Learn provides core SPD operators and neural-network layers, including numerically stable spectral operators, […]

Ver mais

This study proposes a federated contrastive learning based distributed anomaly detection framework to address privacy protection requirements in IoT environments. The framework builds local encoders on each node to embed high-dimensional time series and network behavior features, and uses representation alignment to reduce distribution differences across devices. Based on this, a contrastive learning objective is introduced to strengthen the compactness of normal patterns in the latent space and to enlarge the boundary between normal and abnormal features, which […]

Ver mais

Amortized Bayesian Inference (ABI) enables efficient posterior estimation using generative neural networks trained on simulated data, but often suffers from performance degradation under model misspecification. While self-consistency (SC) training on unlabeled empirical data can enhance network robustness, current approaches are limited to static, single-task settings and fail to handle sequentially arriving data or distribution shifts. We propose a continual learning framework for ABI that decouples simulation-based pre-training from unsupervised sequential SC fine-tuning on real-world data. To address the […]

Ver mais

Software development has undergone a significant transformation over the last 20 years, transitioning away from the traditional method of manual coding and moving toward AI-assisted programming and collaborative version control, or AI copilots that provide code snippet suggestions and significantly increase productivity. Agentic coding in 2026 is expected to be fully developed and available to all development environments, providing the basis for the next phase of advancement in AI-assisted software development. Traditional AI coding tools, which required significant […]

Ver mais

This paper proposes a graph-structured temporal dynamic learning model to address the challenges of backend load prediction in cloud computing and microservice environments, including dynamic topology changes, complex dependency structures, and multi-source heterogeneous monitoring data. The model constructs a time-varying service dependency graph to adaptively model structural relationships among nodes and integrates a temporal encoding mechanism to capture multi-scale temporal features, achieving joint representation of load characteristics in both spatial and temporal dimensions. It consists of four main […]

Ver mais

The identification and property prediction of chemical molecules is of central importance in the advancement of drug discovery and material science, where the tandem mass spectrometry technology gives valuable fragmentation cues in the form of mass-to-charge ratio peaks. However, the lack of experimental spectra hinders the attachment of each molecular identification, and thus urges the establishment of prediction approaches for computational models. Deep learning models appear promising for predicting molecular structure spectra, but overall assessment remains challenging as […]

Ver mais