Neural Networks

Meta's FAIR AI team unveiled Brain2Qwerty v2, a non-invasive system that translates brain activity into typed sentences by reading magnetic signals from outside the skull. The technology aims to provide a communication a…

A developer built AdaBoost from scratch and created an interactive demo showing how the algorithm combines hundreds of weak classifiers—decision stumps—into a strong ensemble. The implementation uses weighted error minim…

Meta's FAIR team has developed Brain2Qwerty v2, a non-invasive AI system that translates brain activity into typed sentences by reading magnetic signals outside the skull. The system's accuracy is improving with each add…

Researchers tasked LLM agents with inventing concept-erasure algorithms that outperform existing methods like LEACE and QLEACE. The best agent-discovered algorithm reduced nonlinear probe accuracy from 99% to 70%, compar…

Steve Lucas, CEO of Boomi, told ZDNET that the frontier engineer—a specialist with advanced data and neural networking skills—will become the key enterprise AI expert needed to optimize frontier models and unlock competi…

Deep learning, the core technology behind AI applications like ChatGPT, image recognition, and self-driving cars, is explained as a method of teaching computers to learn patterns from large datasets using layered neural …

Researchers introduced MPL-MAE, a new 3D autoencoding framework that reduces positional bias to enhance semantic learning in point cloud processing. The framework uses a gated positional interface module to balance spati…

Optimizing joint OFDM-RIS systems for 6G networks faces challenges including a lack of standardized benchmarks and scalability issues, with 78 attempts between 2021 and 2026 falling into four paradigms. ML-based methods …

Researchers have developed a domain-decomposed neural network framework that solves partial differential equations on unbounded domains with improved accuracy and flexibility. The method assigns separate randomized subne…

Researchers introduced LINet, a multi-stream neural network for RGB-D scene classification that outperforms existing methods through a continuous integration strategy. LINet uses a novel Linear Integration Convolution op…

TF-MoE, a sparse Mixture-of-Experts framework, enhances speech separation model capacity without increasing computational costs, outperforming existing methods. It achieves a +3.8 dB SDR improvement on Libri2Mix with onl…

Researchers improved associative recall in recurrent neural networks by orthogonalizing the mLSTM memory matrix during reads, inspired by the Muon optimizer. In noisy associative recall tasks, the method boosted success …

Researchers at MIT found that large language models develop a modular cognitive architecture mirroring the human brain, with distinct neuron sets supporting language, formal reasoning, social reasoning, and physical reas…

Researchers introduce Depth-wise Gradient Augmentation, a new optimization paradigm that transforms layer-wise updates to exploit structured relationships across deep neural network layers. Their instantiation, Gradient …

Researchers propose the Quality Representation Module (QRM), a lightweight transformer module that learns a quality-aware representation to modulate diffusion transformers (DiT) for text-to-image generation, improving im…

Researchers introduced the Neuro-Bayesian-Symbolic Residual Attention Shallow Network (NBS-RASN), a hybrid neural architecture for explainable cybersecurity risk assessment in open-source ecosystems. The shallow network …

Researchers introduced MCO-PDE, a competitive optimization framework that discovers shared partial differential equations from multi-source datasets by fusing as few as 50 observations per dataset. The framework uses neu…

SKT AI Labs has released a new experimental technique called Weight Manifold Fusion (WMF) and invites researchers and developers to test it on the SKT-SURYA-H model. The company seeks community feedback to refine the arc…

Researchers introduced GAT, a purely transformer-based Generative Adversarial Network trained in a compact Variational Autoencoder latent space, achieving state-of-the-art single-step class-conditional generation on Imag…

A single resting-state functional MRI scan can predict depression and anxiety symptoms in adolescents one year later, according to a study of 150 participants. The predictive marker, derived from a whole-brain generative…