Neural Networks

Researchers propose a frequency-guided feature representation learner for small object detection that shifts from spatial to spectral processing. The method, instantiated via lightweight plug-and-play modules, achieves c…

Researchers propose EMRFormer, a spiking neural network architecture that applies spike-driven transformers to automatic modulation recognition, achieving state-of-the-art accuracy while reducing theoretical energy consu…

Researchers developed an automated pipeline to explore heterogeneous 4-expert mixture-of-experts architectures, generating 4,463 candidate models over 28 days. The search was biased toward the AirNet family due to an enu…

Researchers introduced the apothem measure for computing robustness certifications in neural networks, achieving optimal certifications with linear oracle calls. They proved volume-optimal oracle-based algorithms are int…

Researchers propose ModTGCN, a modularity-aware graph neural network for text classification that jointly optimizes cross-entropy and a modularity-based auxiliary objective to promote class-coherent document communities.…

Researchers argue that vanilla conditional diffusion models are fundamentally incapable of compositional generation when target distributions are out-of-distribution, as score estimation errors cause catastrophic perform…

Researchers introduced Neuro-Symbolic Drive, a framework that supervises driving vision-language-action (VLA) models with rule-grounded reasoning traces from classical rule-based planners, improving trajectory prediction…

A new framework proposes four dimensions—diffuseness, agency, severity, and probability—to quantify and tame fear, mapping emotions like dread and panic onto numerical vectors. The model aims to help people study and red…

Engineers can now understand the transformer architecture behind large language models through a plain-English guide that explains how words become vectors, layers move points, and attention mechanisms allow context-awar…

The field of artificial intelligence marks its 70th anniversary since being formally established at the 1956 Dartmouth Summer Research Project. AI has evolved from early neural network concepts by McCulloch and Pitts and…

A study found that positive face-to-face conversations between mothers and children temporarily align their neural activity even after the interaction ends, suggesting everyday social experiences may shape brain developm…

SonicLAB released SSNN, a live audio processor plugin embedding a spiking neural network with 960 neurons across 32 layers, driving eight synthesis engines in real time. The plugin is available for Windows and Mac in VST…

Researchers trained deep MLPs to memorize secret binary strings and found that naturally learned behaviors resist detection by both human-made input-optimization algorithms and learned algorithms, posing a challenge for …

Engineers are increasingly using machine learning surrogate models to accelerate computational fluid dynamics (CFD) and finite element analysis (FEA) simulations, replacing traditional hour-long solver runs with faster p…

A developer traces the 70-year history of AI from symbolic systems to modern transformer-based models, highlighting key milestones like the 1956 Dartmouth Workshop, expert systems, the first AI winter, the rise of statis…

MiniMax introduced sparse key-value caches that reduce attention scaling from quadratic to near-linear, enabling practical multi-hundred-kilobyte context windows on a single GPU. The method cuts per-token attention compu…

Independent AI researcher YoungSeong Kim published 'Window is Everything' (2025), a unified theory of neural operations that decomposes any operation into Path, Shape, and Weight components. The paper gained over 17,000 …

Chinese researchers developed a flexible brain implant electrode array thinner than a human hair that records neural activity with exceptional clarity and remains stable for at least 18 months, overcoming the rigidity mi…

A developer implemented an LSTM from scratch using PyTorch and Lightning AI, detailing the initialization of weights and biases for the forget, input, cell candidate, and output gates. The tutorial uses the Adam optimize…

Researchers at GSI/FAIR and international collaborators developed RHINE, a machine learning model using deep neural networks to simulate r-process nucleosynthesis in extreme astrophysical events like neutron star mergers…