Nvidia's Nemotron-Cascade 2 wins math and coding gold medals with 3B active parameters — and its post-training recipe is now open-source

Nvidia's Nemotron-Cascade 2 is a 30B MoE model that activates only 3B parameters at inference time, yet achieved gold medal-level performance at the 2025 IMO, IOI, and ICPC World Finals. Nvidia has ...
eLife

The Locomoting State Selectively Amplifies Activity of Sensitizing Neurons in Primary Visual Cortex

This manuscript presents a valuable analysis of how locomotion modulates the activity of different subtypes of cortical neurons in the mouse primary visual cortex, showing that locomotion more ...

Train-to-Test scaling explained: How to optimize your end-to-end AI compute budget for inference

AI reasoning does not necessarily require spending huge amounts on frontier models. Instead, smaller models can yield ...
EurekAlert!

Teaching neural networks where errors matter

Physics-informed neural networks (PINNs) have shown remarkable prospects in solving forward and inverse problems involving ...
eLife

Serotonergic modulation of motor subspace dynamics drives a sleep-independent quiescent state

The dorsal raphe nucleus (DRN) serotonergic (5-HT) system has been implicated in regulating sleep and motor control; however, its specific role remains controversial. In this study, we found that ...
BBC

Claude Code users hitting usage limits 'way faster than expected'

Anthropic says it is looking to resolve an issue which is blocking users of its AI coding tool. Claude Code, the AI-powered helper for writing computer code, has become popular in recent months. The ...
IEEE

Electromagnetic Parameter and Thickness Estimation Using Pulsed Eddy Current and Physics-Informed Transformer Networks

Abstract: Pulsed eddy current (PEC) has been widely employed in the metal industry for estimating the electromagnetic parameter and thickness of steel plates, owing to its unique advantages, such as ...
IEEE

An Efficient Multi-Estimation-Based Parameter Centroid Decision Via Linear Regression Approach

Abstract: We propose a novel post-processing approach for the local optimization of Locally Optimized RANdom SAmple Consensus (LO-RANSAC), called the Multi-Estimation-based Parameter Centroid (MEPC) ...
GitHub

train_validate.m

Rx = real_imag_consits(fake_samples_list_test(i,:),ones(1,M),M); %Rx = real_imag_consits(Test_R_clean(i,:),ones(1,M),M); legend({'Jammer','Target','Train'},'FontSize ...