EurekAlert!

AI boosts understanding of ocean dynamics and marine structure safety

Fluid–structure interaction (FSI) governs how flowing water and air interact with marine structures—from wind turbines to ...
marktechpost

Google Introduces T5Gemma 2: Encoder Decoder Models with Multimodal Inputs via SigLIP and 128K Context

T5Gemma 2 follows the same adaptation idea introduced in T5Gemma, initialize an encoder-decoder model from a decoder-only checkpoint, then adapt with UL2. In the above figure the research team show ...
MarketWatch

Ampace and DG Matrix Announce Strategic Partnership to Deliver First UL-Certified Grid-Active BESS Integrated with Multi-Port Solid-State Transformer Architecture for AI Data ...

The MarketWatch News Department was not involved in the creation of this content. RALEIGH, N.C., Dec. 16, 2025 /PRNewswire/ -- Ampace, a global leader in advanced lithium-ion energy storage, today ...
VentureBeat

Nvidia debuts Nemotron 3 with hybrid MoE and Mamba-Transformer to drive efficient agentic AI

Nvidia launched the new version of its frontier models, Nemotron 3, by leaning in on a model architecture that the world’s most valuable company said offers more accuracy and reliability for agents.
Frontiers

A hybrid Spiking Neural Network–Transformer architecture for motor imagery and sleep apnea detection

Department of Computer Science and Engineering, Faculty of Applied Sciences, University of West Bohemia in Pilsen, Pilsen, Czechia Introduction: Motor imagery (MI) classification and sleep apnea (SA) ...
Wall Street Journal

An AI Startup Looks Toward the Post-Transformer Era

Most of the worries about an AI bubble involve investments in businesses that built their large language models and other forms of generative AI on the concept of the transformer, an innovative type ...
Geeky Gadgets

IBM Granite 4.0 : Tiny AI Runs Offline, Slicing Costs for Real Workflows

What if the future of artificial intelligence wasn’t just about being bigger, faster, or more powerful, but smaller, smarter, and more accessible? That’s the promise of IBM’s Granite 4, a new leap in ...
IEEE

HiTrans-SAM: Hierarchical Transformer Encoder and SAM-Augmented Inputs for Multi-Scale Remote Sensing Image Segmentation

Abstract: Semantic segmentation of remote sensing images is challenging due to complex scenes, substantial variations in object scales, and ambiguous boundaries. In this study, we propose a novel ...
marktechpost

This AI Paper Proposes a Novel Dual-Branch Encoder-Decoder Architecture for Unsupervised Speech Enhancement (SE)

Most learning-based speech enhancement pipelines depend on paired clean–noisy recordings, which are expensive or impossible to collect at scale in real-world conditions. Unsupervised routes like ...
InfoWorld

IBM launches Granite 4.0 to cut AI infra costs with hybrid Mamba-transformer models

Built for long-context tasks and edge deployments, Granite 4.0 combines Mamba’s linear scaling with transformer precision, offering enterprises lower memory usage, faster inference, and ISO ...