Results 101 to 110 of about 104,311 (276)
Insertion-based Decoding with Automatically Inferred Generation Order
Conventional neural autoregressive decoding commonly assumes a fixed left-to-right generation order, which may be sub-optimal. In this work, we propose a novel decoding algorithm— InDIGO—which supports flexible sequence generation in arbitrary orders ...
Gu, Jiatao, Liu, Qi, Cho, Kyunghyun
doaj +1 more source
Grounding Large Language Models for Robot Task Planning Using Closed‐Loop State Feedback
BrainBody‐Large Language Model (LLM) introduces a hierarchical, feedback‐driven planning framework where two LLMs coordinate high‐level reasoning and low‐level control for robotic tasks. By grounding decisions in real‐time state feedback, it reduces hallucinations and improves task reliability.
Vineet Bhat +4 more
wiley +1 more source
Tiny TR-CAP: A novel small-scale benchmark dataset for general-purpose image captioning tasks
In the last decade, the outstanding performance of deep learning has also led to a rapid and inevitable rise in automatic image captioning, as well as the need for large amounts of data.
Abbas Memiş, Serdar Yıldız
doaj +1 more source
This work develops dynamically softening polyacrylamide hydrogels for time‐resolved imaging during continuous mechanical transitions. The study revealed that mechanotransduction is biphasic; YAP/TAZ inactivation is driven by early loss of the nucleocytoskeletal continuum connecting subnuclear adhesions, F‐actin, and the nuclear envelope, coupled with ...
Alessandro Gandin +12 more
wiley +1 more source
An entity‐centric foundation model, GloPath, is introduced for comprehensive glomerular lesion assessment from routine renal biopsy images. Trained on over one million glomeruli, the framework enables robust lesion recognition, grading, and cross modality diag nosis, while uncovering large‐scale clinicopathological associations.
Qiming He +28 more
wiley +1 more source
Cap2Seg: leveraging caption generation for enhanced segmentation of COVID-19 medical images
Incorporating medical text annotations compensates for the quality deficiencies of image data, effectively overcoming the limitations of medical image segmentation.
Wanlong Zhao +9 more
doaj +1 more source
Grounded Video Caption Generation
We propose a new task, dataset and model for grounded video caption generation. This task unifies captioning and object grounding in video, where the objects in the caption are grounded in the video via temporally consistent bounding boxes. We introduce the following contributions.
Kazakos, Evangelos +2 more
openaire +2 more sources
Growth‐Like Swelling of 2D Precursors for Efficient Fabrication of Complex 3D Microstructures
The growth‐like swelling process fills photocured polymers into the elastomeric microstructures, inducing a significant swelling deformation of microstructures that can be maintained in the absence of solvents. The growth‐like swelling process breaks through the dependence of swelling deformation on the solvents, thereby expanding the application of 3D
Xinkai Zhu +13 more
wiley +1 more source
Lymphoid‐Tissue‐on‐Chip Recapitulates Human Antibody Responses In Vitro
The presented lymphoid‐tissue‐on‐chip system allows culture of primary human tonsil cells at organotypic high density under perfusion for up to 4 weeks, emulates immune response to soluble vaccines and vaccination via peripheral antigen‐presenting cells and represents a useful tool to assess cellular interactions during homeostasis, immune responses ...
Claudia Teufel +15 more
wiley +1 more source
A bifunctional lactate oxidase‐like and catalase‐like artificial enzyme (Metazyme) is integrated into a rod‐shaped microgel (MetaRgel) to enable cascade lactate oxidation and oxygen regeneration. By reprogramming the wound metabolic microenvironment, MetaRgel alleviates excessive lactate accumulation, oxidative stress, hypoxia, and inflammation ...
Yongyuan Kang +9 more
wiley +1 more source

