Tagged articles

7 articles

Page 1 of 1

May 13, 2026 · Artificial Intelligence

How a 0.5 MB AI Model Tackles Global Supply‑Chain Challenges: Li‑Net in Action

Li‑Net, a 0.5 MB multi‑channel time‑series model co‑developed by SF Technology and Chinese universities, achieves state‑of‑the‑art accuracy with linear‑complexity attention, runs on edge devices, and has been deployed across SF's global supply‑chain for demand forecasting, inventory optimization, and capacity planning, delivering measurable cost reductions.

AILi-NetSupply Chain

0 likes · 4 min read

How a 0.5 MB AI Model Tackles Global Supply‑Chain Challenges: Li‑Net in Action

AntTech

Feb 5, 2026 · Artificial Intelligence

How Triple Alignment and Rationale Generation Supercharge Knowledge‑Based VQA

This paper presents a lightweight, high‑efficiency framework called Triple Alignment with Rationale Generation (TAG) that transforms knowledge‑based visual question answering into a contrastive learning task, dramatically reducing trainable parameters while achieving state‑of‑the‑art performance on major KVQA benchmarks.

CLIPVQAcontrastive learning

0 likes · 7 min read

How Triple Alignment and Rationale Generation Supercharge Knowledge‑Based VQA

Old Zhang's AI Learning

Jan 31, 2026 · Artificial Intelligence

How a 0.1B‑Parameter OCR Model Beats Multi‑Billion‑Parameter Vision‑Language Models

UniRec‑0.1B, a lightweight OCR model with only 0.1 B parameters, achieves accuracy comparable to or better than multi‑billion‑parameter visual‑language models across text, formula, and mixed‑content tasks, thanks to hierarchical supervision training, a semantic‑decoupled tokenizer, and a large 40 M‑sample dataset, while delivering 2‑9× faster inference and full open‑source availability.

Hierarchical SupervisionOCRSemantic Decoupled Tokenizer

0 likes · 12 min read

How a 0.1B‑Parameter OCR Model Beats Multi‑Billion‑Parameter Vision‑Language Models

Xiaohongshu Tech REDtech

Feb 27, 2025 · Artificial Intelligence

SAFE: A Lightweight General AI Image Detection Method Achieving 96.7% Accuracy Across 33 Test Subsets

SAFE is a lightweight AI‑image detection framework using only 1.44 M parameters and 2.30 B FLOPs that preserves fine‑grained artifacts through crop‑based preprocessing, invariant augmentations, and high‑frequency wavelet features, achieving an average 96.7 % accuracy across 33 test subsets and strong generalization to unseen GAN and diffusion generators.

AI image detectionComputer VisionDeep Learning

0 likes · 11 min read

SAFE: A Lightweight General AI Image Detection Method Achieving 96.7% Accuracy Across 33 Test Subsets

AIWalker

Jan 12, 2025 · Artificial Intelligence

CubeFormer: A Simple Yet Effective Lightweight Image Super‑Resolution Baseline

CubeFormer introduces a novel cube attention mechanism and dual transformer blocks that dramatically improve feature diversity, enabling a lightweight image super‑resolution model to achieve state‑of‑the‑art PSNR and visual detail across multiple benchmarks while keeping parameters low.

Computer VisionDeep Learningcube attention

0 likes · 21 min read

CubeFormer: A Simple Yet Effective Lightweight Image Super‑Resolution Baseline

AntTech

Aug 25, 2023 · Artificial Intelligence

LayoutGCN: A Lightweight Graph Convolutional Network for Visually Rich Document Understanding

LayoutGCN is a lightweight, graph‑based framework that jointly encodes text, layout, and image features of visually rich documents, achieving competitive performance on multiple downstream tasks while drastically reducing model size and computational cost, making it suitable for edge deployment.

Graph Neural NetworkLayoutGCNdocument understanding

0 likes · 24 min read

LayoutGCN: A Lightweight Graph Convolutional Network for Visually Rich Document Understanding

Bilibili Tech

Sep 9, 2022 · Artificial Intelligence

Visual Lossless Deep Learning Pre‑processing for Video Transcoding Using DCT‑Based Low‑Rank Loss and a Lightweight Model

A visual‑lossless deep‑learning pre‑processor that employs a DCT‑based low‑rank loss and an ultra‑lightweight CPU‑friendly model achieves up to 20% bitrate reduction for 1080p videos while preserving high‑frequency details, enabling real‑time transcoding and bandwidth savings for popular content on Bilibili.

AIDCTDeep Learning