KAIST researchers introduced Agent‑X, a pure‑software acceleration framework that eliminates prefill and decode bottlenecks on mobile devices, achieving a 1.61× end‑to‑end speedup for on‑device AI agents without any loss in task accuracy.
PhoneClaw and PokeClaw are open‑source, on‑device AI agents for iOS and Android that run Gemma 4 locally, offering offline privacy, zero‑cost operation, and native tool calling through iOS APIs or Android Accessibility services.
Google’s open‑source AI Edge Gallery lets Android and iOS devices run large language models such as Gemma 4 entirely offline, eliminating network latency and privacy concerns; the app showcases six modular AI features, offers a simple install path, and signals Google’s push toward a standardized edge‑AI ecosystem.
This article explains how to implement on‑device Retrieval‑Augmented Generation (RAG) for large language models, covering embedding, vector indexing, model selection, quantization, data chunking, incremental updates, hybrid search, and agentic RAG to deliver fast, private, and personalized AI experiences on mobile devices.
OPPO AI Center introduces AndesVL, an open‑source, fully‑adapted multimodal large model ranging from 0.6B to 4B parameters, designed for high‑performance, privacy‑preserving, low‑latency AI on mobile devices, with advanced architecture, training pipelines, on‑device optimizations, and state‑of‑the‑art benchmark results.
This article walks through the practical steps of integrating Google ML Kit into an Android app, covering its privacy‑first, zero‑learning‑curve advantages and providing complete code examples for barcode scanning, OCR, error handling, CameraX setup, and performance tuning.
Apple recently open‑sourced its FastVLM and MobileCLIP2 models, showcasing a multimodal vision‑language system that runs up to 85 times faster than comparable models, enabling real‑time AI on iPhones and other edge devices while illustrating Apple’s broader “B‑plan” of on‑device small‑model AI strategy.
This article details a complete engineering solution for bringing on‑device AI to WeChat mini‑programs, comparing TensorFlow.js and WeChat native inference, covering model conversion, package‑size optimization, integration steps, performance metrics, and a hybrid strategy that boosts recommendation click‑through rates by 30%.
The article analyzes Google’s Made by Google 2025 event, showing how the Pixel 10 lineup, the Tensor G5 chip, Gemini Nano, and a full‑stack AI infrastructure—including custom TPUs, AI Hypercomputer, and Vertex AI—form a coordinated on‑device AI strategy that challenges Apple and builds a long‑term economic moat.
An in‑depth analysis of Apple Intelligence and the broader LLM ecosystem, covering recent model scaling breakthroughs, data and compute requirements, pricing dynamics, hardware trends, on‑device versus cloud deployment, and strategic implications for developers, product managers, and AI practitioners.
This article outlines the evolution from 1G to 6G communications, explains the third AI wave driven by big data, theory, and compute, introduces federated learning (horizontal, vertical, transfer), and details on‑device AI architectures, decision tree and neural network models, and real‑world use cases such as video preloading and autonomous driving.
Device intelligence brings on-device reasoning and real-time inference to smartphones and IoT gateways, delivering low-latency, privacy-preserving, personalized services such as AR/VR enhancements and recommendation re-ranking, while confronting challenges of hardware fragmentation and model size, and complementing cloud AI through architectures like Hala’s MNN-based pipeline.
LiteKit, Baidu’s mobile AI deployment framework built on Paddle Lite, delivers out‑of‑the‑box video super‑resolution, human segmentation and gesture‑recognition SDKs that reduce integration complexity to three simple steps across Objective‑C, Java and C++, achieving real‑time performance (25 FPS) while lowering development effort and platform barriers.
This article examines the limitations of cloud‑based recommendation pipelines, explains how edge computing can provide localized user perception and rapid re‑ranking, describes the EdgeRec on‑device model architecture—including heterogeneous behavior sequence modeling and behavior‑aware attention reranking—and presents offline and online experimental results that demonstrate significant gains in click‑through and conversion rates.
This article examines the evolution of smart terminals, outlines the sensor and computing trends driving new mobile experiences, and details Alibaba’s xMedia SDK—including its rich‑media foundation, on‑device deep‑learning engine (xNN), SLAM positioning (xSLAM), 3D rendering (xAnt3D), and cross‑platform capabilities—showcasing how these technologies enable more intelligent, decentralized user interactions.
