Tag
The paper introduces Lie–Taylor jet matching, a solver-free framework for training Universal Differential Equations on microcontrollers, achieving accuracy comparable to traditional methods with minimal memory usage.
The author embedded the NAM A2-Lite AI model into an iRig HD X guitar interface by replacing its firmware, enabling real-time amp modelling on the device's microcontroller without external computers.
A comprehensive book by David Such that teaches how to integrate AI into embedded systems, featuring over 25 hands-on projects with practical engineering guidance.
AMD details its new Ryzen AI Embedded X100 Series processors and Kria AI platforms, aiming to lead in the physical AI market with high-performance compute and unified AI acceleration for edge systems.
Cactus releases Needle 2, a 14MB agentic LLM for phones, wearables, smart home devices, and robots, achieving fast inference on low-end hardware and supporting structured extraction and fine-tuning.
A developer successfully ran a 5.2 million parameter MoE LLM quantized to INT4 on an ESP32 Dev Kit V1 using only 81KB of SRAM by streaming experts from flash, achieving about 5 tokens per second.
bootai is an open-source UEFI application that boots directly into an AI chat/code REPL, running Qwen2.5 and SmolLM model inference on bare metal without an OS, with hand-written network drivers and a TCP/IP stack.
This paper presents a lightweight 1D CNN framework for classifying affective touches in soft plush companions, achieving 75% test accuracy with a 13.2k-parameter model and proposing a hybrid heuristic-CNN inference pipeline for embedded deployment.
The developer spent 14 months creating an AI physical prototype device named Keito, based on the ESP32 chip. It supports features such as voice conversation, real-time lip-sync animation, capacitive touch interaction, music playback, and weather query, aiming to liberate AI from the text box.