artificial intelligence 2020, #tfworld, 32-bit computing, abstraction, accelerometer, accessibility, ai, android (operating system), app debugging, app developers, application programming interface, application software, arduino, arm architecture, arm cortex-m, artificial intelligence, artificial neural network, automation, binary multiplier, bluetooth, booting, c (programming language), c++, camera trap, census, central processing unit, christmas lights, cloud computing, cloud edge tpu, coffee, compiler, computer, computer data storage, computer file, computer graphics, computer network, computer programming, computer science 2020, computer science 2021, computer science conference 2021, computer science event, computer science event 2021, computer vision, computing, concept, coral board, coral dev board, cuda, data center, data type, decision-making, deep learning, developing country, digital signal processor, double-precision floating-point format, drinking water, edge tpu, edge tpu demo, electrical cable, electrical grid, electricity, embedded platforms, embedded system, energy harvesting, engine, engineering, expert, fairy, field-programmable gate array, file format, flash memory, gesture, github, googel coral board, google, google ai, google assistant, google edge tpu, gpu, graphical user interface, graphics processing unit, guitar, hardware acceleration, hope, information technology 2020, infrastructure, instruction pipelining, instruction set architecture, integrated circuit, intel, internet, internet of things, interpreter (computing), ios, iridium satellite constellation, jaw, json, k-nearest neighbors algorithm, kernel (operating system), language, laptop, learning, library (computing), light-emitting diode, linux, machine learning, machine learning 2021, machine learning android, machine learning developers, machine learning for apps, machine learning for embedded devices, machine learning for mobile apps, machine learning ios, machine learning mobile, machine learning technologies , machine learning tools, make (software), mathematical optimization, matthew du puy, menu, metadata, microcontroller, mind, ml, ml for mobile, ml framework, ml kit, ml tutorials, mobile app ml, mobile devices, mobile machine learning, neural network, nintendo entertainment system, noise, numerical control, nupur garg, nvidia, o'reilly, o'reilly tensorflow world, opencl, operating system, optical character recognition, past, pete warden, pleasure, privacy, program optimization, protein, prototype, radio, random-access memory, raspberry pi, read-only memory, real-time computing, reason, research, santa clara convention center, serial port, server (computing), signal, simulation, siri, smartphone, software , software framework, solar panel, solution for running ml, solution for running ml on device, solution for running ml on-device, sparkfun, sparse matrix, species, speech recognition, spotify, stack (abstract data type), string (computer science), system call, system on a chip, technology, tensor processing unit, tensorflow, tensorflow 2.0, tensorflow android, tensorflow apis, tensorflow applications, tensorflow dev summit, tensorflow dev summit 2019, tensorflow dev summit 2020, tensorflow dev summit keynote, tensorflow developer summit, tensorflow developer summit 2019, tensorflow for apps, tensorflow for embedded devices, tensorflow for mobile, tensorflow in mobile, tensorflow ios, tensorflow lite, tensorflow lite demo, tensorflow world, tensorflow world 19, tensorflow world 2019, tf, tf dev summit 20, tf dev summit 2019, tf dev summit 2020, tf lite, tf lite 1.0, tf lite micro, tf world, tf world 19, tf world 2019, tfds, tflite, tfworld, tfworld 19, tfworld 2019, tinyml, tinyml conference 2020, tinyml foundation, tinyml talks 2021, tinyml2020, tinyml2021 panel discussion, tpu, unit testing, usability, vector processor, verb, vicenza, voicemail, wand, wearable computer, wi-fi, world wide web