The Decade of AI Development: The Most Noteworthy Moments of the 2010s
The 2010s will be known for the evolution of one of the most advanced technologies in the world — artificial intelligence. Over the years, as more resources are made available for AI development, and it is made more accepted by consumers and companies alike, it is worth revisiting some of the key milestones over the last decade that have made this development possible.
So, let us take some time to go through an overlook of important events that have happened in the last decade in the field of AI development.
The decade of AI development
2010
· ImageNet competition — the ImageNet Large Scale Visual Recognition Challenge (ILSCRV) is used to evaluate the algorithms for object detection and image classification at a very large scale. This was the benchmark used to evaluate the performance of the classification model.
· Apple acquired Siri — Siri was one of the first prototypes of the virtual assistants that we have today. Apple had actively involved in its AI development to make the AI assistant user-friendly.
· Microsoft launches Kinect for Xbox 360 — it is the first gaming device that tracked human body movement with the help of a 3D camera and infrared detection.
2011
· IBM Watson wins Jeopardy — IBM Watson was a natural language question-answering computer and it competed on Jeopardy! consequently defeating two former champions.
· Growth of CNN — a convolutional neural network (CNN) won the German Traffic Sign Recognition competition with 99.46% accuracy.
2012
· Cat vs Dog — the researchers in the AI development Google X lab trained a neural network of 16,000 processors to identify the images of cats by presenting 10 million unlabelled images from YouTube videos.
· CNN at ImageNet Competition — researchers at the University of Toronto designed a convolutional neural network that showcased an error rate of only 16% in the ImageNet Large Scale Visual Recognition Challenge.
2013
· The Never Ending Image Learner — NEIL is a computer program that functions 24/7 learning about images that it finds on the internet. The objective of the software development program is to learn common sense relationships found in everyday life.
2014
· Alexa was born — the well-recognized AI voice, that was invented by Rohit Prasad, is a small device that turns on when its name is called out. Once on, it will record and store everything you say on the cloud — to learn and respond.
· Tesla autopilot — Tesla Motors announced their first version of AutoPilot. The cars equipped with this AI development system were capable of lane control with autonomous steering, braking, and speed limit adjustment based on signals image recognition.
2015
· Tensorflow Release — Google made its deep learning framework Tensorflow open-source. It is an important milestone in AI software development as it gave everyone the tool to build extraordinary models. It is now an end-to-end open-source platform for machine learning.
· Facenet paper — the technology that initiated face recognition, Facenet was released by Google. It marked the beginning of face recognition to a large set of users.
· YOLO paper — You Only Look Once (YOLO) was a new approach to object detection. It is a single neural network that predicts class probabilities and bounding boxes directly from complete images in just one evaluation.
2016
· Release of TPU — a Tensor Processing Unit is an AI development accelerator application-specific integrated circuit (ASIC) that was developed by Google. It was exclusively for neural network machine learning.
· Birth of Sophia — Sophia is a social humanoid robot that was developed by the Hong Kong-based company Hanson Robotics. Sophia can imitate facial expressions, and human gestures, answer certain questions and make simple conversations on predefined topics.
· PyTorch — it is an open-source ML library that is based on the Torch library. It is used for the software development of applications such as computer vision and NLP. It was primarily developed by Facebook’s AI research lab (FAIR).
2017
· AlphaZero — AlphaZero was launched by the company DeepMind to master the games of chess, go, and shogi. In just 24 hours of its release, it reached a superhuman level of play, defeating the world champion Stockfish. Without any human guidance except the basic game rules, AlphaZero learned how to play chess by playing against itself in just 4 hours.
· ONNX — Microsoft and Facebook joined together to enable AI framework interoperability. Along with the help of partner communities, the tech giants developed the Open Neural Network Exchange (ONNX). It is an open format AI development technology that represents deep learning models that can be trained in one framework and transferred to another for inference.
2018
· Waymo covered 10 million miles — Waymo’s self-driving vehicles cross over 10 million miles on public roads and almost 7 billion in simulation.
· Deepfake technology — deepfakes started combining and superimposing existing media onto source media with the help of machine learning technology known as autoencoders and generative adversarial networks (GANs).
· AlphaFold — AlphaFold relies on years of research in utilizing vast genomic data to predict protein structure.
· BERT — Google successfully develops the first Bidirectional Unsupervised Language Representation (BERT) that can be used on a range of natural language software development tasks using transfer learning.
2019
· Deepfakes detection challenge — this challenge invited people around the world to create new technologies that can help detect deepfakes and manipulated media.
· AI development in medicine — Google researchers worked with Northwestern Medicine to build an AI system that detects lung cancer more accurately than humans.
· Smart homes — Google, Amazon, and Apple teamed up to develop an open-source smart home standard that ensures that devices work together, keeps everything secure, and makes the AI development of new devices easier.
2020
· LinearFold — Baidu’s AI-based LinearFold algorithm helped scientific and medical teams to fight the COVID-19 outbreak. It is faster than the traditional RNA folding algorithms at predicting a virus’s secondary RNA structure.
· AI development also took a big part in speeding up the development of vaccines (that would usually take decades), by helping researchers in analyzing vast amounts of data about coronavirus.
Future of AI development
AI development does not seem to slow down anytime. With an increase in data, research, and computation, AI is now evolving at a rapid pace.
One thing is for sure, AI is here to stay. So, if you want to stay ahead in the digital race, make sure that you adopt the technology at the earliest in your software development processes.
