Paul Werbos

American engineer and social scientist whose 1974 Harvard doctoral thesis, 'Beyond Regression', showed how to compute the gradients of a multi-stage model efficiently by working backward from its output — the procedure now called backpropagationbackpropagation, n.: The method neural networks use to learn from mistakes: it works backward from the error to figure out how much each internal connection contributed, so every connection can be adjusted in the right direction. — and proposed using it to fit models including neural networks. He applied it explicitly to multilayer neural networksmultilayer perceptron, n.: A neural network made of several layers of artificial neurons, each connected to every neuron in the next layer. It is the simplest kind of deep network and a component of almost every modern model. in 1982, before its popularization by David RumelhartDavid Rumelhart, prop. n.: An American psychologist who, with Geoffrey Hinton and Ronald Williams, showed in 1986 how to train multilayer neural networks, launching the modern revival of the field., Geoffrey HintonGeoffrey Hinton, prop. n.: A British-Canadian scientist whose decades of persistent work on brain-inspired neural networks laid the foundations of modern AI; often called a 'godfather of AI', he now warns publicly about its dangers. and Ronald Williams in 1986, and popularized the name and method of backpropagation through timebackpropagation through time, n.: The way recurrent networks are trained: the network is unrolled into a long chain, one copy per time step, and errors are traced back through every step to adjust the shared weights. (1990) for recurrent systems. He also developed adaptive critic designs, which combine neural networks with approximate dynamic programmingdynamic programming, n.: A problem-solving method that breaks a big problem into smaller overlapping pieces, solves each piece once and reuses the answers — the basis for computing the best long-term strategy step by step. for control and are an early form of actor–criticactor–critic, n.: A way of training an AI agent in which one part, the 'actor', makes the moves while another, the 'critic', learns to predict how well they will work out, and the actor improves by heeding the critic. reinforcement learningreinforcement learning, n.: A way of teaching a computer program by trial and error: it tries actions, earns points for good results and penalties for bad ones, and gradually learns which choices pay off over time.. As a long-serving program director at the US National Science Foundation he funded research on neural networks and neuroengineering. He received the IEEE Neural Network Pioneer Award and served as president of the International Neural Network Society.