Abstract: A computer-implemented method, a machine learning system, and non-transitory computer-readable storage medium for designing electrical circuits are provided. In the computer-implemented method input data is received and processed to generate a representation of the electrical circuit. A plurality of candidate routes for connecting a first and second circuit element of the electrical circuit are identified. A candidate route is then selected by iteratively selecting candidate sub-routes. Selecting candidate sub-routes is performed by using a Sequential Monte Carlo process to perform a look ahead search of a subset of the plurality of candidate routes, the Sequential Monte Carlo process being guided by a neural network. The representation of the electrical circuit is then updated with an action selection signal representing a selection of a candidate sub-route.

Abstract: A computer-implemented method, a system, and non-transitory computer-readable storage medium for designing electrical circuits are provided. In the method input data is received and processed to generate a representation of the electrical circuit. A first process is repeatedly performed to identify a plurality of candidate routes for connecting a first and second circuit element based on the representation. A candidate route is selected from the plurality of candidate routes based on a look ahead search. The first process includes selecting a first point in the representation, executing a second process to identify a set of candidate points, and selecting a second point from the set of candidate points. The second process comprises evaluating at least one candidate path extending in a linear direction from the first point to identify the set of candidate points based on at least a constraint and a topology of the electrical circuit.

Abstract: A computer-implemented method, a machine learning system, and non-transitory computer-readable storage medium for designing electrical circuits are provided. In the method input data, comprising an indication of a plurality of connections including a first and second connection is processed to generate a representation of the electrical circuit. Routes for the first and second connections are determined using an iterative process that includes defining one or more orders in which to determine routes for the first and second connections. A Sequential Monte Carlo process is used to perform a look ahead search of each defined order by generating simulations in respect of routes to be determined for the connections in the orders, the Sequential Monte Carlo process being guided by a neural network. A connection is selected and a route for the selected connection is determined. The representation is updated by providing an action selection signal representing the determined route.

Abstract: A computer-implemented method, a machine learning system, and non-transitory computer-readable storage medium for training a neural network are provided. The neural network is used to instruct an agent to select actions for interacting with an environment to determine a solution to a specified problem. In the computer-implemented method a state signal representing a current state of the environment is received. A Sequential Monte Carlo process is then used to perform a search to determine target action selection data associated with the current state of the environment. This target action selection data is stored in association with the state signal and the current state of the environment is updated by providing an action selection signal based on the target action selection data. The Sequential Monte Carlo process involves generating a plurality of simulations using the neural network to determine the target action selection data.

Abstract: A computer-implemented method, a machine learning system, and non-transitory computer-readable storage medium for designing electrical circuits are provided. In the computer-implemented method input data is received and processed to generate a representation of the electrical circuit. A plurality of candidate routes for connecting a first and second circuit element of the electrical circuit are identified. A candidate route is then selected by iteratively selecting candidate sub-routes. Selecting candidate sub-routes is performed by using a Sequential Monte Carlo process to perform a look ahead search of a subset of the plurality of candidate routes, the Sequential Monte Carlo process being guided by a neural network. The representation of the electrical circuit is then updated with an action selection signal representing a selection of a candidate sub-route.

Abstract: A computer-implemented method, a machine learning system, and non-transitory computer-readable storage medium for designing electrical circuits are provided. In the method input data, comprising an indication of a plurality of connections including a first and second connection is processed to generate a representation of the electrical circuit. Routes for the first and second connections are determined using an iterative process that includes defining one or more orders in which to determine routes for the first and second connections. A Sequential Monte Carlo process is used to perform a look ahead search of each defined order by generating simulations in respect of routes to be determined for the connections in the orders, the Sequential Monte Carlo process being guided by a neural network. A connection is selected and a route for the selected connection is determined. The representation is updated by providing an action selection signal representing the determined route.

Abstract: A computer-implemented method, a system, and non-transitory computer-readable storage medium for designing electrical circuits are provided. In the method input data is received and processed to generate a representation of the electrical circuit. A first process is repeatedly performed to identify a plurality of candidate routes for connecting a first and second circuit element based on the representation. A candidate route is selected from the plurality of candidate routes based on a look ahead search. The first process includes selecting a first point in the representation, executing a second process to identify a set of candidate points, and selecting a second point from the set of candidate points. The second process comprises evaluating at least one candidate path extending in a linear direction from the first point to identify the set of candidate points based on at least a constraint and a topology of the electrical circuit.