Abstract: Techniques for using transfer learning to address label data shortage in seniority modeling for an online service are disclosed herein. In some embodiments, a computer-implemented method comprises training an initialized neural network using training examples comprising profile data and labels for the profile data, where each label comprises a standardized position title, and the training of the initialized neural network forms a pre-trained neural network. Next, the computer system may train the pre-trained neural network using training examples comprising profile data and labels for the profile data, where the labels comprise a position seniority, and the training of the pre-trained neural network forms a fine-tuned neural network. The computer system may then compute the position seniority for a user based on profile data of the user using the fine-tuned neural network, and use the position seniority of the user in an application of an online service.

Abstract: A method and system for generating neural architectures to perform a particular task. An actor neural network, as part of a continuous action reinforcement learning (RL) agent, generates a randomized continuous actions parameters to encourage exploration of a search space to generate candidate architectures without bias. The continuous action parameters are discretized and applied to a search space to generate candidate architectures, the performance of which for performing the particular task is evaluated. Corresponding reward and state are determined based on the performance. A critic neural network, as part of the continuous action RL agent, learns a mapping of the continuous action to a reward using modified Deep Deterministic Policy Gradient (DDPG) with quantile loss function by sampling a list of top performing architectures. The actor neural network is updated with the learned mapping.

Abstract: A method and system for neural architectural search (NAS) for performing a task. A generative adversarial network comprising a generator and a discriminator receives, from a user device, a query for neural network architecture, the query including a search space. The generator of the generative adversarial network generates a plurality of generated neural network architectures responsive to the received search space. The discriminator of the generative adversarial network selects an optimal neural network architecture from among the plurality of generated neural network architectures. The optimal generated neural network architecture is transmitted to the user device.

Abstract: A method and system for generating neural architectures to perform a particular task. An actor neural network, as part of a continuous action reinforcement learning (RL) agent, generates a randomized continuous actions parameters to encourage exploration of a search space to generate candidate architectures without bias. The continuous action parameters are discretized and applied to a search space to generate candidate architectures, the performance of which for performing the particular task is evaluated. Corresponding reward and state are determined based on the performance. A critic neural network, as part of the continuous action RL agent, learns a mapping of the continuous action to a reward using modified Deep Deterministic Policy Gradient (DDPG) with quantile loss function by sampling a list of top performing architectures. The actor neural network is updated with the learned mapping.