We presented a framework for evolutionary robotics in both simulation world and the real world by integrating Gym, Gazebo, and ROS. The interface provided by Gym allows users and researchers to focus on the learners, instead of wasting time on the s…

With Contextual BERT, we presented novel ways of conditioning the BERT model. The strong performance on a real-world use-case provides evidence for the superiority of using a global state to inject context into the Transformer-based architecture. Ou…

High and low frequency components present in an image play a vital role when they are processed by deep learning models. Several recent research works have also highlighted that CNN models are highly sensitive towards high and low-frequency componen…

Converting sequential data into a format which implicitly encodes temporal information as spatial information is an effective way of generating samples of such data as whole pieces. Here, we explored this approach for melody generation of fixed-leng…

In this paper, we propose a simple single-loop algorithm for nonconvex min-max problems (1.1). For an important family of problems (1.2), the algorithm is even more efficient due to the dual error bound, and it is well-suited for problems in large-s…