Writing computer programs is a skill that remains inaccessible to most due to the barrier of programming language (PL) syntax. While large language models (LLMs) have been proposed to translate natural language pseudocode to PL code, they are costly in terms of data and compute. We propose a lightweight alternative to LLMs that exploits the property of code wherein most tokens can be simply copied from the pseudocode. We divide the problem into three phases: Copy, Generate, and Combine. In the Copy Phase, a binary classifier is employed to determine and mask the pseudocode tokens that can be directly copied into the code. In the Generate Phase, a Sequence-to-Sequence model is used to generate the masked PL code equivalent. In the Combine Phase, the generated sequence is combined with the tokens that the Copy Phase had masked. We show that our C3PO models achieve similar performance to non-C3PO models while reducing the computational cost of training as well as the vocabulary sizes.

Cross-lingual summarization(CLS) is the process of generating a summary in one particular language for a source document in a different language. Low resource languages like Kannada greatly benefit from such systems because they help in delivering a concise representation of the same information in a different popular language. We propose a novel dataset generation pipeline and a first of its kind dataset that will aid in CLS for Kannada-English language pair. This work is also an attempt to inspect the existing systems and extend them to the Kannada-English language pair using our dataset.