A language model that can generate utterances that are appraised as being within a specific age of a young child who is beginning their language learning journey can be useful in scenarios where child-level language is needed, for example in virtual avatars, interactions with individuals who have disabilities, or developmental robotics. In this paper, we focus on an age range that is not represented in prior work: emergent speakers. We use the CHILDES database to train and tune language models of different parameter sizes using a group relative policy optimization reinforcement learning regime. Our goal is to find the most coherent, yet child-like language model while keeping the number of parameters to as few as possible. We evaluate using metrics of coherency, “toddlerality,” and an evaluation using human subjects who interact with two robot platforms. Our experiments show that even small language models (under 1 billion parameters) can be used effectively to generate child-like utterances.

This demo will showcase updates made to the ‘robot-ready spoken dialogue system’ built on the Retico framework. Updates include new modules, logging and real-time monitoring tools, integrations with the Coppelia Sim virtual robot platfrom, integrations with a benchmark, improved documentation, and pypi environment usage.

My main research interests lies in the application of Reinforcement Learning (RL) alignment of LLMs in human robot dialogue. More specifically, my latest research aims to use RL alignment as an efficient training regime to train a newly initialized tiny LM to behave like a toddler. Previous research expresses the difficulty of building a robust tiny LM with an educated adult level understanding. Our hypothesis is that the cognitive barrier to train a tiny LM to at-least behave as a child is achievable with a very small number of parameters especially if training efficiently using RL LLM training regime. My interests also extend to apply RL to LLM training for dialogue management and planning.