TY - RPRT TI - Human-like compositional learning of visually-grounded concepts using synthetic environments AU - Zijun Lin AU - M Ganesh Kumar AU - Cheston Tan PY - 2025 UR - https://arxiv.org/abs/2504.06618 ID - 2504.06618 ER -