Learning to Solve Complex Problems via Dataset Decomposition
This addresses the challenge of learning from complex data for AI systems, though it appears incremental as it builds on existing curriculum learning methods.
The paper tackles the problem of training models on complex datasets by proposing a reverse curriculum generation approach that recursively decomposes complex examples into simpler components, resulting in superior performance on math and code generation datasets compared to standard training.
Curriculum learning is a class of training strategies that organizes the data being exposed to a model by difficulty, gradually from simpler to more complex examples. This research explores a reverse curriculum generation approach that recursively decomposes complex datasets into simpler, more learnable components. We propose a teacher-student framework where the teacher is equipped with the ability to reason step-by-step, which is used to recursively generate easier versions of examples, enabling the student model to progressively master difficult tasks. We propose a novel scoring system to measure data difficulty based on its structural complexity and conceptual depth, allowing curriculum construction over decomposed data. Experiments on math datasets (MATH and AIME) and code generation datasets demonstrate that models trained with curricula generated by our approach exhibit superior performance compared to standard training on original datasets.