One primary driver for artificial intelligence research in mathematical reasoning is that it may further increase model understanding and problem-solving abilities on complex mathematical problems.

Math-OCR-Zero is based on the VeRL framework and leverages the synthetic OCR-Math dataset to enhance the reasoning capabilities of multimodal large models. It serves as a reproduction of ...

Language models have made significant strides in mathematical reasoning, with synthetic data playing a crucial role in their development. However, the field faces significant challenges due to the ...

The training steps of the model include 2 stages: Continue pretraining: Using text corpus collected from external data about elementary school mathematics knowledge, some types of exercises to help ...

Abstract: We present a multi-way parallel corpus of Math Word Problems (MWPs) in nine languages, including six low-resource languages. To date, this is the largest multilingual MWP dataset available.

Mathematics is the foundation of countless sciences, allowing us to model things like planetary orbits, atomic motion, signal frequencies, protein folding, and more. Moreover, it’s a valuable testbed ...

Mathematics has always posed a significant challenge for AI models. Mastering math requires complex reasoning skills, and for AI, this task is anything but straightforward. That creates a huge problem ...