Google's Minerva Will Solve a Third of University-Level Problems in Mathematics, Physics, Chemistry, Economics, and Biology!

Google's Minerva Will Solve a Third of University-Level Problems in Mathematics, Physics, Chemistry, Economics, and Biology!

We have another innovation for fans of language models. The Minerva model is based on Google's Pathways language model, which boasts 540 billion parameters. It was initially trained on general language data and then fine-tuned on technical content. It appears to be the best in its class in its current focus. It tackles mathematical and scientific questions using a "step-by-step" reasoning approach. The model was trained on a 118GB dataset of scientific papers from arXiv and websites containing mathematical expressions using LaTeX, MathJax, or other mathematical formats.

Sources: [1]: https://ai.googleblog.com/2022/06/minerva-solving-quantitative-reasoning.html
[2]: Research paper on Minerva: https://arxiv.org/abs/2206.14858
[3]: Chain of Thought explanation from Google: https://ai.googleblog.com/2022/05/language-models-perform-reasoning-via.html
[4]: Research on Chain of Thought: https://arxiv.org/abs/2201.11903
[5]: Meier, RD, Cheney, DL, and Seyfarth, RM (1990). Do Animals Remember the Past and Plan for the Future? Behavioral and Brain Sciences, 13(1), 1–16.
[6]: https://pub.towardsai.net/how-minerva-a-new-nlp-model-performs-better-than-gpt-3-f7713458b59b

Originally published on Facebook — link to post