ScienceWorld: Is your Agent Smarter than a 5th Grader?
This addresses the challenge of assessing true reasoning versus memorization in AI for scientific education and research, though it is incremental as it builds on existing interactive environment benchmarks.
The authors tackled the problem of evaluating scientific reasoning in AI agents by introducing ScienceWorld, a benchmark based on elementary school science, and found that large models trained statically struggle with novel contexts, while a smaller agent trained interactively outperforms them, with the interactive agent (1.5M parameters) beating an 11B parameter model.
We present ScienceWorld, a benchmark to test agents' scientific reasoning abilities in a new interactive text environment at the level of a standard elementary school science curriculum. Despite the transformer-based progress seen in question-answering and scientific text processing, we find that current models cannot reason about or explain learned science concepts in novel contexts. For instance, models can easily answer what the conductivity of a known material is but struggle when asked how they would conduct an experiment in a grounded environment to find the conductivity of an unknown material. This begs the question of whether current models are simply retrieving answers by way of seeing a large number of similar examples or if they have learned to reason about concepts in a reusable manner. We hypothesize that agents need to be grounded in interactive environments to achieve such reasoning capabilities. Our experiments provide empirical evidence supporting this hypothesis -- showing that a 1.5 million parameter agent trained interactively for 100k steps outperforms a 11 billion parameter model statically trained for scientific question-answering and reasoning from millions of expert demonstrations.