Table 3: Comparison of models trained with different representations of protein structure across various tasks, on a random data split . The optimal choice of representation depends on the task. Shown are mean and standard deviation across four runs with different seeds. Table 4: Comparison of models trained with different representations of protein structure across various tasks, on a sequence data split . Table 5: Comparison of models trained with different representations of protein structure across various tasks, on a structure data split .

However, the heterogeneity of protein data constitutes a problem for reproducible machine learning.

We release all the benchmark, making it more efficient and friendly to researchers.

In the context of reinforcement learning from human feedback (RLHF), the reward function is generally derived from maximum likelihood estimation of a random utility model based on pairwise comparisons made by humans. The problem of learning a reward function is one of preference aggregation that, we argue, largely falls within the scope of social choice theory. From this perspective, we can evaluate different aggregation methods via established axioms, examining whether these methods meet or fail well-known standards. We demonstrate that both the Bradley-Terry-Luce Model and its broad generalizations fail to meet basic axioms. In response, we develop novel rules for learning reward functions with strong axiomatic guarantees. A key innovation from the standpoint of social choice is that our problem has a linear structure, which greatly restricts the space of feasible rules and leads to a new paradigm that we call linear social choice .