Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Learning to Extrapolate to New Tasks: A Relational Approach to Task Extrapolation

About

Modern learning systems excel at interpolation but struggle to generalize to unseen tasks outside the training distribution's support. This failure occurs even in simple settings, such as handling task parameters beyond the training range, and persists despite advances in foundation models. To this end, we develop the Relational Task Extrapolator (RTE), an algorithm designed to enable systematic extrapolation to novel tasks. The key observation is that extrapolation is inherently relational: extrapolating to unseen tasks requires learning how tasks transform into one another. If a model learns the transformation between tasks A and B during training, it can apply that same transformation to relate known tasks to unseen ones at test time. RTE operationalizes this idea by decomposing each target task into a known anchor task and a transformation linking the anchor and target. It then learns a relational operator, mapping an anchor-transformation pair to predictions for the target task. We instantiate RTE across multiple task extrapolation regimes in function prediction, e.g. where target tasks use out-of-range parameters (parameter extrapolation), have greater compositional depth (length extrapolation), and/or recombine function primitives in unseen ways (compositional extrapolation). We further extend RTE to sequence prediction, integrating it into fine-tuning algorithms for foundation models. Across empirical studies, we find that RTE substantially outperforms existing approaches on extrapolation to novel, unseen tasks.

Adam Ousherovitch, Yixin Wang• 2026

Related benchmarks

TaskDatasetResultRank
Parameter ExtrapolationSin Trend F2
MSE0.051
4
Parameter ExtrapolationTri Trend F2
MSE0.048
4
Parameter ExtrapolationExp F2
MSE0.8
4
Compositional String TransformationsCodeIO (200 unseen compositions)
Accuracy45.3
4
Parameter ExtrapolationCubic F2
MSE1.53
4
Parameter ExtrapolationQuadratic F2
MSE7.33
4
Composition ExtrapolationUnseen combinations of known primitives
Aggregate MSE0.2867
3
Length ExtrapolationSynthetic Degree-9 Polynomials Length Extrapolation (F2)
MSE0.3712
3
Sparse Parity ExtrapolationSparse Parity subset size 6 (test)
Accuracy66.07
3
Showing 9 of 9 rows

Other info

Follow for update