Random sampling maps (with-replacement, binning, species) induce metrics that give uniform any-dimensional generalization and sketching rates for continuous functions on sequences, graphs and tensors.
Length Generalization Bounds for Transformers
1 Pith paper cite this work. Polarity classification is still indexing.
abstract
Length generalization is a key property of a learning algorithm that enables it to make correct predictions on inputs of any length, given finite training data. To provide such a guarantee, one needs to be able to compute a length generalization bound, beyond which the model is guaranteed to generalize. This paper concerns the open problem of the computability of such generalization bounds for C-RASP, a class of languages which is closely linked to transformers. A positive partial result was recently shown by Chen et al. for C-RASP with only one layer and, under some restrictions, also with two layers. We provide complete answers to the above open problem. Our main result is the non-existence of computable length generalization bounds for C-RASP (already with two layers) and hence for transformers. To complement this, we provide a computable bound for the positive fragment of C-RASP, which we show equivalent to fixed-precision transformers. For both positive C-RASP and fixed-precision transformers, we show that the length complexity is exponential, and prove optimality of the bounds.
fields
math.ST 1years
2026 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Any-Dimensional Learning by Sampling
Random sampling maps (with-replacement, binning, species) induce metrics that give uniform any-dimensional generalization and sketching rates for continuous functions on sequences, graphs and tensors.