Search papers, labs, and topics across Lattice.
This paper investigates the limits of language generation with bounded memory, contrasting it with traditional learning theory assumptions of full history access. It demonstrates that memoryless generators can handle any countable collection of infinite languages under a mild enumeration restriction, and provides an exact characterization of memoryless generation without this restriction. The paper also derives a combinatorial bound for the optimal minimax density achievable by memoryless generators for finite language collections, showing that a sliding window does not improve worst-case density, while adaptive memory does.
Even without remembering everything, you can still generate infinite languages, but hitting density targets or identifying the right language becomes much harder.
We study language generation in the limit under bounded memory. In this task, a learner observes examples from an unknown target language one at a time and must eventually output only new valid examples. Prior work assumes access to the entire history, a strong assumption since realistic algorithms retain limited past information. Classical work in learning theory shows memory constraints dramatically alter learnability; we extend this to language generation. First, we study memoryless generators. Under a mild enumeration restriction, every countable collection of infinite languages remains generable without memory. Without this restriction, we exactly characterize when memoryless generation is possible. For finite collections, we characterize the optimal minimax density achievable by memoryless generators -- the best density guaranteed against any collection of a given size. This combinatorial bound relies on Sperner's theorem and symmetric chain decompositions. We further show that a sliding window of the last $W$ examples does not improve this worst-case density, whereas allowing it to store $b$ adaptively chosen past examples improves the achievable density for every $b \geq 1$. Finally, we revisit identification in the limit, where the learner must converge to a single correct hypothesis for the target language. We focus on its incremental variant, where the learner remembers only its previous guess. Here, although exact identification fails on a collection of just three languages, a mild relaxation requiring convergence to an ``approximate''version of the target is achievable for every finite collection. These results show bounded memory affects these tasks differently: generation remains achievable for every countable collection, while density and identification are confined to finite collections, with guarantees weakening as the collection grows.
