Technology Category
Machine Learning: Neuro-Symbolic LearningNatural Language Processing: Code Generation / Program Synthesis from Natural LanguageConstraint Satisfaction and Optimization: Satisfiability Modulo Theories
Synthesizing DSLs for Few-Shot Learning
🤖 AI Summary
This work addresses the automatic synthesis of domain-specific languages (DSLs) for symbolic few-shot learning: given a base language and a small set of input-output training/test pairs, the goal is to synthesize a syntactically constrained DSL such that concise expressions satisfying the training examples generalize provably to test examples. We formally define this problem for the first time and establish its decidability under a tree-automata–based semantic framework, showing that feasible DSLs correspond precisely to regular tree languages. Methodologically, we integrate formal verification with grammar induction, extend expressivity via macro grammars, and measure expression complexity by parse-tree depth. Experimentally, we realize automatic DSL synthesis for a class of symbolic languages, ensuring that correctness of small expressions on training data implies strict generalization to test data. Our approach further yields decidability results for several natural variants of the problem.
Technology Category
Application Category
📝 Abstract
We study the problem of synthesizing domain-specific languages (DSLs) for few-shot learning in symbolic domains. Given a base language and instances of few-shot learning problems, where each instance is split into training and testing samples, the DSL synthesis problem asks for a grammar over the base language that guarantees that small expressions solving training samples also solve corresponding testing samples. We prove that the problem is decidable for a class of languages whose semantics over fixed structures can be evaluated by tree automata and when expression size corresponds to parse tree depth in the grammar, and, furthermore, the grammars solving the problem correspond to a regular set of trees. We also prove decidability results for variants of the problem where DSLs are only required to express solutions for input learning problems and where DSLs are defined using macro grammars.
Problem
Research questions and friction points this paper is trying to address.
Synthesizing domain-specific languages for few-shot symbolic learning
Finding grammars ensuring training solutions generalize to testing
Proving decidability for automata-evaluated languages with size constraints
Innovation
Methods, ideas, or system contributions that make the work stand out.
Synthesizing domain-specific languages for few-shot learning
Grammar over base language ensuring training-to-testing generalization
Decidable for tree automata semantics with regular tree sets
🔎 Similar Papers
No similar papers found.
