Is Programming by Example solved by LLMs?

Programming-by-Examples (PBE) aims to generate an algorithm from input-output examples. Such systems are practically and theoretically important: from an end-user perspective, they are deployed to millions of people, and from an AI perspective, PBE corresponds to a very general form of few-shot inductive inference. Given the success of Large Language Models (LLMs) in code-generation tasks, we investigate here the extent to which LLMs can be said to have `solved' PBE. We experiment on classic domains such as lists and strings, and an uncommon graphics programming domain not well represented in typical pretraining data. We find that pretrained models are not effective at PBE, but that they can be fine-tuned for much higher performance, provided the test problems are in-distribution. We analyze empirically what causes these models to succeed and fail, and take steps toward understanding how to achieve better out-of-distribution generalization. Collectively these results suggest that LLMs make strong progress toward solving the typical suite of PBE tasks, potentially increasing the flexibility and applicability of PBE systems, while also identifying ways in which LLMs still fall short.

Code Repair with LLMs gives an Exploration-Exploitation Tradeoff

Iteratively improving and repairing source code with large language models (LLMs), known as refinement, has emerged as a popular way of generating programs that would be too complex to construct in one shot. Given a bank of test cases, together with a candidate program, an LLM can improve that program by being prompted with failed test cases. But it remains an open question how to best iteratively refine code, with prior work employing simple greedy or breadth-first strategies. We show here that refinement exposes an explore-exploit tradeoff: exploit by refining the program that passes the most test cases, or explore by refining a lesser considered program. We frame this as an arm-acquiring bandit problem, which we solve with Thompson Sampling. The resulting LLM-based program synthesis algorithm is broadly applicable: Across loop invariant synthesis, visual reasoning puzzles, and competition programming problems, we find that our new method can solve more problems using fewer language model calls.

WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environment

We give a model-based agent that builds a Python program representing its knowledge of the world based on its interactions with the environment. The world model tries to explain its interactions, while also being optimistic about what reward it can achieve. We do this by extending work on program synthesis via LLMs. We study our agent on gridworlds, finding our approach is more sample-efficient compared to deep RL, and more compute-efficient compared to ReAct-style agents.

Doing Experiments and Revising Rules with Natural Language and Probabilistic Reasoning

We build a computational model of how humans actively infer hidden rules by doing experiments. The basic principles behind the model is that, even if the rule is deterministic, the learner considers a broader space of fuzzy probabilistic rules, which it represents in natural language, and updates its hypotheses online after each experiment according to approximately Bayesian principles. In the same framework we also model experiment design according to information-theoretic criteria. We find that the combination of these three principles -- explicit hypotheses, probabilistic rules, and online updates -- can explain human performance on a Zendo-style task, and that removing any of these components leaves the model unable to account for the data.

Active Preference Inference using Language Models and Probabilistic Reasoning

Actively inferring user preferences, for example by asking good questions, is important for any human-facing decision-making system. Active inference allows such systems to adapt and personalize themselves to nuanced individual preferences. To enable this ability for instruction-tuned large language models (LLMs), one may prompt them to ask users questions to infer their preferences, transforming the language models into more robust, interactive systems. However, out of the box, these models are not efficient at extracting preferences: the questions they generate are not informative, requiring a high number of user interactions and impeding the usability of the downstream system. In this work, we introduce an inference-time algorithm that helps LLMs quickly infer preferences by using more informative questions. Our algorithm uses a probabilistic model whose conditional distributions are defined by prompting an LLM, and returns questions that optimize expected entropy and expected model change. Results in a simplified interactive web shopping setting with real product items show that an LLM equipped with our entropy reduction algorithm outperforms baselines with the same underlying LLM on task performance while using fewer user interactions.

Rapid Motor Adaptation for Robotic Manipulator Arms

Developing generalizable manipulation skills is a core challenge in embodied AI. This includes generalization across diverse task configurations, encompassing variations in object shape, density, friction coefficient, and external disturbances such as forces applied to the robot. Rapid Motor Adaptation (RMA) offers a promising solution to this challenge. It posits that essential hidden variables influencing an agent's task performance, such as object mass and shape, can be effectively inferred from the agent's action and proprioceptive history. Drawing inspiration from RMA in locomotion and in-hand rotation, we use depth perception to develop agents tailored for rapid motor adaptation in a variety of manipulation tasks. We evaluated our agents on four challenging tasks from the Maniskill2 benchmark, namely pick-and-place operations with hundreds of objects from the YCB and EGAD datasets, peg insertion with precise position and orientation, and operating a variety of faucets and handles, with customized environment variations. Empirical results demonstrate that our agents surpass state-of-the-art methods like automatic domain randomization and vision-based policies, obtaining better generalization performance and sample efficiency.

Modeling Human-like Concept Learning with Bayesian Inference over Natural Language

We model learning of abstract symbolic concepts by performing Bayesian inference over utterances in natural language. For efficient inference, we use a large language model as a proposal distribution. We fit a prior to human data to better model human learners, and evaluate on both generative and logical concepts.

LambdaBeam: Neural Program Search with Higher-Order Functions and Lambdas

Search is an important technique in program synthesis that allows for adaptive strategies such as focusing on particular search directions based on execution results. Several prior works have demonstrated that neural models are effective at guiding program synthesis searches. However, a common drawback of those approaches is the inability to handle iterative loops, higher-order functions, or lambda functions, thus limiting prior neural searches from synthesizing longer and more general programs. We address this gap by designing a search algorithm called LambdaBeam that can construct arbitrary lambda functions that compose operations within a given DSL. We create semantic vector representations of the execution behavior of the lambda functions and train a neural policy network to choose which lambdas to construct during search, and pass them as arguments to higher-order functions to perform looping computations. Our experiments show that LambdaBeam outperforms neural, symbolic, and LLM-based techniques in an integer list manipulation domain.

Top-Down Synthesis for Library Learning

This paper introduces corpus-guided top-down synthesis as a mechanism for synthesizing library functions that capture common functionality from a corpus of programs in a domain specific language (DSL). The algorithm builds abstractions directly from initial DSL primitives, using syntactic pattern matching of intermediate abstractions to intelligently prune the search space and guide the algorithm towards abstractions that maximally capture shared structures in the corpus. We present an implementation of the approach in a tool called Stitch and evaluate it against the state-of-the-art deductive library learning algorithm from DreamCoder. Our evaluation shows that Stitch is 3-4 orders of magnitude faster and uses 2 orders of magnitude less memory while maintaining comparable or better library quality (as measured by compressivity). We also demonstrate Stitch's scalability on corpora containing hundreds of complex programs that are intractable with prior deductive approaches and show empirically that it is robust to terminating the search procedure early -- further allowing it to scale to challenging datasets by means of early stopping.

I Speak, You Verify: Toward Trustworthy Neural Program Synthesis

We develop an approach for improving the trustworthiness and overall accuracy of program synthesizers based on large language models for source code. Given a natural language description of a programming problem, our method samples both candidate programs as well as candidate predicates specifying how the program should behave. We learn to analyze the agreement between programs and predicates to judge both which program is most likely to be correct, and also judge whether the language model is able to solve the programming problem in the first place. This latter capacity allows favoring high precision over broad recall: fostering trust by only proposing a program when the system is certain that it is correct.