Sketch-based program synthesis allows users to guide the synthesizer by writing partial programs (sketches). Traditionally, the specifications of these sketches are safety properties expressed either as assertions or semantic equivalences. These specifications, however, lack expressiveness when the user wants to establish how the execution of the desired program evolves over time. This is especially important when synthesizing reactive programs, where the whole specification is about how the computation evolves over time. We explore an alternative method letting the user specify desired program executions as linear temporal logic (LTL) formulae. We define a method for transforming sketches with LTL assertions into sketches with only standard assertions. Specifically, for terminating programs, our method transforms and implements, within the sketch, the LTL formulae as runtime monitors. For non-terminating programs, our procedure defines and implements, into the sketch, fairness conditions (analyzing when an acceptance state of the Büchi automata equivalent to these LTL formulae occurs infinitely often). We prove the correctness of both constructions. Evaluation of our implementation in Sketch shows that our method enables the system to synthesize non-terminating programs, such as a round-robin arbiter for a variable number of devices and a lift controller for a variable number of floors. For terminating programs, our approach improves the synthesizer performance by restricting the search of program candidates without modifying the sketches structures.

Large Language Models (LLMs) applied to code-related applications have emerged as a prominent field, attracting significant interest from academia and industry. However, as new and improved LLMs are developed, existing evaluation benchmarks (e.g., HumanEvla, MBPP) are no longer sufficient for assessing their capabilities suffering from data contamination, overfitting, saturation, and focus on merely code generation. In this work, we propose LiveCodeBench, a comprehensive and contamination-free evaluation of LLMs for code, which collects new problems over time from contests across three competition platforms, Leetcode, Atcoder, and Codeforces. Notably, our benchmark also focuses on a broader range of code-related capabilities, such as self-repair, code execution, and test output prediction, beyond just code generation. Currently, LiveCodeBench hosts over six hundred coding problems that were published between May 2023 and Aug 2024. We evaluate over 50 LLMs on LiveCodeBench (LCB for brevity) presenting the largest evaluation study of code LLMs on competition problems. Based on the study, we present novel empirical findings on contamination, overfitting, and holistic evaluations. We demonstrate that time-segmented evaluations serve as a robust approach to evade contamination; they are successful at detecting contamination across a wide range of open and closed models including GPT-4O, Claude, Deepseek, and Codestral. Next, we highlight overfitting and saturation of traditional coding benchmarks like HumanEvla and demonstrate LCB allows more reliable evaluations. Finally, our holistic evaluation scenarios allow for measuring the different capabilities of programming agents in isolation.

Procedural materials, represented as functional node graphs, are ubiquitous in computer graphics for photorealistic material appearance design. They allow users to perform intuitive and precise editing to achieve desired visual appearances. However, creating a procedural material given an input image requires professional knowledge and significant effort. In this work, we leverage the ability to convert procedural materials into standard Python programs and fine-tune a large pre-trained vision-language model (VLM) to generate such programs from input images. To enable effective fine-tuning, we also contribute an open-source procedural material dataset and propose to perform program-level augmentation by prompting another pre-trained large language model (LLM). Through extensive evaluation, we show that our method outperforms previous methods on both synthetic and real-world examples.

We present Structured World Modeling for Policy Optimization (SWMPO), a framework for unsupervised learning of neurosymbolic Finite State Machines (FSM) that capture environmental structure for policy optimization. Traditional unsupervised world modeling methods rely on unstructured representations, such as neural networks, that do not explicitly represent high-level patterns within the system (e.g., patterns in the dynamics of regions such as \emph{water} and \emph{land}). Instead, SWMPO models the environment as a finite state machine (FSM), where each state corresponds to a specific region with distinct dynamics. This structured representation can then be leveraged for tasks like policy optimization. Previous works that synthesize FSMs for this purpose have been limited to discrete spaces, not continuous spaces. Instead, our proposed FSM synthesis algorithm operates in an unsupervised manner, leveraging low-level features from unprocessed, non-visual data, making it adaptable across various domains. The synthesized FSM models are expressive enough to be used in a model-based Reinforcement Learning scheme that leverages offline data to efficiently synthesize environment-specific world models. We demonstrate the advantages of SWMPO by benchmarking its environment modeling capabilities in simulated environments.

We introduce a new approach to *agent programming*, the development of LLM-based agents. Current approaches to agent programming often entangle two aspects of agent design: the core workflow logic and the inference-time strategy (e.g., tree search). We introduce *probabilistic angelic nondeterminism* (PAN), a programming model that disentangles these two concerns, allowing the programmer to describe the agent workflow and independently experiment with different inference-time strategies by simply changing a few inputs. We provide an implementation of PAN in Python as the EnCompass framework, which uses a Python decorator to compile agent workflow programs into a search space. We present three case studies that demonstrate how the framework lets the programmer quickly improve the reliability of an agent and easily switch between different inference-time strategies, all with little additional coding.

As AI becomes more closely integrated with peoples' daily activities, socially intelligent AI that can understand and interact seamlessly with humans in daily lives is increasingly important. However, current works in AI social reasoning all rely on language-only or language-dominant approaches to benchmark and training models, resulting in systems that are improving in verbal communication but struggle with nonverbal social understanding. To address this limitation, we tap into a novel data source rich in nonverbal social interactions -- mime videos. Mimes refer to the art of expression through gesture and movement without spoken words, which presents unique challenges and opportunities in interpreting nonverbal social communication. We contribute a new dataset called MimeQA, obtained by sourcing ~8 hours of videos clips from YouTube and developing a comprehensive video question-answering benchmark comprising 806 carefully annotated and verified question-answer pairs, designed to probe nonverbal social reasoning capabilities. Using MimeQA, we evaluate state-of-the-art video large language models (VideoLLMs) and find that they achieve low accuracy, generally ranging from 20-30%, while humans score 86\%. Our analysis reveals that VideoLLMs often fail to ground imagined objects and over-rely on the text prompt while ignoring subtle nonverbal interactions. We hope to inspire future work in AI models that embody true social intelligence capable of interpreting non-verbal human interactions.

We present new techniques for exact and approximate inference in discrete probabilistic programs, based on two new ways of exploiting lazy evaluation. First, we show how knowledge compilation, a state-of-the art technique for exact inference in discrete probabilistic programs, can be made lazy, enabling asymptotic speed-ups. Second, we show how a probabilistic program’s lazy semantics naturally give rise to a division of its random choices into subproblems, which can be solved in sequence by sequential Monte Carlo with locallyoptimal proposals automatically computed via lazy knowledge compilation. We implement our approach in a new tool, Pluck , and evaluate its performance against state-of-the-art approaches to inference in discrete probabilistic languages. We !nd that on a suite of inference benchmarks, lazy knowledge compilation can be faster than state-of-the-art approaches, sometimes by orders of magnitude.

We present batch-based consistency, a new approach for consistency optimization that allows programmers to specialize consistency with application-level integrity properties. We implement the approach with a two-step process: we statically infer optimal consistency requirements for executions of bounded sets of operations, and then, use the inferred requirements to parameterize a new distributed protocol to relax operation reordering at run time when it is safe to do so. Our approach supports standard notions of consistency. We implement batch-based consistency in Peepco , demonstrate its expressiveness for partial data replication, and examine Peepco’s run-time performance impact in different settings.

Large language models have shown remarkable aptitude in code generation, but still struggle to perform complex tasks. Self-repair---in which the model debugs and repairs its own code---has recently become a popular way to boost performance in these settings. However, despite its increasing popularity, existing studies of self-repair have been limited in scope; in many settings, its efficacy thus remains poorly understood. In this paper, we analyze Code Llama, GPT-3.5 and GPT-4's ability to perform self-repair on problems taken from HumanEval and APPS. We find that when the cost of carrying out repair is taken into account, performance gains are often modest, vary a lot between subsets of the data, and are sometimes not present at all. We hypothesize that this is because self-repair is bottlenecked by the model's ability to provide feedback on its own code; using a stronger model to artificially boost the quality of the feedback, we observe substantially larger performance gains. Similarly, a small-scale study in which we provide GPT-4 with feedback from human participants suggests that even for the strongest models, self-repair still lags far behind what can be achieved with human-level debugging.

We present Code Reasoning, Understanding, and eXecution Evaluation, a benchmark consisting of 800 Python functions (3-13 lines). Each function comes with an input-output pair, leading to two natural tasks: input prediction and output prediction. First, we propose a general recipe for generating our execution benchmark by sampling from a model, which can be used for more challenging versions of the benchmark if needed. Second, we evaluate twenty code models on our benchmark and discover that many recent high-scoring models on HumanEval show no improvements on our benchmark. Third, we show that simple CoT and fine-tuning schemes can improve performance on our benchmark but remain far from solving it. The best setup, GPT-4 with chain of thought (CoT), achieves a pass@1 of 75% and 81% on input and output prediction, respectively. In contrast, Code Llama 34B achieves a pass@1 of 50% and 46% on input and output prediction. When it comes to reasoning about code, GPT-4 has a huge edge over other models but still fails consistently on some surprisingly simple Python programs.

Skills are temporal abstractions that are intended to improve reinforcement learning (RL) performance through hierarchical RL. Despite our intuition about the properties of an environment that make skills useful, a precise characterization has been absent. We provide the first such characterization, focusing on the utility of deterministic skills in deterministic sparse-reward environments with finite action spaces. We show theoretically and empirically that RL performance gain from skills is worse in environments where solutions to states are less compressible. Additional theoretical results suggest that skills benefit exploration more than they benefit learning from existing experience, and that using unexpressive skills such as macroactions may worsen RL performance. We hope our findings can guide research on automatic skill discovery and help RL practitioners better decide when and how to use skills.

Robots are often built from standardized assemblies, (e.g. arms, legs, or fingers), but each robot must be trained from scratch to control all the actuators of all the parts together. In this paper we demonstrate a new approach that takes a single robot and its controller as input and produces a set of modular controllers for each of these assemblies such that when a new robot is built from the same parts, its control can be quickly learned by reusing the modular controllers. We achieve this with a framework called MeMo which learns (Me)aningful, (Mo)dular controllers. Specifically, we propose a novel modularity objective to learn an appropriate division of labor among the modules. We demonstrate that this objective can be optimized simultaneously with standard behavior cloning loss via noise injection. We benchmark our framework in locomotion and grasping environments on simple to complex robot morphology transfer. We also show that the modules help in task transfer. On both structure and task transfer, MeMo achieves improved training efficiency to graph neural network and Transformer baselines.

Pure is a new programming model and runtime system explicitly designed to take advantage of shared memory within nodes in the context of a mostly message passing interface enhanced with the ability to use tasks to make use of idle cores. Pure leverages shared memory in two ways: (a) by allowing cores to steal work from each other while waiting on messages to arrive, and, (b) by leveraging efficient lock-free data structures in shared memory to achieve high-performance messaging and collective operations between the ranks within nodes. We use microbenchmarks to evaluate Pure's key messaging and collective features and also show application speedups up to 2.1× on the CoMD molecular dynamics and the miniAMR adaptive mesh refinement applications scaling up to 4,096 cores.

We present a new dataset for learning to solve, explain, and generate university-level STEM questions from 27 courses across a dozen departments in seven universities. We scale up previous approaches to questions from courses in the departments of Mechanical Engineering, Materials Science and Engineering, Chemistry, Electrical Engineering, Computer Science, Physics, Earth Atmospheric and Planetary Sciences, Economics, Mathematics, Biological Engineering, Data Systems, and Society, and Statistics. We visualize similarities and differences between questions across courses. We demonstrate that a large foundation model is able to generate questions that are as appropriate and at the same difficulty level as human-written questions.

Logical reasoning, i.e., deductively inferring the truth value of a conclusion from a set of premises, is an important task for artificial intelligence with wide potential impacts on science, mathematics, and society.While many prompting-based strategies have been proposed to enable Large Language Models (LLMs) to do such reasoning more effectively, they still appear unsatisfactory, often failing in subtle and unpredictable ways.In this work, we investigate the validity of instead reformulating such tasks as modular neurosymbolic programming, which we call LINC: Logical Inference via Neurosymbolic Computation.In LINC, the LLM acts as a semantic parser, translating premises and conclusions from natural language to expressions in first-order logic.These expressions are then offloaded to an external theorem prover, which symbolically performs deductive inference.Leveraging this approach, we observe significant performance gains on FOLIO and a balanced subset of ProofWriter for three different models in nearly all experimental conditions we evaluate.On ProofWriter, augmenting the comparatively small open-source StarCoder+ (15.5B parameters) with LINC even outperforms GPT-3.5 and GPT-4 with Chain-of-Thought (CoT) prompting by an absolute 38% and 10%, respectively.When used with GPT-4, LINC scores 26% higher than CoT on ProofWriter while performing comparatively on FOLIO.Further analysis reveals that although both methods on average succeed roughly equally often on this dataset, they exhibit distinct and complementary failure modes.We thus provide promising evidence for how logical reasoning over natural language can be tackled through jointly leveraging LLMs alongside symbolic provers.All corresponding code is publicly available.

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.

We present a new algorithm that synthesizes functional reactive programs from observation data. The key novelty is to iterate between a functional synthesis step, which attempts to generate a transition function over observed states, and an automata synthesis step, which adds any additional latent state necessary to fully account for the observations. We develop a functional reactive DSL called Autumn that can express a rich variety of causal dynamics in time-varying, Atari-style grid worlds, and apply our method to synthesize Autumn programs from data. We evaluate our algorithm on a benchmark suite of 30 Autumn programs as well as a third-party corpus of grid-world-style video games. We find that our algorithm synthesizes 27 out of 30 programs in our benchmark suite and 21 out of 27 programs from the third-party corpus, including several programs describing complex latent state transformations, and from input traces containing hundreds of observations. We expect that our approach will provide a template for how to integrate functional and automata synthesis in other induction domains.

We present a new general-purpose synthesis technique for generating programs from input-output examples. Our method, called metric program synthesis, relaxes the observational equivalence idea (used widely in bottom-up enumerative synthesis) into a weaker notion of observational similarity, with the goal of reducing the search space that the synthesizer needs to explore. Our method clusters programs into equivalence classes based on an expert-provided distance metric and constructs a version space that compactly represents “approximately correct” programs. Then, given a “close enough” program sampled from this version space, our approach uses a distance-guided repair algorithm to find a program that exactly matches the given input-output examples. We have implemented our proposed metric program synthesis technique in a tool called SyMetric and evaluate it in three different domains considered in prior work. Our evaluation shows that SyMetric outperforms other domain-agnostic synthesizers that use observational equivalence and that it achieves results competitive with domain-specific synthesizers that are either designed for or trained on those domains.

Physical products are often complex assemblies combining a multitude of 3D parts modeled in computer-aided design (CAD) software. CAD designers build up these assemblies by aligning individual parts to one another using constraints called joints. In this paper we introduce JoinABLe, a learning-based method that assembles parts together to form joints. JoinABLe uses the weak supervision available in standard parametric CAD files without the help of object class labels or human guidance. Our results show that by making network predictions over a graph representation of solid models we can outperform multiple baseline methods with an accuracy (79.53%) that approaches human performance (80%). Finally, to support future research we release the Fusion 360 Gallery assembly dataset, containing assemblies with rich information on joints, contact surfaces, holes, and the underlying assembly graph structure.

We present a framework that allows programmers to specify replicated data stores through application logic, data replication scheme, and high-level invariants that needs to be satisfied. From such specifications all the needed consistency requirements can be inferred from traces of executions of the potential data store, to determine the optimal data store coordination. The framework supports arbitrarily complex data store operations and partial data replication. This leads to expressiveness for a wide range of data stores, with significant run-time performance benefits.

Many problem domains, including program synthesis and rewrite-based optimization, require searching astronomically large spaces of programs. Existing approaches often rely on building specialized data structures—version-space algebras, finite tree automata, or e-graphs—to compactly represent such spaces. At their core, all these data structures exploit independence of subterms; as a result, they cannot efficiently represent more complex program spaces, where the choices of subterms are entangled. We introduce equality-constrained tree automata (ECTAs), a new data structure, designed to compactly represent large spaces of programs with entangled subterms. We present efficient algorithms for extracting programs from ECTAs, implemented in a performant Haskell library, ecta. Using the ecta library, we construct Hectare, a type-driven program synthesizer for Haskell. Hectare significantly outperforms a state-of-the-art synthesizer Hoogle+—providing an average speedup of 8×—despite its implementation being an order of magnitude smaller.

We develop the first theory of control-flow graphs from first principles, and use it to create an algorithm for automatically synthesizing many variants of control-flow graph generators from a language’s operational semantics. Our approach first introduces a new algorithm for converting a large class of small-step operational semantics to an abstract machine. It next uses a technique called ”abstract rewriting” to automatically abstract the semantics of a language, which is used both to directly generate a CFG from a program (”interpreted mode”) and to generate standalone code, similar to a human-written CFG generator, for any program in a language. We show how the choice of two abstraction and projection parameters allow our approach to synthesize several families of CFG-generators useful for different kinds of tools. We prove the correspondence between the generated graphs and the original semantics. We provide and prove an algorithm for automatically proving the termination of interpreted-mode generators. In addition to our theoretical results, we have implemented this algorithm in a tool called Mandate, and show that it produces human-readable code on two medium-size languages with 60−80 rules, featuring nearly all intraprocedural control constructs common in modern languages. We then show these CFG-generators were sufficient to build two static analyses atop them. Our work is a promising step towards the grand vision of being able to synthesize all desired tools from the semantics of a programming language.

Abstract Sketch is a popular program synthesis tool that solves for unknowns in a sketch or partial program. However, while Sketch is powerful, it does not directly support modular synthesis of dependencies, potentially limiting scalability. In this paper, we introduce Sketcham, a new technique that modularizes a regular sketch by automatically generating mocks—functions that approximate the behavior of complete implementations—from the sketch’s test suite. For example, if the function f originally calls g, Sketcham creates a mock $$g_m$$ g m from g’s tests and augments the sketch with a version of f that calls $$g_m$$ g m . This change allows the unknowns in f and g to be solved separately, enabling modular synthesis with no extra work from the Sketch user. We evaluated Sketcham on ten benchmarks, performing enough runs to show at a 95% confidence level that Sketcham improves median synthesis performance on six of our ten benchmarks by a factor of up to 5 $$\times $$ × compared to plain Sketch, including one benchmark that times out on Sketch, while exhibiting similar performance on the remaining four. Our results show that Sketcham can achieve modular synthesis by automatically generating mocks from tests.

Deploying deep reinforcement learning in safety-critical settings requires developing algorithms that obey hard constraints during exploration. This paper contributes a first approach toward enforcing formal safety constraints on end-to-end policies with visual inputs. Our approach draws on recent advances in object detection and automated reasoning for hybrid dynamical systems. The approach is evaluated on a novel benchmark that emphasizes the challenge of safely exploring in the presence of hard constraints. Our benchmark draws from several proposed problem sets for safe learning and includes problems that emphasize challenges such as reward signals that are not aligned with safety constraints. On each of these benchmark problems, our algorithm completely avoids unsafe behavior while remaining competitive at optimizing for as much reward as is safe. We characterize safety constraints in terms of a refinement relation on Markov decision processes - rather than directly constraining the reinforcement learning algorithm so that it only takes safe actions, we instead refine the environment so that only safe actions are defined in the environment's transition structure. This has pragmatic system design benefits and, more importantly, provides a clean conceptual setting in which we are able to prove important safety and efficiency properties. These allow us to transform the constrained optimization problem of acting safely in the original environment into an unconstrained optimization in a refined environment.

A landmark challenge for AI is to learn flexible, powerful representations from small numbers of examples. On an important class of tasks, hypotheses in the form of programs provide extreme generalization capabilities from surprisingly few examples. However, whereas large natural few-shot learning image benchmarks have spurred progress in meta-learning for deep networks, there is no comparably big, natural program-synthesis dataset that can play a similar role. This is because, whereas images are relatively easy to label from internet meta-data or annotated by non-experts, generating meaningful input-output examples for program induction has proven hard to scale. In this work, we propose a new way of leveraging unit tests and natural inputs for small programs as meaningful input-output examples for each sub-program of the overall program. This allows us to create a large-scale naturalistic few-shot program-induction benchmark and propose new challenges in this domain. The evaluation of multiple program induction and synthesis algorithms points to shortcomings of current methods and suggests multiple avenues for future work.

We present Omega, a probabilistic programming language with support for counterfactual inference. Counterfactual inference means to observe some fact in the present, and infer what would have happened had some past intervention been taken, e.g. “given that medication was not effective at dose x, what is the probability that it would have been effective at dose 2x?.” We accomplish this by introducing a new operator to probabilistic programming akin to Pearl’s do, define its formal semantics, provide an implementation, and demonstrate its utility through examples in a variety of simulation models.

A key challenge for reinforcement learning is solving long-horizon planning problems. Recent work has leveraged programs to guide reinforcement learning in these settings. However, these approaches impose a high manual burden on the user since they must provide a guiding program for every new task. Partially observed environments further complicate the programming task because the program must implement a strategy that correctly, and ideally optimally, handles every possible configuration of the hidden regions of the environment. We propose a new approach, model predictive program synthesis (MPPS), that uses program synthesis to automatically generate the guiding programs. It trains a generative model to predict the unobserved portions of the world, and then synthesizes a program based on samples from this model in a way that is robust to its uncertainty. In our experiments, we show that our approach significantly outperforms non-program-guided approaches on a set of challenging benchmarks, including a 2D Minecraft-inspired environment where the agent must complete a complex sequence of subtasks to achieve its goal, and achieves a similar performance as using handcrafted programs to guide the agent. Our results demonstrate that our approach can obtain the benefits of program-guided reinforcement learning without requiring the user to provide a new guiding program for every new task.

We present a system for inductive program synthesis called DreamCoder, which inputs a corpus of synthesis problems each specified by one or a few examples, and automatically derives a library of program components and a neural search policy that can be used to efficiently solve other similar synthesis problems. The library and search policy bootstrap each other iteratively through a variant of "wake-sleep" approximate Bayesian learning. A new refactoring algorithm based on E-graph matching identifies common sub-components across synthesized programs, building a progressively deepening library of abstractions capturing the structure of the input domain. We evaluate on eight domains including classic program synthesis areas and AI tasks such as planning, inverse graphics, and equation discovery. We show that jointly learning the library and neural search policy leads to solving more problems, and solving them more quickly.

Parametric computer-aided design (CAD) is a standard paradigm used to design manufactured objects, where a 3D shape is represented as a program supported by the CAD software. Despite the pervasiveness of parametric CAD and a growing interest from the research community, currently there does not exist a dataset of realistic CAD models in a concise programmatic form. In this paper we present the Fusion 360 Gallery , consisting of a simple language with just the sketch and extrude modeling operations, and a dataset of 8,625 human design sequences expressed in this language. We also present an interactive environment called the Fusion 360 Gym , which exposes the sequential construction of a CAD program as a Markov decision process, making it amendable to machine learning approaches. As a use case for our dataset and environment, we define the CAD reconstruction task of recovering a CAD program from a target geometry. We report results of applying state-of-the-art methods of program synthesis with neurally guided search on this task.

We study the problem of inferring communication structures that can solve cooperative multi-agent planning problems while minimizing the amount of communication. We quantify the amount of communication as the maximum degree of the communication graph; this metric captures settings where agents have limited bandwidth. Minimizing communication is challenging due to the combinatorial nature of both the decision space and the objective; for instance, we cannot solve this problem by training neural networks using gradient descent. We propose a novel algorithm that synthesizes a control policy that combines a programmatic communication policy used to generate the communication graph with a transformer policy network used to choose actions. Our algorithm first trains the transformer policy, which implicitly generates a "soft" communication graph; then, it synthesizes a programmatic communication policy that "hardens" this graph, forming a neurosymbolic transformer. Our experiments demonstrate how our approach can synthesize policies that generate low-degree communication graphs while maintaining near-optimal performance.

Many aspects of human reasoning, including language, require learning rules from very little data. Humans can do this, often learning systematic rules from very few examples, and combining these rules to form compositional rule-based systems. Current neural architectures, on the other hand, often fail to generalize in a compositional manner, especially when evaluated in ways that vary systematically from training. In this work, we present a neuro-symbolic model which learns entire rule systems from a small set of examples. Instead of directly predicting outputs from inputs, we train our model to induce the explicit system of rules governing a set of previously seen examples, drawing upon techniques from the neural program synthesis literature. Our rule-synthesis approach outperforms neural meta-learning techniques in three domains: an artificial instruction-learning domain used to evaluate human learning, the SCAN challenge datasets, and learning rule-based translations of number words into integers for a wide range of human languages.

Program synthesis techniques construct or infer programs from user-provided specifications, such as input-output examples. Yet most specifications, especially those given by end-users, leave the synthesis problem radically ill-posed, because many programs may simultaneously satisfy the specification. Prior work resolves this ambiguity by using various inductive biases, such as a preference for simpler programs. This work introduces a new inductive bias derived by modeling the program synthesis task as rational communication, drawing insights from recursive reasoning models of pragmatics. Given a specification, we score a candidate program both on its consistency with the specification, and also whether a rational speaker would chose this particular specification to communicate that program. We develop efficient algorithms for such an approach when learning from input-output examples, and build a pragmatic program synthesizer over a simple grid-like layout domain. A user study finds that end-user participants communicate more effectively with the pragmatic program synthesizer over a non-pragmatic one.

We present a new approach to semantic code search based on equational reasoning, and the Yogo tool implementing this approach. Our approach works by considering not only the dataflow graph of a function, but also the dataflow graphs of all equivalent functions reachable via a set of rewrite rules. In doing so, it can recognize an operation even if it uses alternate APIs, is in a different but mathematically-equivalent form, is split apart with temporary variables, or is interleaved with other code. Furthermore, it can recognize when code is an instance of some higher-level concept such as iterating through a file. Because of this, from a single query, Yogo can find equivalent code in multiple languages. Our evaluation further shows the utility of Yogo beyond code search: encoding a buggy pattern as a Yogo query, we found a bug in Oracle’s Graal compiler which had been missed by a hand-written static analyzer designed for that exact kind of bug. Yogo is built on the Cubix multi-language infrastructure, and currently supports Java and Python.

Optimizing the physical data storage and retrieval of data are two key database management problems. In this paper, we propose a language that can express both a relational query and the layout of its data. Our language can express a wide range of physical database layouts, going well beyond the row- and column-based methods that are widely used in database management systems. We use deductive program synthesis to turn a high-level relational representation of a database query into a highly optimized low-level implementation which operates on a specialized layout of the dataset. We build an optimizing compiler for this language and conduct experiments using a popular database benchmark, which shows that the performance of our specialized queries is better than a state-of-the-art in memory compiled database system while achieving an order-of-magnitude reduction in memory use.

We present Lifty, a domain-specific language for data-centric applications that manipulate sensitive data. A Lifty programmer annotates the sources of sensitive data with declarative security policies, and the language statically and automatically verifies that the application handles the data according to the policies. Moreover, if verification fails, Lifty suggests a provably correct repair, thereby easing the programmer burden of implementing policy enforcing code throughout the application. The main insight behind Lifty is to encode information flow control using liquid types, an expressive yet decidable type system. Liquid types enable fully automatic checking of complex, data dependent policies, and power our repair mechanism via type-driven error localization and patch synthesis. Our experience using Lifty to implement three case studies from the literature shows that (1) the Lifty policy language is sufficiently expressive to specify many real-world policies, (2) the Lifty type checker is able to verify secure programs and find leaks in insecure programs quickly, and (3) even if the programmer leaves out all policy enforcing code, the Lifty repair engine is able to patch all leaks automatically within a reasonable time.

We formulate numerically-robust inductive proof rules for unbounded stability and safety properties of continuous dynamical systems. These induction rules robustify standard notions of Lyapunov functions and barrier certificates so that they can tolerate small numerical errors. In this way, numerically-driven decision procedures can establish a sound and relative-complete proof system for unbounded properties of very general nonlinear systems. We demonstrate the effectiveness of the proposed rules for rigorously verifying unbounded properties of various nonlinear systems, including a challenging powertrain control model.

Our goal is to build systems which write code automatically from the kinds of specifications humans can most easily provide, such as examples and natural language instruction. The key idea of this work is that a flexible combination of pattern recognition and explicit reasoning can be used to solve these complex programming problems. We propose a method for dynamically integrating these types of information. Our novel intermediate representation and training algorithm allow a program synthesis system to learn, without direct supervision, when to rely on pattern recognition and when to perform symbolic search. Our model matches the memorization and generalization performance of neural synthesis and symbolic search, respectively, and achieves state-of-the-art performance on a dataset of simple English description-to-code programming problems.

Programming languages allow us to express complex predicates, but existing inference methods are unable to condition probabilistic models on most of them. To support a broader class of predicates, we develop an inference procedure called predicate exchange, which softens predicates. A soft predicate quantifies the extent to which values of model variables are consistent with its hard counterpart. We substitute the likelihood term in the Bayesian posterior with a soft predicate, and develop a variant of replica exchange MCMC to draw posterior samples. We implement predicate exchange as a language agnostic tool which performs a nonstandard execution of a probabilistic program. We demonstrate the approach on sequence models of health and inverse rendering.

The great promise of modular robots is the ability to create on demand robots; however, choosing the “right” design based on a task is still a challenging problem. In this paper, we present an approach to automatically synthesize both the design and control for modular robots from a task description. In particular, we focus on manipulators composed of one degree-of-freedom (DoF) modules. Our approach is able to handle partially infeasible tasks by either identifying the infeasible part and finding a design that satisfies the feasible part or searching for multiple designs that together satisfy the entire task. We compare our approach to a baseline genetic algorithm in a series of increasingly complex environments.

We present a neural program synthesis approach integrating components which write, execute, and assess code to navigate the search space of possible programs. We equip the search process with an interpreter or a read-eval-print-loop (REPL), which immediately executes partially written programs, exposing their semantics. The REPL addresses a basic challenge of program synthesis: tiny changes in syntax can lead to huge changes in semantics. We train a pair of models, a policy that proposes the new piece of code to write, and a value function that assesses the prospects of the code written so-far. At test time we can combine these models with a Sequential Monte Carlo algorithm. We apply our approach to two domains: synthesizing text editing programs and inferring 2D and 3D graphics programs.

As machine learning systems are increasingly used to make real world legal and financial decisions, it is of paramount importance that we develop algorithms to verify that these systems do not discriminate against minorities. We design a scalable algorithm for verifying fairness specifications. Our algorithm obtains strong correctness guarantees based on adaptive concentration inequalities; such inequalities enable our algorithm to adaptively take samples until it has enough data to make a decision. We implement our algorithm in a tool called VeriFair, and show that it scales to large machine learning models, including a deep recurrent neural network that is more than five orders of magnitude larger than the largest previously-verified neural network. While our technique only gives probabilistic guarantees due to the use of random samples, we show that we can choose the probability of error to be extremely small.

A key challenge in program synthesis is synthesizing programs that use libraries, which most real-world software does. The current state of the art is to model libraries with mock library implementations that perform the same function in a simpler way. However, mocks may still be large and complex, and must include many implementation details, both of which could limit synthesis performance. To address this problem, we introduce JLibSketch, a Java program synthesis tool that allows library behavior to be described with algebraic specifications, which are rewrite rules for sequences of method calls, e.g., encryption followed by decryption (with the same key) is the identity. JLibSketch implements rewrite rules by compiling JLibSketch problems into problems for the Sketch program synthesis tool. More specifically, after compilation, library calls are represented by abstract data types (ADTs), and rewrite rules manipulate those ADTs. We formalize compilation and prove it sound and complete if the rewrite rules are ordered and non-unifiable. We evaluated JLibSketch by using it to synthesize nine programs that use libraries from three domains: data structures, cryptography, and file systems. We found that algebraic specifications are, on average, about half the size of mocks. We also found that algebraic specifications perform better than mocks on seven of the nine programs, sometimes significantly so, and perform equally well on the last two programs. Thus, we believe that JLibSketch takes an important step toward synthesis of programs that use libraries.

We propose $$\delta $$ -complete decision procedures for solving satisfiability of nonlinear SMT problems over real numbers that contain universal quantification and a wide range of nonlinear functions. The methods combine interval constraint propagation, counterexample-guided synthesis, and numerical optimization. In particular, we show how to handle the interleaving of numerical and symbolic computation to ensure delta-completeness in quantified reasoning. We demonstrate that the proposed algorithms can handle various challenging global optimization and control synthesis problems that are beyond the reach of existing solvers.

Program synthesis is a class of regression problems where one seeks a solution, in the form of a source-code program, mapping the inputs to their corresponding outputs exactly. Due to its precise and combinatorial nature, program synthesis is commonly formulated as a constraint satisfaction problem, where input-output examples are encoded as constraints and solved with a constraint solver. A key challenge of this formulation is scalability: while constraint solvers work well with a few well-chosen examples, a large set of examples can incur significant overhead in both time and memory. We describe a method to discover a subset of examples that is both small and representative: the subset is constructed iteratively, using a neural network to predict the probability of unchosen examples conditioned on the chosen examples in the subset, and greedily adding the least probable example. We empirically evaluate the representativeness of the subsets constructed by our method, and demonstrate such subsets can significantly improve synthesis time and stability.

While deep reinforcement learning has successfully solved many challenging control tasks, its real-world applicability has been limited by the inability to ensure the safety of learned policies. We propose an approach to verifiable reinforcement learning by training decision tree policies, which can represent complex policies (since they are nonparametric), yet can be efficiently verified using existing techniques (since they are highly structured). The challenge is that decision tree policies are difficult to train. We propose VIPER, an algorithm that combines ideas from model compression and imitation learning to learn decision tree policies guided by a DNN policy (called the oracle) and its Q-function, and show that it substantially outperforms two baselines. We use VIPER to (i) learn a provably robust decision tree policy for a variant of Atari Pong with a symbolic state space, (ii) learn a decision tree policy for a toy game based on Pong that provably never loses, and (iii) learn a provably stable decision tree policy for cart-pole. In each case, the decision tree policy achieves performance equal to that of the original DNN policy.

Successful approaches to program induction require a hand-engineered domain-specific language (DSL), constraining the space of allowed programs and imparting prior knowledge of the domain. We contribute a program induction algorithm that learns a DSL while jointly training a neural network to efficiently search for programs in the learned DSL. We use our model to synthesize functions on lists, edit text, and solve symbolic regression problems, showing how the model learns a domain-specific library of program components for expressing solutions to problems in the domain.

We introduce a model that learns to convert simple hand drawings into graphics programs written in a subset of \LaTeX.~The model combines techniques from deep learning and program synthesis. We learn a convolutional neural network that proposes plausible drawing primitives that explain an image. These drawing primitives are a specification (spec) of what the graphics program needs to draw. We learn a model that uses program synthesis techniques to recover a graphics program from that spec. These programs have constructs like variable bindings, iterative loops, or simple kinds of conditionals. With a graphics program in hand, we can correct errors made by the deep network and extrapolate drawings.

We present a new algorithm to generate minimal, stable, and symbolic corrections to an input that will cause a neural network with ReLU activations to change its output. We argue that such a correction is a useful way to provide feedback to a user when the network's output is different from a desired output. Our algorithm generates such a correction by solving a series of linear constraint satisfaction problems. The technique is evaluated on three neural network models: one predicting whether an applicant will pay a mortgage, one predicting whether a first-order theorem can be proved efficiently by a solver using certain heuristics, and the final one judging whether a drawing is an accurate rendition of a canonical drawing of a cat.

We present a new approach for building source-to-source transformations that can run on multiple programming languages, based on a new way of representing programs called incremental parametric syntax. We implement this approach in Haskell in our Cubix system, and construct incremental parametric syntaxes for C, Java, JavaScript, Lua, and Python. We demonstrate a whole-program refactoring tool that runs on all of them, along with three smaller transformations that each run on several. Our evaluation shows that (1) once a transformation is written, little work is required to configure it for a new language (2) transformations built this way output readable code which preserve the structure of the original, according to participants in our human study, and (3) our transformations can still handle language corner-cases, as validated on compiler test suites.

Delimited continuations are the mother of all monads! So goes the slogan inspired by Filinski’s 1994 paper, which showed that delimited continuations can implement any monadic effect, letting the programmer use an effect as easily as if it was built into the language. It’s a shame that not many languages have delimited continuations. Luckily, exceptions and state are also the mother of all monads! In this Pearl, we show how to implement delimited continuations in terms of exceptions and state, a construction we call thermometer continuations . While traditional implementations of delimited continuations require some way of ”capturing” an intermediate state of the computation, the insight of thermometer continuations is to reach this intermediate state by replaying the entire computation from the start, guiding it using a recording so that the same thing happens until the captured point. Along the way, we explain delimited continuations and monadic reflection, show how the Filinski construction lets thermometer continuations express any monadic effect, share an elegant special-case for nondeterminism, and discuss why our construction is not prevented by theoretical results that exceptions and state cannot macro-express continuations.

While computer-aided design is a major part of many modern manufacturing pipelines, the design files typically generated describe raw geometry. Lost in this representation is the procedure by which these designs were generated. In this paper, we present a method for reverse-engineering the process by which 3D models may have been generated, in the language of constructive solid geometry (CSG). Observing that CSG is a formal grammar, we formulate this inverse CSG problem as a program synthesis problem. Our solution is an algorithm that couples geometric processing with state-of-the-art program synthesis techniques. In this scheme, geometric processing is used to convert the mixed discrete and continuous domain of CSG trees to a pure discrete domain where modern program synthesizers excel. We demonstrate the efficiency and scalability of our algorithm on several different examples, including those with over 100 primitive parts. We show that our algorithm is able to find simple programs which are close to the ground truth, and demonstrate our method's applicability in mesh re-editing. Finally, we compare our method to prior state-of-the-art. We demonstrate that our algorithm dominates previous methods in terms of resulting CSG compactness and runtime, and can handle far more complex input meshes than any previous method.

We introduce DemoMatch, a tool for API discovery that allows the user to discover how to implement functionality using a software framework by demonstrating the functionality in existing applications built with the same framework. DemoMatch matches the demonstrations against a database of execution traces called Semeru and generates code snippets explaining how to use the functionality. We evaluated DemoMatch on several case studies involving Java Swing and Eclipse RCP.

Entity matching (EM) is a critical part of data integration and cleaning. In many applications, the users need to understand why two entities are considered a match, which reveals the need for interpretable and concise EM rules. We model EM rules in the form of General Boolean Formulas (GBFs) that allows arbitrary attribute matching combined by conjunctions (∨), disjunctions (∧), and negations. (¬) GBFs can generate more concise rules than traditional EM rules represented in disjunctive normal forms (DNFs). We use program synthesis, a powerful tool to automatically generate rules (or programs) that provably satisfy a high-level specification, to automatically synthesize EM rules in GBF format, given only positive and negative matching examples.

This paper presents natural synthesis, which generalizes the proof-theoretic synthesis technique to support very expressive logic theories. This approach leverages the natural proof methodology and reduces an intractable, unbounded-size synthesis problem to a tractable, bounded-size synthesis problem, which is amenable to be handled by modern inductive synthesis engines. The synthesized program admits a natural proof and is a provably-correct solution to the original synthesis problem. We explore the natural synthesis approach in the domain of imperative data-structure manipulations and present a novel syntax-guided synthesizer based on natural synthesis. The input to our system is a program template together with a rich functional specification that the synthesized program must meet. Our system automatically produces a program implementation along with necessary proof artifacts, namely loop invariants and ranking functions, and guarantees the total correctness with a natural proof. Experiments show that our natural synthesizer can efficiently produce provably-correct implementations for sorted lists and binary search trees. To our knowledge, this is the first system that can automatically synthesize these programs, their functional correctness and their termination in tandem from bare-bones control flow skeletons.

Entity matching (EM) is a critical part of data integration. We study how to synthesize entity matching rules from positive-negative matching examples. The core of our solution is program synthesis , a powerful tool to automatically generate rules (or programs) that satisfy a given high-level specification, via a predefined grammar. This grammar describes a General Boolean Formula ( GBF ) that can include arbitrary attribute matching predicates combined by conjunctions (∧), disjunctions (∨) and negations (¬), and is expressive enough to model EM problems, from capturing arbitrary attribute combinations to handling missing attribute values. The rules in the form of GBF are more concise than traditional EM rules represented in Disjunctive Normal Form ( DNF ). Consequently, they are more interpretable than decision trees and other machine learning algorithms that output deep trees with many branches. We present a new synthesis algorithm that, given only positive-negative examples as input, synthesizes EM rules that are effective over the entire dataset. Extensive experiments show that we outperform other interpretable rules (e.g., decision trees with low depth) in effectiveness, and are comparable with non-interpretable tools (e.g., decision trees with high depth, gradient-boosting trees, random forests and SVM).

This paper addresses the problem of creating simplifiers for logic formulas based on conditional term rewriting. In particular, the paper focuses on a program synthesis application where formula simplifications have been shown to have a significant impact. We show that by combining machine learning techniques with constraint-based synthesis, it is possible to synthesize a formula simplifier fully automatically from a corpus of representative problems, making it possible to create formula simplifiers tailored to specific problem domains. We demonstrate the benefits of our approach for synthesis benchmarks from the SyGuS competition and automated grading.

Symbolic execution is a powerful program analysis technique, but it is difficult to apply to programs built using frameworks such as Swing and Android, because the framework code itself is hard to symbolically execute. The standard solution is to manually create a framework model that can be symbolically executed, but developing and maintaining a model is difficult and error-prone. In this paper, we present Pasket, a new system that takes a first step toward automatically generating Java framework models to support symbolic execution. Pasket's focus is on creating models by instantiating design patterns. Pasket takes as input class, method, and type information from the framework API, together with tutorial programs that exercise the framework. From these artifacts and Pasket's internal knowledge of design patterns, Pasket synthesizes a framework model whose behavior on the tutorial programs matches that of the original framework. We evaluated Pasket by synthesizing models for subsets of Swing and Android. Our results show that the models derived by Pasket are sufficient to allow us to use off-the-shelf symbolic execution tools to analyze Java programs that rely on frameworks.

Towards learning programs from data, we introduce the problem of sampling programs from posterior distributions conditioned on that data. Within this setting, we propose an algorithm that uses a symbolic solver to efficiently sample programs. The proposal combines constraint-based program synthesis with sampling via random parity constraints. We give theoretical guarantees on how well the samples approximate the true posterior, and have empirical results showing the algorithm is efficient in practice, evaluating our approach on 22 program learning problems in the domains of text editing and computer-aided programming.

We introduce a framework allowing domain experts to manipulate computational terms in the interest of deriving better, more efficient implementations.It employs deductive reasoning to generate provably correct efficient implementations from a very high-level specification of an algorithm, and inductive constraint-based synthesis to improve automation. Semantic information is encoded into program terms through the use of refinement types.

This paper demonstrates a novel combination of program synthesis and verification to lift stencil computations from low-level Fortran code to a high-level summary expressed using a predicate language. The technique is sound and mostly automated, and leverages counter-example guided inductive synthesis (CEGIS) to find provably correct translations. Lifting existing code to a high-performance description language has a number of benefits, including maintainability and performance portability. For example, our experiments show that the lifted summaries can enable domain specific compilers to do a better job of parallelization as compared to an off-the-shelf compiler working on the original code, and can even support fully automatic migration to hardware accelerators such as GPUs. We have implemented verified lifting in a system called STNG and have evaluated it using microbenchmarks, mini-apps, and real-world applications. We demonstrate the benefits of verified lifting by first automatically summarizing Fortran source code into a high-level predicate language, and subsequently translating the lifted summaries into Halide, with the translated code achieving median performance speedups of 4.1X and up to 24X for non-trivial stencils as compared to the original implementation.

We present a method for synthesizing recursive functions that provably satisfy a given specification in the form of a polymorphic refinement type. We observe that such specifications are particularly suitable for program synthesis for two reasons. First, they offer a unique combination of expressive power and decidability, which enables automatic verification—and hence synthesis—of nontrivial programs. Second, a type-based specification for a program can often be effectively decomposed into independent specifications for its components, causing the synthesizer to consider fewer component combinations and leading to a combinatorial reduction in the size of the search space. At the core of our synthesis procedure is a newalgorithm for refinement type checking, which supports specification decomposition. We have evaluated our prototype implementation on a large set of synthesis problems and found that it exceeds the state of the art in terms of both scalability and usability. The tool was able to synthesize more complex programs than those reported in prior work (several sorting algorithms and operations on balanced search trees), as well as most of the benchmarks tackled by existing synthesizers, often starting from a more concise and intuitive user input.

We present an approach for dynamic information flow control across the application and database. Our approach reduces the amount of policy code required, yields formal guarantees across the application and database, works with existing relational database implementations, and scales for realistic applications. In this paper, we present a programming model that factors out information flow policies from application code and database queries, a dynamic semantics for the underlying $^JDB$ core language, and proofs of termination-insensitive non-interference and policy compliance for the semantics. We implement these ideas in Jacqueline, a Python web framework, and demonstrate feasibility through three application case studies: a course manager, a health record system, and a conference management system used to run an academic workshop. We show that in comparison to traditional applications with hand-coded policy checks, Jacqueline applications have 1) a smaller trusted computing base, 2) fewer lines of policy code, and 2) reasonable, often negligible, additional overheads.

We present AUTOGEN---an algorithm that for a wide class of dynamic programming (DP) problems automatically discovers highly efficient cache-oblivious parallel recursive divide-and-conquer algorithms from inefficient iterative descriptions of DP recurrences. AUTOGEN analyzes the set of DP table locations accessed by the iterative algorithm when run on a DP table of small size, and automatically identifies a recursive access pattern and a corresponding provably correct recursive algorithm for solving the DP recurrence. We use AUTOGEN to autodiscover efficient algorithms for several well-known problems. Our experimental results show that several autodiscovered algorithms significantly outperform parallel looping and tiled loop-based algorithms. Also these algorithms are less sensitive to fluctuations of memory and bandwidth compared with their looping counterparts, and their running times and energy profiles remain relatively more stable. To the best of our knowledge, AUTOGEN is the first algorithm that can automatically discover new nontrivial divide-and-conquer algorithms.

Many web applications store persistent data in databases. During execution, such applications spend a significant amount of time communicating with the database for retrieval and storing of persistent data over the network. These network round-trips represent a significant fraction of the overall execution time for many applications (especially those that issue a lot of database queries) and, as a result, increase their latency. While there has been prior work that aims to eliminate round-trips by batching queries, they are limited by (1) a requirement that developers manually identify batching opportunities, or (2) the fact that they employ static program analysis techniques that cannot exploit many opportunities for batching, as many of these opportunities require knowing precise information about the state of the running program. In this article, we present S loth , a new system that extends traditional lazy evaluation to expose query batching opportunities during application execution, even across loops, branches, and method boundaries. Many such opportunities often require expensive and sophisticated static analysis to recognize from the application source code. Rather than doing so, S loth instead makes use of dynamic analysis to capture information about the program state and, based on that information, decides how to batch queries and when to issue them to the database. We formalize extended lazy evaluation and prove that it preserves program semantics when executed under standard semantics. Furthermore, we describe our implementation of S loth and our experience in evaluating S loth using over 100 benchmarks from two large-scale open-source applications, in which S loth achieved up to a 3 × reduction in page load time by delaying computation using extended lazy evaluation.

We introduce an unsupervised learning algorithmthat combines probabilistic modeling with solver-based techniques for program synthesis.We apply our techniques to both a visual learning domain and a language learning problem,showing that our algorithm can learn many visual concepts from only a few examplesand that it can recover some English inflectional morphology.Taken together, these results give both a new approach to unsupervised learning of symbolic compositional structures,and a technique for applying program synthesis tools to noisy data.

Sketch-based synthesis, epitomized by the Sketch tool, lets developers synthesize software starting from a partial program, also called a sketch or template. This paper presents JSketch, a tool that brings sketch-based synthesis to Java. JSketch's input is a partial Java program that may include holes, which are unknown constants, expression generators, which range over sets of expressions, and class generators, which are partial classes. JSketch then translates the synthesis problem into a Sketch problem; this translation is complex because Sketch is not object-oriented. Finally, JSketch synthesizes an executable Java program by interpreting the output of Sketch.

This article studies undefined behavior arising in systems programming languages such as C/C++. Undefined behavior bugs lead to unpredictable and subtle systems behavior, and their effects can be further amplified by compiler optimizations. Undefined behavior bugs are present in many systems, including the Linux kernel and the Postgres database. The consequences range from incorrect functionality to missing security checks. This article proposes a formal and practical approach that finds undefined behavior bugs by finding “unstable code” in terms of optimizations that leverage undefined behavior. Using this approach, we introduce a new static checker called S tack that precisely identifies undefined behavior bugs. Applying S tack to widely used systems has uncovered 161 new bugs that have been confirmed and fixed by developers.

We present a new technique for parameter synthesis under boolean and quantitative objectives. The input to the technique is a "sketch" --- a program with missing numerical parameters --- and a probabilistic assumption about the program's inputs. The goal is to automatically synthesize values for the parameters such that the resulting program satisfies: (1) a {boolean specification}, which states that the program must meet certain assertions, and (2) a {quantitative specification}, which assigns a real valued rating to every program and which the synthesizer is expected to optimize. Our method --- called smoothed proof search --- reduces this task to a sequence of unconstrained smooth optimization problems that are then solved numerically. By iteratively solving these problems, we obtain parameter values that get closer and closer to meeting the boolean specification; at the limit, we obtain values that provably meet the specification. The approximations are computed using a new notion of smoothing for program abstractions, where an abstract transformer is approximated by a function that is continuous according to a metric over abstract states. We present a prototype implementation of our synthesis procedure, and experimental results on two benchmarks from the embedded control domain. The experiments demonstrate the benefits of smoothed proof search over an approach that does not meet the boolean and quantitative synthesis goals simultaneously.

This paper demonstrates how ideas from generative programming and software synthesis can help support the development of bulk-synchronous distributed memory kernels. These ideas are realized in a new language called MSL, a C-like language that combines synthesis features with high level notations for array manipulation and bulk-synchronous parallelism to simplify the semantic analysis required for synthesis. The paper shows that by leveraging these high level notations, it is possible to scale the synthesis and automated bug-finding technologies that underlie MSL to realistic computational kernels. Specifically, we demonstrate the methodology through case studies implementing non-trivial distributed kernels -- both regular and irregular -- from the NAS parallel benchmarks. We show that our approach can automatically infer many challenging details from these benchmarks and can enable high level implementation ideas to be reused between similar kernels. We also demonstrate that these high level notations map easily to low level C code and show that the performance of this generated code matches that of handwritten Fortran.

Many web applications store persistent data in databases. During execution, such applications spend a significant amount of time communicating with the database for retrieval and storing of persistent data over the network. These network round trips represent a significant fraction of the overall execution time for many applications and as a result increase their latency. While there has been prior work that aims to eliminate round trips by batching queries, they are limited by 1) a requirement that developers manually identify batching opportunities, or 2) the fact that they employ static program analysis techniques that cannot exploit many opportunities for batching. In this paper, we present Sloth, a new system that extends traditional lazy evaluation to expose query batching opportunities during application execution, even across loops, branches, and method boundaries. We evaluated Sloth using over 100 benchmarks from two large-scale open-source applications, and achieved up to a 3x reduction in page load time by delaying computation.

Object-relational mapping libraries are a popular way for applications to interact with databases because they provide transparent access to the database using the same language as the application. Unfortunately, using such frameworks often leads to poor performance, as modularity concerns encourage developers to implement relational operations in application code. Such application code does not take advantage of the optimized relational implementations that database systems provide, such as efficient implementations of joins or push down of selection predicates. In this paper we present QBS, a system that automatically transforms fragments of application logic into SQL queries. QBS differs from traditional compiler optimizations as it relies on synthesis technology to generate invariants and postconditions for a code fragment. The postconditions and invariants are expressed using a new theory of ordered relations that allows us to reason precisely about both the contents and order of the records produced complex code fragments that compute joins and aggregates. The theory is close in expressiveness to SQL, so the synthesized postconditions can be readily translated to SQL queries. Using 75 code fragments automatically extracted from over 120k lines of open-source code written using the Java Hibernate ORM, we demonstrate that our approach can convert a variety of imperative constructs into relational specifications and significantly improve application performance asymptotically by orders of magnitude.

We present a new method for automatically providing feedback for introductory programming problems. In order to use this method, we need a reference implementation of the assignment, and an error model consisting of potential corrections to errors that students might make. Using this information, the system automatically derives minimal corrections to student's incorrect solutions, providing them with a measure of exactly how incorrect a given solution was, as well as feedback about what they did wrong.

This paper studies an emerging class of software bugs called optimization-unstable code: code that is unexpectedly discarded by compiler optimizations due to undefined behavior in the program. Unstable code is present in many systems, including the Linux kernel and the Postgres database. The consequences of unstable code range from incorrect functionality to missing security checks.

This paper presents a new approach to select events of interest to users in a social media setting where events are generated from mobile devices. We argue that the problem is best solved by inductive learning, where the goal is to first generalize from the users' expressed "likes" and "dislikes" of specific events, then to produce a program that can be used to collect only data of interest.

It is becoming increasingly important for applications to protect sensitive data. With current techniques, the programmer bears the burden of ensuring that the application's behavior adheres to policies about where sensitive values may flow. Unfortunately, privacy policies are difficult to manage because their global nature requires coordinated reasoning and enforcement. To address this problem, we describe a programming model that makes the system responsible for ensuring adherence to privacy policies. The programming model has two components: 1) core programs describing functionality independent of privacy concerns and 2) declarative, decentralized policies controlling how sensitive values are disclosed. Each sensitive value encapsulates multiple views; policies describe which views are allowed based on the output context. The system is responsible for automatically ensuring that outputs are consistent with the policies. We have implemented this programming model in a new functional constraint language named Jeeves. In Jeeves, sensitive values are introduced as symbolic variables and policies correspond to constraints that are resolved at output channels. We have implemented Jeeves as a Scala library using an SMT solver as a model finder. In this paper we describe the dynamic and static semantics of Jeeves and the properties about policy enforcement that the semantics guarantees. We also describe our experience implementing a conference management system and a social network.

Software construction today often involves the use of large frameworks. The challenge in this type of programming is that object-oriented frameworks tend to grow exceedingly intricate; they spread functionality among numerous classes, and any use of the framework requires knowledge of many interacting components. We present a system named MATCHMAKER that from a simple query synthesizes code that interacts with the framework. The query consists of names of two framework classes, and our system produces code enabling interaction between them. MATCHMAKER relies on a database of dynamic program traces called DELIGHT that uses novel abstraction-based indexing techniques to answer queries about the evolution of heap connectivity in a matter of seconds.

Bugs in deployed software can be extremely difficult to track down. Invasive logging techniques, such as logging all non-deterministic inputs, can incur substantial runtime overheads. This paper shows how symbolic analysis can be used to re-create path equivalent executions for very long running programs such as databases and web servers. The goal is to help developers debug such long-running programs by allowing them to walk through an execution of the last few requests or transactions leading up to an error. The challenge is to provide this functionality without the high runtime overheads associated with traditional replay techniques based on input logging or memory snapshots. Our approach achieves this by recording a small amount of information about program execution, such as the direction of branches taken, and then using symbolic analysis to reconstruct the execution of the last few inputs processed by the application, as well as the state of memory before these inputs were executed.

We present the Storyboard Programming framework, a new synthesis system designed to help programmers write imperative low-level data-structure manipulations. The goal of this system is to bridge the gap between the "boxes-and-arrows" diagrams that programmers often use to think about data-structure manipulation algorithms and the low-level imperative code that implements them. The system takes as input a set of partial input-output examples, as well as a description of the high-level structure of the desired solution. From this information, it is able to synthesize low-level imperative implementations in a matter of minutes.

We present smooth interpretation, a method to systematically approximate numerical imperative programs by smooth mathematical functions. This approximation facilitates the use of numerical search techniques like gradient descent for program analysis and synthesis. The method extends to programs the notion of Gaussian smoothing, a popular signal-processing technique that filters out noise and discontinuities from a signal by taking its convolution with a Gaussian function.

We describe PSketch, a program synthesizer that helps programmers implement concurrent data structures. The system is based on the concept of sketching, a form of synthesis that allows programmers to express their insight about an implementation as a partial program: a sketch. The synthesizer automatically completes the sketch to produce an implementation that matches a given correctness criteria.

Performance of stencil computations can be significantly improved through smart implementations that improve memory locality, computation reuse, or parallelize the computation. Unfortunately, efficient implementations are hard to obtain because they often involve non-traditional transformations, which means that they cannot be produced by optimizing the reference stencil with a compiler. In fact, many stencils are produced by code generators that were tediously handcrafted.

Sketching is a software synthesis approach where the programmer develops a partial implementation - a sketch - and a separate specification of the desired functionality. The synthesizer then completes the sketch to behave like the specification. The correctness of the synthesized implementation is guaranteed by the compiler, which allows, among other benefits, rapid development of highly tuned implementations without the fear of introducing bugs.We develop SKETCH, a language for finite programs with linguistic support for sketching. Finite programs include many highperformance kernels, including cryptocodes. In contrast to prior synthesizers, which had to be equipped with domain-specific rules, SKETCH completes sketches by means of a combinatorial search based on generalized boolean satisfiability. Consequently, our combinatorial synthesizer is complete for the class of finite programs: it is guaranteed to complete any sketch in theory, and in practice has scaled to realistic programming problems.Freed from domain rules, we can now write sketches as simpleto-understand partial programs, which are regular programs in which difficult code fragments are replaced with holes to be filled by the synthesizer. Holes may stand for index expressions, lookup tables, or bitmasks, but the programmer can easily define new kinds of holes using a single versatile synthesis operator.We have used SKETCH to synthesize an efficient implementation of the AES cipher standard. The synthesizer produces the most complex part of the implementation and runs in about an hour.