Search

Among approaches for provably safe reinforcement learning, Model Predictive Shielding (MPS) has proven effective at complex tasks in continuous, high-dimensional state spaces, by leveraging a backup policy to ensure safety when the learned policy attempts to take risky actions. However, while MPS can ensure safety both during and after training, it often hinders task progress due to the conservative and task-oblivious nature of backup policies. This paper introduces Dynamic Model Predictive Shielding (DMPS), which optimizes reinforcement learning objectives while maintaining provable safety. DMPS employs a local planner to dynamically select safe recovery actions that maximize both short-term progress as well as long-term rewards. Crucially, the planner and the neural policy play a synergistic role in DMPS. When planning recovery actions for ensuring safety, the planner utilizes the neural policy to estimate long-term rewards, allowing it to observe beyond its short-term planning horizon. Conversely, the neural policy under training learns from the recovery plans proposed by the planner, converging to policies that are both high-performing and safe in practice. This approach guarantees safety during and after training, with bounded recovery regret that decreases exponentially with planning horizon depth. Experimental results demonstrate that DMPS converges to policies that rarely require shield interventions after training and achieve higher rewards compared to several state-of-the-art baselines.

Online streaming algorithms, tailored for continuous data processing, offer substantial benefits but are often more intricate to design than their offline counterparts. This paper introduces a novel approach for automatically synthesizing online streaming algorithms from their offline versions. In particular, we propose a novel methodology, based on the notion of relational function signature (RFS), for deriving an online algorithm given its offline version. Then, we propose a concrete synthesis algorithm that is an instantiation of the proposed methodology. Our algorithm uses the RFS to decompose the synthesis problem into a set of independent subtasks and uses a combination of symbolic reasoning and search to solve each subproblem. We implement the proposed technique in a new tool called Opera and evaluate it on over 50 tasks spanning two domains: statistical computations and online auctions. Our results show that Opera can automatically derive the online version of the original algorithm for 98% of the tasks. Our experiments also demonstrate that Opera significantly outperforms alternative approaches, including adaptations of SyGuS solvers to this problem as well as two of Opera's own ablations.

This paper addresses the problem of preference learning, which aims to learn user-specific preferences (e.g., "good parking spot", "convenient drop-off location") from visual input. Despite its similarity to learning factual concepts (e.g., "red cube"), preference learning is a fundamentally harder problem due to its subjective nature and the paucity of person-specific training data. We address this problem using a new framework called Synapse, which is a neuro-symbolic approach designed to efficiently learn preferential concepts from limited demonstrations. Synapse represents preferences as neuro-symbolic programs in a domain-specific language (DSL) that operates over images, and leverages a novel combination of visual parsing, large language models, and program synthesis to learn programs representing individual preferences. We evaluate Synapse through extensive experimentation including a user case study focusing on mobility-related concepts in mobile robotics and autonomous driving. Our evaluation demonstrates that Synapse significantly outperforms existing baselines as well as its own ablations. The code and other details can be found on the project website https://amrl.cs.utexas.edu/synapse ., Comment: 25 pages, 7 tables, 9 figures; Preprint; Updated figures and appendix, added VLM ablations

Due to the availability of increasingly large amounts of visual data, there is a growing need for tools that can help users find relevant images. While existing tools can perform image retrieval based on similarity or metadata, they fall short in scenarios that necessitate semantic reasoning about the content of the image. This paper explores a new multi-modal image search approach that allows users to conveniently specify and perform semantic image search tasks. With our tool, PhotoScout, the user interactively provides natural language descriptions, positive and negative examples, and object tags to specify their search tasks. Under the hood, PhotoScout is powered by a program synthesis engine that generates visual queries in a domain-specific language and executes the synthesized program to retrieve the desired images. In a study with 25 participants, we observed that PhotoScout allows users to perform image retrieval tasks more accurately and with less manual effort.

This paper lays down the research agenda for a domain-specific foundation model for operating systems (OSes). Our case for a foundation model revolves around the observations that several OS components such as CPU, memory, and network subsystems are interrelated and that OS traces offer the ideal dataset for a foundation model to grasp the intricacies of diverse OS components and their behavior in varying environments and workloads. We discuss a wide range of possibilities that then arise, from employing foundation models as policy agents to utilizing them as generators and predictors to assist traditional OS control algorithms. Our hope is that this paper spurs further research into OS foundation models and creating the next generation of operating systems for the evolving computing landscape., Comment: Machine Learning for Systems Workshop at 37th NeurIPS Conference, 2023, New Orleans, LA, USA

Developers often dedicate significant time to maintaining and refactoring existing code. However, most prior work on generative models for code focuses solely on creating new code, overlooking the distinctive needs of editing existing code. In this work, we explore a multi-round code auto-editing setting, aiming to predict edits to a code region based on recent changes within the same codebase. Our model, Coeditor, is a fine-tuned language model specifically designed for code editing tasks. We represent code changes using a line diff format and employ static analysis to form large customized model contexts, ensuring the availability of appropriate information for prediction. We collect a code editing dataset from the commit histories of 1650 open-source Python projects for training and evaluation. In a simplified single-round, single-edit task, Coeditor significantly outperforms GPT-3.5 and SOTA open-source code completion models (bringing exact-match accuracy from 34.7 up to 60.4), demonstrating the benefits of incorporating editing history for code completion. In a multi-round, multi-edit setting, we observe substantial gains by iteratively conditioning on additional user edits. We have open-sourced our code, data, and model weights to encourage future research and have released a VSCode extension powered by our model for interactive IDE usage., Comment: The Twelfth International Conference on Learning Representations (2024)

Many data extraction tasks of practical relevance require not only syntactic pattern matching but also semantic reasoning about the content of the underlying text. While regular expressions are very well suited for tasks that require only syntactic pattern matching, they fall short for data extraction tasks that involve both a syntactic and semantic component. To address this issue, we introduce semantic regexes, a generalization of regular expressions that facilitates combined syntactic and semantic reasoning about textual data. We also propose a novel learning algorithm that can synthesize semantic regexes from a small number of positive and negative examples. Our proposed learning algorithm uses a combination of neural sketch generation and compositional type-directed synthesis for fast and effective generalization from a small number of examples. We have implemented these ideas in a new tool called Smore and evaluated it on representative data extraction tasks involving several textual datasets. Our evaluation shows that semantic regexes can better support complex data extraction tasks than standard regular expressions and that our learning algorithm significantly outperforms existing tools, including state-of-the-art neural networks and program synthesis tools.

Prior work has combined chain-of-thought prompting in large language models (LLMs) with programmatic representations to perform effective and transparent reasoning. While such an approach works well for tasks that only require forward reasoning (e.g., straightforward arithmetic), it is less effective for constraint solving problems that require more sophisticated planning and search. In this paper, we propose a new satisfiability-aided language modeling (SatLM) approach for improving the reasoning capabilities of LLMs. We use an LLM to generate a declarative task specification rather than an imperative program and leverage an off-the-shelf automated theorem prover to derive the final answer. This approach has two key advantages. The declarative specification is closer to the problem description than the reasoning steps are, so the LLM can parse it out of the description more accurately. Furthermore, by offloading the actual reasoning task to an automated theorem prover, our approach can guarantee the correctness of the answer with respect to the parsed specification and avoid planning errors in the solving process. We evaluate SATLM on 8 different datasets and show that it consistently outperforms program-aided LMs in the imperative paradigm. In particular, SATLM outperforms program-aided LMs by 23% on a challenging subset of the GSM arithmetic reasoning dataset; SATLM also achieves a new SoTA on LSAT and BoardgameQA, surpassing previous models that are trained on the respective training sets., Comment: NeurIPS 2023

The goal of programmatic Learning from Demonstration (LfD) is to learn a policy in a programming language that can be used to control a robot's behavior from a set of user demonstrations. This paper presents a new programmatic LfD algorithm that targets long-horizon robot tasks which require synthesizing programs with complex control flow structures, including nested loops with multiple conditionals. Our proposed method first learns a program sketch that captures the target program's control flow and then completes this sketch using an LLM-guided search procedure that incorporates a novel technique for proving unrealizability of programming-by-demonstration problems. We have implemented our approach in a new tool called PROLEX and present the results of a comprehensive experimental evaluation on 120 benchmarks involving complex tasks and environments. We show that, given a 120 second time limit, PROLEX can find a program consistent with the demonstrations in 80% of the cases. Furthermore, for 81% of the tasks for which a solution is returned, PROLEX is able to find the ground truth program with just one demonstration. In comparison, CVC5, a syntax guided synthesis tool, is only able to solve 25% of the cases even when given the ground truth program sketch, and an LLM-based approach, GPT-Synth, is unable to solve any of the tasks due to the environment complexity., Comment: 40 Pages, Extended Version of POPL 2024 paper

Zero-knowledge (ZK) proof systems have emerged as a promising solution for building security-sensitive applications. However, bugs in ZK applications are extremely difficult to detect and can allow a malicious party to silently exploit the system without leaving any observable trace. This paper presents Coda, a novel statically-typed language for building zero-knowledge applications. Critically, Coda makes it possible to formally specify and statically check properties of a ZK application through a rich refinement type system. One of the key challenges in formally verifying ZK applications is that they require reasoning about polynomial equations over large prime fields that go beyond the capabilities of automated theorem provers. Coda mitigates this challenge by generating a set of Coq lemmas that can be proven in an interactive manner with the help of a tactic library. We have used Coda to re-implement 79 arithmetic circuits from widely-used Circom libraries and applications. Our evaluation shows that Coda makes it possible to specify important and formally verify correctness properties of these circuits. Our evaluation also revealed 6 previously-unknown vulnerabilities in the original Circom projects., Comment: This paper was incorrectly submitted, and should be submitted to Cryptology ePrint Archive instead

This paper presents a new synthesis-based approach for batch image processing. Unlike existing tools that can only apply global edits to the entire image, our method can apply fine-grained edits to individual objects within the image. For example, our method can selectively blur or crop specific objects that have a certain property. To facilitate such fine-grained image editing tasks, we propose a neuro-symbolic domain-specific language (DSL) that combines pre-trained neural networks for image classification with other language constructs that enable symbolic reasoning. Our method can automatically learn programs in this DSL from user demonstrations by utilizing a novel synthesis algorithm. We have implemented the proposed technique in a tool called ImageEye and evaluated it on 50 image editing tasks. Our evaluation shows that ImageEye is able to automate 96% of these tasks.

There has been growing interest in automatically predicting missing type annotations in programs written in Python and JavaScript. While prior methods have achieved impressive accuracy when predicting the most common types, they often perform poorly on rare or complex types. In this paper, we present a new type inference method that treats type prediction as a code infilling task by leveraging CodeT5, a state-of-the-art seq2seq pre-trained language model for code. Our method uses static analysis to construct dynamic contexts for each code element whose type signature is to be predicted by the model. We also propose an iterative decoding scheme that incorporates previous type predictions in the model's input context, allowing information exchange between related code elements. Our evaluation shows that the proposed approach, TypeT5, not only achieves a higher overall accuracy (particularly on rare and complex types) but also produces more coherent results with fewer type errors -- while enabling easy user intervention., Comment: Published as a conference paper at ICLR 2023

Imitation Learning (IL) is a promising paradigm for teaching robots to perform novel tasks using demonstrations. Most existing approaches for IL utilize neural networks (NN), however, these methods suffer from several well-known limitations: they 1) require large amounts of training data, 2) are hard to interpret, and 3) are hard to repair and adapt. There is an emerging interest in programmatic imitation learning (PIL), which offers significant promise in addressing the above limitations. In PIL, the learned policy is represented in a programming language, making it amenable to interpretation and repair. However, state-of-the-art PIL algorithms assume access to action labels and struggle to learn from noisy real-world demonstrations. In this paper, we propose PLUNDER, a novel PIL algorithm that integrates a probabilistic program synthesizer in an iterative Expectation-Maximization (EM) framework to address these shortcomings. Unlike existing PIL approaches, PLUNDER synthesizes probabilistic programmatic policies that are particularly well-suited for modeling the uncertainties inherent in real-world demonstrations. Our approach leverages an EM loop to simultaneously infer the missing action labels and the most likely probabilistic policy. We benchmark PLUNDER against several established IL techniques, and demonstrate its superiority across five challenging imitation learning tasks under noise. PLUNDER policies achieve 95% accuracy in matching the given demonstrations, outperforming the next best baseline by 19%. Additionally, policies generated by PLUNDER successfully complete the tasks 17% more frequently than the nearest baseline.

In reinforcement learning for safety-critical settings, it is often desirable for the agent to obey safety constraints at all points in time, including during training. We present a novel neurosymbolic approach called SPICE to solve this safe exploration problem. SPICE uses an online shielding layer based on symbolic weakest preconditions to achieve a more precise safety analysis than existing tools without unduly impacting the training process. We evaluate the approach on a suite of continuous control benchmarks and show that it can achieve comparable performance to existing safe learning techniques while incurring fewer safety violations. Additionally, we present theoretical results showing that SPICE converges to the optimal safe policy under reasonable assumptions.

We propose a new technique based on program synthesis for automatically generating visualizations from natural language queries. Our method parses the natural language query into a refinement type specification using the intents-and-slots paradigm and leverages type-directed synthesis to generate a set of visualization programs that are most likely to meet the user's intent. Our refinement type system captures useful hints present in the natural language query and allows the synthesis algorithm to reject visualizations that violate well-established design guidelines for the input data set. We have implemented our ideas in a tool called Graphy and evaluated it on NLVCorpus, which consists of 3 popular datasets and over 700 real-world natural language queries. Our experiments show that Graphy significantly outperforms state-of-the-art natural-language-based visualization tools, including transformer and rule-based ones., Comment: 39 pages

We present a new domain-agnostic synthesis technique for generating programs from input-output examples. Our method, called metric program synthesis, relaxes the well-known 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 a 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.

We present a study of crash-consistency bugs in persistent-memory (PM) file systems and analyze their implications for file-system design and testing crash consistency. We develop FlyTrap, a framework to test PM file systems for crash-consistency bugs. FlyTrap discovered 18 new bugs across four PM file systems; the bugs have been confirmed by developers and many have been already fixed. The discovered bugs have serious consequences such as breaking the atomicity of rename or making the file system unmountable. We present a detailed study of the bugs we found and discuss some important lessons from these observations. For instance, one of our findings is that many of the bugs are due to logic errors, rather than errors in using flushes or fences; this has important applications for future work on testing PM file systems. Another key finding is that many bugs arise from attempts to improve efficiency by performing metadata updates in-place and that recovery code that deals with rebuilding in-DRAM state is a significant source of bugs. These observations have important implications for designing and testing PM file systems. Our code is available at https://github.com/utsaslab/flytrap ., Comment: 16 pages

While many mainstream languages such as Java, Python, and C# increasingly incorporate functional APIs to simplify programming and improve parallelization/performance, there are no effective techniques that can be used to automatically translate existing imperative code to functional variants using these APIs. Motivated by this problem, this paper presents a transpilation approach based on inductive program synthesis for modernizing existing code. Our method is based on the observation that the overwhelming majority of source/target programs in this setting satisfy an assumption that we call trace-compatibility: not only do the programs share syntactically identical low-level expressions, but these expressions also take the same values in corresponding execution traces. Our method leverages this observation to design a new neural-guided synthesis algorithm that (1) uses a novel neural architecture called cognate grammar network (CGN) and (2) leverages a form of concolic execution to prune partial programs based on intermediate values that arise during a computation. We have implemented our approach in a tool called NGST2 and use it to translate imperative Java and Python code to functional variants that use the Stream and functools APIs respectively. Our experiments show that NGST2 significantly outperforms several baselines and that our proposed neural architecture and pruning techniques are vital for achieving good results., Comment: Fixed some incorrectly rendered citations

Accurate kinodynamic models play a crucial role in many robotics applications such as off-road navigation and high-speed driving. Many state-of-the-art approaches in learning stochastic kinodynamic models, however, require precise measurements of robot states as labeled input/output examples, which can be hard to obtain in outdoor settings due to limited sensor capabilities and the absence of ground truth. In this work, we propose a new technique for learning neural stochastic kinodynamic models from noisy and indirect observations by performing simultaneous state estimation and dynamics learning. The proposed technique iteratively improves the kinodynamic model in an expectation-maximization loop, where the E Step samples posterior state trajectories using particle filtering, and the M Step updates the dynamics to be more consistent with the sampled trajectories via stochastic gradient ascent. We evaluate our approach on both simulation and real-world benchmarks and compare it with several baseline techniques. Our approach not only achieves significantly higher accuracy but is also more robust to observation noise, thereby showing promise for boosting the performance of many other robotics applications., Comment: Accepted for publication in the Proceedings of IROS 2022

A monitor is a widely-used concurrent programming abstraction that encapsulates all shared state between threads. Monitors can be classified as being either implicit or explicit depending on the primitives they provide. Implicit monitors are much easier to program but typically not as efficient. To address this gap, there has been recent research on automatically synthesizing explicit-signal monitors from an implicit specification, but prior work does not exploit all paralellization opportunities due to the use of a single lock for the entire monitor. This paper presents a new technique for synthesizing fine-grained explicit-synchronization protocols from implicit monitors. Our method is based on two key innovations: First, we present a new static analysis for inferring safe interleavings that allow violating mutual exclusion of monitor operations without changing its semantics. Second, we use the results of this static analysis to generate a MaxSAT instance whose models correspond to correct-by-construction synchronization protocols. We have implemented our approach in a tool called Cortado and evaluate it on monitors that contain parallelization opportunities. Our evaluation shows that Cortado can synthesize synchronization policies that are competitive with, or even better than, expert-written ones on these benchmarks., Comment: Change title to include "(Extended Version)"

Machine learning (ML) models play an increasingly prevalent role in many software engineering tasks. However, because most models are now powered by opaque deep neural networks, it can be difficult for developers to understand why the model came to a certain conclusion and how to act upon the model's prediction. Motivated by this problem, this paper explores counterfactual explanations for models of source code. Such counterfactual explanations constitute minimal changes to the source code under which the model "changes its mind". We integrate counterfactual explanation generation to models of source code in a real-world setting. We describe considerations that impact both the ability to find realistic and plausible counterfactual explanations, as well as the usefulness of such explanation to the user of the model. In a series of experiments we investigate the efficacy of our approach on three different models, each based on a BERT-like architecture operating over source code., Comment: 10 pages, 6 listings, 2 algorithms, 2 tables, 1 figure

As smart contracts gain adoption in financial transactions, it becomes increasingly important to ensure that they are free of bugs and security vulnerabilities. Of particular relevance in this context are arithmetic overflow bugs, as integers are often used to represent financial assets like account balances. Motivated by this observation, this paper presents SolType, a refinement type system for Solidity that can be used to prevent arithmetic over- and under-flows in smart contracts. SolType allows developers to add refinement type annotations and uses them to prove that arithmetic operations do not lead to over- and under-flows. SolType incorporates a rich vocabulary of refinement terms that allow expressing relationships between integer values and aggregate properties of complex data structures. Furthermore, our implementation, called Solid, incorporates a type inference engine and can automatically infer useful type annotations, including non-trivial contract invariants. To evaluate the usefulness of our type system, we use Solid to prove arithmetic safety of a total of 120 smart contracts. When used in its fully automated mode (i.e., using Solid's type inference capabilities), Solid is able to eliminate 86.3% of redundant runtime checks used to guard against overflows. We also compare Solid against a state-of-the-art arithmetic safety verifier called VeriSmart and show that Solid has a significantly lower false positive rate, while being significantly faster in terms of verification time., Comment: To appear in POPL '22. This is the extended version of the paper with the proofs, after the main text went through peer review. 51 pages, 15 figures

Understanding the semantics of tables at scale is crucial for tasks like data integration, preparation, and search. Table understanding methods aim at detecting a table's topic, semantic column types, column relations, or entities. With the rise of deep learning, powerful models have been developed for these tasks with excellent accuracy on benchmarks. However, we observe that there exists a gap between the performance of these models on these benchmarks and their applicability in practice. In this paper, we address the question: what do we need for these models to work in practice? We discuss three challenges of deploying table understanding models and propose a framework to address them. These challenges include 1) difficulty in customizing models to specific domains, 2) lack of training data for typical database tables often found in enterprises, and 3) lack of confidence in the inferences made by models. We present SigmaTyper which implements this framework for the semantic column type detection task. SigmaTyper encapsulates a hybrid model trained on GitTables and integrates a lightweight human-in-the-loop approach to customize the model. Lastly, we highlight avenues for future research that further close the gap towards making table understanding effective in practice., Comment: Submitted to CIDR'22

We present a novel bottom-up method for the synthesis of functional recursive programs. While bottom-up synthesis techniques can work better than top-down methods in certain settings, there is no prior technique for synthesizing recursive programs from logical specifications in a purely bottom-up fashion. The main challenge is that effective bottom-up methods need to execute sub-expressions of the code being synthesized, but it is impossible to execute a recursive subexpression of a program that has not been fully constructed yet. In this paper, we address this challenge using the concept of angelic semantics. Specifically, our method finds a program that satisfies the specification under angelic semantics (we refer to this as angelic synthesis), analyzes the assumptions made during its angelic execution, uses this analysis to strengthen the specification, and finally reattempts synthesis with the strengthened specification. Our proposed angelic synthesis algorithm is based on version space learning and therefore deals effectively with many incremental synthesis calls made during the overall algorithm. We have implemented this approach in a prototype called Burst and evaluate it on synthesis problems from prior work. Our experiments show that Burst is able to synthesize a solution to 94% of the benchmarks in our benchmark suite, outperforming prior work.

In this paper, we propose a new technique based on program synthesis for extracting information from webpages. Given a natural language query and a few labeled webpages, our method synthesizes a program that can be used to extract similar types of information from other unlabeled webpages. To handle websites with diverse structure, our approach employs a neurosymbolic DSL that incorporates both neural NLP models as well as standard language constructs for tree navigation and string manipulation. We also propose an optimal synthesis algorithm that generates all DSL programs that achieve optimal F1 score on the training examples. Our synthesis technique is compositional, prunes the search space by exploiting a monotonicity property of the DSL, and uses transductive learning to select programs with good generalization power. We have implemented these ideas in a new tool called WebQA and evaluate it on 25 different tasks across multiple domains. Our experiments show that WebQA significantly outperforms existing tools such as state-of-the-art question answering models and wrapper induction systems., Comment: 20 pages

We propose a new algorithm to simplify the controller development for distributed robotic systems subject to external observations, disturbances, and communication delays. Unlike prior approaches that propose specialized solutions to handling communication latency for specific robotic applications, our algorithm uses an arbitrary centralized controller as the specification and automatically generates distributed controllers with communication management and delay compensation. We formulate our goal as nonlinear optimal control -- using a regret minimizing objective that measures how much the distributed agents behave differently from the delay-free centralized response -- and solve for optimal actions w.r.t. local estimations of this objective using gradient-based optimization. We analyze our proposed algorithm's behavior under a linear time-invariant special case and prove that the closed-loop dynamics satisfy a form of input-to-state stability w.r.t. unexpected disturbances and observations. Our experimental results on both simulated and real-world robotic tasks demonstrate the practical usefulness of our approach and show significant improvement over several baseline approaches., Comment: 8 pages, 4 figures

Modern visualization tools aim to allow data analysts to easily create exploratory visualizations. When the input data layout conforms to the visualization design, users can easily specify visualizations by mapping data columns to visual channels of the design. However, when there is a mismatch between data layout and the design, users need to spend significant effort on data transformation. We propose Falx, a synthesis-powered visualization tool that allows users to specify visualizations in a similarly simple way but without needing to worry about data layout. In Falx, users specify visualizations using examples of how concrete values in the input are mapped to visual channels, and Falx automatically infers the visualization specification and transforms the data to match the design. In a study with 33 data analysts on four visualization tasks involving data transformation, we found that users can effectively adopt Falx to create visualizations they otherwise cannot implement., Comment: CHI 2021

Several real-world libraries (e.g., reentrant locks, GUI frameworks, serialization libraries) require their clients to use the provided API in a manner that conforms to a context-free specification. Motivated by this observation, this paper describes a new technique for verifying the correct usage of context-free API protocols. The key idea underlying our technique is to over-approximate the program's feasible API call sequences using a context-free grammar (CFG) and then check language inclusion between this grammar and the specification. However, since this inclusion check may fail due to imprecision in the program's CFG abstraction, we propose a novel refinement technique to progressively improve the CFG. In particular, our method obtains counterexamples from CFG inclusion queries and uses them to introduce new non-terminals and productions to the grammar while still over-approximating the program's relevant behavior. We have implemented the proposed algorithm in a tool called CFPChecker and evaluate it on 10 popular Java applications that use at least one API with a context-free specification. Our evaluation shows that CFPChecker is able to verify correct usage of the API in clients that use it correctly and produces counterexamples for those that do not. We also compare our method against three relevant baselines and demonstrate that CFPChecker enables verification of safety properties that are beyond the reach of existing tools.

Multimodal program synthesis, which leverages different types of user input to synthesize a desired program, is an attractive way to scale program synthesis to challenging settings; however, it requires integrating noisy signals from the user, like natural language, with hard constraints on the program's behavior. This paper proposes an optimal neural synthesis approach where the goal is to find a program that satisfies user-provided constraints while also maximizing the program's score with respect to a neural model. Specifically, we focus on multimodal synthesis tasks in which the user intent is expressed using a combination of natural language (NL) and input-output examples. At the core of our method is a top-down recurrent neural model that places distributions over abstract syntax trees conditioned on the NL input. This model not only allows for efficient search over the space of syntactically valid programs, but it allows us to leverage automated program analysis techniques for pruning the search space based on infeasibility of partial programs with respect to the user's constraints. The experimental results on a multimodal synthesis dataset (StructuredRegex) show that our method substantially outperforms prior state-of-the-art techniques in terms of accuracy and efficiency, and finds model-optimal programs more frequently., Comment: Findings of EMNLP 2021

We present Revel, a partially neural reinforcement learning (RL) framework for provably safe exploration in continuous state and action spaces. A key challenge for provably safe deep RL is that repeatedly verifying neural networks within a learning loop is computationally infeasible. We address this challenge using two policy classes: a general, neurosymbolic class with approximate gradients and a more restricted class of symbolic policies that allows efficient verification. Our learning algorithm is a mirror descent over policies: in each iteration, it safely lifts a symbolic policy into the neurosymbolic space, performs safe gradient updates to the resulting policy, and projects the updated policy into the safe symbolic subset, all without requiring explicit verification of neural networks. Our empirical results show that Revel enforces safe exploration in many scenarios in which Constrained Policy Optimization does not, and that it can discover policies that outperform those learned through prior approaches to verified exploration.

Existing datasets for regular expression (regex) generation from natural language are limited in complexity; compared to regex tasks that users post on StackOverflow, the regexes in these datasets are simple, and the language used to describe them is not diverse. We introduce StructuredRegex, a new regex synthesis dataset differing from prior ones in three aspects. First, to obtain structurally complex and realistic regexes, we generate the regexes using a probabilistic grammar with pre-defined macros observed from real-world StackOverflow posts. Second, to obtain linguistically diverse natural language descriptions, we show crowdworkers abstract depictions of the underlying regex and ask them to describe the pattern they see, rather than having them paraphrase synthetic language. Third, we augment each regex example with a collection of strings that are and are not matched by the ground truth regex, similar to how real users give examples. Our quantitative and qualitative analysis demonstrates the advantages of StructuredRegex over prior datasets. Further experimental results using various multimodal synthesis techniques highlight the challenge presented by our dataset, including non-local constraints and multi-modal inputs., Comment: ACL2020

As gradual typing becomes increasingly popular in languages like Python and TypeScript, there is a growing need to infer type annotations automatically. While type annotations help with tasks like code completion and static error catching, these annotations cannot be fully determined by compilers and are tedious to annotate by hand. This paper proposes a probabilistic type inference scheme for TypeScript based on a graph neural network. Our approach first uses lightweight source code analysis to generate a program abstraction called a type dependency graph, which links type variables with logical constraints as well as name and usage information. Given this program abstraction, we then use a graph neural network to propagate information between related type variables and eventually make type predictions. Our neural architecture can predict both standard types, like number or string, as well as user-defined types that have not been encountered during training. Our experimental results show that our approach outperforms prior work in this space by $14\%$ (absolute) on library types, while having the ability to make type predictions that are out of scope for existing techniques., Comment: Accepted as a poster at ICLR 2020

This paper presents a new technique for migrating data between different schemas. Our method expresses the schema mapping as a Datalog program and automatically synthesizes a Datalog program from simple input-output examples to perform data migration. This approach can transform data between different types of schemas (e.g., relational-to-graph, document-to-relational) and performs synthesis efficiently by leveraging the semantics of Datalog. We implement the proposed technique as a tool called Dynamite and show its effectiveness by evaluating Dynamite on 28 realistic data migration scenarios.

While visualizations play a crucial role in gaining insights from data, generating useful visualizations from a complex dataset is far from an easy task. Besides understanding the functionality provided by existing visualization libraries, generating the desired visualization also requires reshaping and aggregating the underlying data as well as composing different visual elements to achieve the intended visual narrative. This paper aims to simplify visualization tasks by automatically synthesizing the required program from simple visual sketches provided by the user. Specifically, given an input data set and a visual sketch that demonstrates how to visualize a very small subset of this data, our technique automatically generates a program that can be used to visualize the entire data set. Automating visualization poses several challenges. First, because many visualization tasks require data wrangling in addition to generating plots, we need to decompose the end-to-end synthesis task into two separate sub-problems. Second, because the intermediate specification that results from the decomposition is necessarily imprecise, this makes the data wrangling task particularly challenging in our context. In this paper, we address these problems by developing a new compositional visualization-by-example technique that (a) decomposes the end-to-end task into two different synthesis problems over different DSLs and (b) leverages bi-directional program analysis to deal with the complexity that arises from having an imprecise intermediate specification. We implemented our visualization-by-example algorithm and evaluate it on 83 visualization tasks collected from on-line forums and tutorials. Viser can solve 84% of these benchmarks within a 600 second time limit, and, for those tasks that can be solved, the desired visualization is among the top-5 generated by Viser in 70% of the cases.

Recent systems for converting natural language descriptions into regular expressions (regexes) have achieved some success, but typically deal with short, formulaic text and can only produce simple regexes. Realworld regexes are complex, hard to describe with brief sentences, and sometimes require examples to fully convey the user's intent. We present a framework for regex synthesis in this setting where both natural language (NL) and examples are available. First, a semantic parser (either grammar-based or neural) maps the natural language description into an intermediate sketch, which is an incomplete regex containing holes to denote missing components. Then a program synthesizer searches over the regex space defined by the sketch and finds a regex that is consistent with the given string examples. Our semantic parser can be trained purely from weak supervision based on correctness of the synthesized regex, or it can leverage heuristically-derived sketches. We evaluate on two prior datasets (Kushman and Barzilay, 2013; Locascio et al., 2016) and a real-world dataset from Stack Overflow. Our system achieves state-of-the-art performance on the prior datasets and solves 57% of the real-world dataset, which existing neural systems completely fail on., Comment: TACL2020

In this paper, we propose a multi-modal synthesis technique for automatically constructing regular expressions (regexes) from a combination of examples and natural language. Using multiple modalities is useful in this context because natural language alone is often highly ambiguous, whereas examples in isolation are often not sufficient for conveying user intent. Our proposed technique first parses the English description into a so-called hierarchical sketch that guides our programming-by-example (PBE) engine. Since the hierarchical sketch captures crucial hints, the PBE engine can leverage this information to both prioritize the search as well as make useful deductions for pruning the search space. We have implemented the proposed technique in a tool called Regel and evaluate it on over three hundred regexes. Our evaluation shows that Regel achieves 80% accuracy whereas the NLP-only and PBE-only baselines achieve 43% and 26% respectively. We also compare our proposed PBE engine against an adaptation of AlphaRegex, a state-of-the-art regex synthesis tool, and show that our proposed PBE engine is an order of magnitude faster, even if we adapt the search algorithm of AlphaRegex to leverage the sketch. Finally, we conduct a user study involving 20 participants and show that users are twice as likely to successfully come up with the desired regex using Regel compared to without it., Comment: 36 pages

In recent years, the notion of local robustness (or robustness for short) has emerged as a desirable property of deep neural networks. Intuitively, robustness means that small perturbations to an input do not cause the network to perform misclassifications. In this paper, we present a novel algorithm for verifying robustness properties of neural networks. Our method synergistically combines gradient-based optimization methods for counterexample search with abstraction-based proof search to obtain a sound and ({\delta}-)complete decision procedure. Our method also employs a data-driven approach to learn a verification policy that guides abstract interpretation during proof search. We have implemented the proposed approach in a tool called Charon and experimentally evaluated it on hundreds of benchmarks. Our experiments show that the proposed approach significantly outperforms three state-of-the-art tools, namely AI^2 , Reluplex, and Reluval.

Many programs that interact with a database need to undergo schema refactoring several times during their life cycle. Since this process typically requires making significant changes to the program's implementation, schema refactoring is often non-trivial and error-prone. Motivated by this problem, we propose a new technique for automatically synthesizing a new version of a database program given its original version and the source and target schemas. Our method does not require manual user guidance and ensures that the synthesized program is equivalent to the original one. Furthermore, our method is quite efficient and can synthesize new versions of database programs (containing up to 263 functions) that are extracted from real-world web applications with an average synthesis time of 69.4 seconds.

Ensuring correctness of smart contracts is paramount to ensuring trust in blockchain-based systems. This paper studies the safety and security of smart contracts in the \emph{Azure Blockchain Workbench}, an enterprise Blockchain-as-a-Service offering from Microsoft. As part of this study, we formalize \emph{semantic conformance} of smart contracts against a state machine model with access-control policy and develop a highly-automated formal verifier for Solidity that can produce proofs as well as counterexamples. We have applied our verifier {\sc VeriSol} to analyze {\it all} contracts shipped with the Azure Blockchain Workbench, which includes application samples as well as a governance contract for Proof of Authority (PoA). We have found previously unknown bugs in these published smart contracts. After fixing these bugs, {\sc VeriSol} was able to successfully perform full verification for all of these contracts., Comment: 13 pages

This paper proposes relational program synthesis, a new problem that concerns synthesizing one or more programs that collectively satisfy a relational specification. As a dual of relational program verification, relational program synthesis is an important problem that has many practical applications, such as automated program inversion and automatic generation of comparators. However, this relational synthesis problem introduces new challenges over its non-relational counterpart due to the combinatorially larger search space. As a first step towards solving this problem, this paper presents a synthesis technique that combines the counterexample-guided inductive synthesis framework with a novel inductive synthesis algorithm that is based on relational version space learning. We have implemented the proposed technique in a framework called Relish, which can be instantiated to different application domains by providing a suitable domain-specific language and the relevant relational specification. We have used the Relish framework to build relational synthesizers to automatically generate string encoders/decoders as well as comparators, and we evaluate our tool on several benchmarks taken from prior work and online forums. Our experimental results show that the proposed technique can solve almost all of these benchmarks and that it significantly outperforms EUSolver, a generic synthesis framework that won the general track of the most recent SyGuS competition.

Many example-guided program synthesis techniques use abstractions to prune the search space. While abstraction-based synthesis has proven to be very powerful, a domain expert needs to provide a suitable abstract domain, together with the abstract transformers of each DSL construct. However, coming up with useful abstractions can be non-trivial, as it requires both domain expertise and knowledge about the synthesizer. In this paper, we propose a new technique for learning abstractions that are useful for instantiating a general synthesis framework in a new domain. Given a DSL and a small set of training problems, our method uses tree interpolation to infer reusable predicate templates that speed up synthesis in a given domain. Our method also learns suitable abstract transformers by solving a certain kind of second-order constraint solving problem in a data-driven way. We have implemented the proposed method in a tool called ATLAS and evaluate it in the context of the BLAZE meta-synthesizer. Our evaluation shows that (a) ATLAS can learn useful abstract domains and transformers from few training problems, and (b) the abstractions learned by ATLAS allow BLAZE to achieve significantly better results compared to manually-crafted abstractions.

Explicit signaling between threads is a perennial cause of bugs in concurrent programs. While there are several run-time techniques to automatically notify threads upon the availability of some shared resource, such techniques are not widely-adopted due to their run-time overhead. This paper proposes a new solution based on static analysis for automatically generating a performant explicit-signal program from its corresponding implicit-signal implementation. The key idea is to generate verification conditions that allow us to minimize the number of required signals and unnecessary context switches, while guaranteeing semantic equivalence between the source and target programs. We have implemented our method in a tool called Expresso and evaluate it on challenging benchmarks from prior papers and open-source software. Expresso-generated code significantly outperforms past automatic signaling mechanisms (avg. 1.56x speedup) and closely matches the performance of hand-optimized explicit-signal code.

Even though many programmers rely on 3-way merge tools to integrate changes from different branches, such tools can introduce subtle bugs in the integration process. This paper aims to mitigate this problem by defining a semantic notion of confict-freedom, which ensures that the merged program does not introduce new unwanted behaviors. We also show how to verify this property using a novel, compositional algorithm that combines lightweight dependence analysis for shared program fragments and precise relational reasoning for the modifications. We evaluate our tool called SafeMerge on 52 real-world merge scenarios obtained from Github and compare the results against a textual merge tool. The experimental results demonstrate the benefits of our approach over syntactic confict-freedom and indicate that SafeMerge is both precise and practical.

We propose a new conflict-driven program synthesis technique that is capable of learning from past mistakes. Given a spurious program that violates the desired specification, our synthesis algorithm identifies the root cause of the conflict and learns new lemmas that can prevent similar mistakes in the future. Specifically, we introduce the notion of equivalence modulo conflict and show how this idea can be used to learn useful lemmas that allow the synthesizer to prune large parts of the search space. We have implemented a general-purpose CDCL-style program synthesizer called Neo and evaluate it in two different application domains, namely data wrangling in R and functional programming over lists. Our experiments demonstrate the substantial benefits of conflict-driven learning and show that Neo outperforms two state-of-the-art synthesis tools, Morpheus and Deepcoder, that target these respective domains., Comment: This was a miscommunication and one author did not want to make it publicly available right now

While many applications export data in hierarchical formats like XML and JSON, it is often necessary to convert such hierarchical documents to a relational representation. This paper presents a novel programming-by-example approach, and its implementation in a tool called Mitra, for automatically migrating tree-structured documents to relational tables. We have evaluated the proposed technique using two sets of experiments. In the first experiment, we used Mitra to automate 98 data transformation tasks collected from StackOverflow. Our method can generate the desired program for 94% of these benchmarks with an average synthesis time of 3.8 seconds. In the second experiment, we used Mitra to generate programs that can convert real-world XML and JSON datasets to full-fledged relational databases. Our evaluation shows that Mitra can automate the desired transformation for all datasets.

We present a new approach to example-guided program synthesis based on counterexample-guided abstraction refinement. Our method uses the abstract semantics of the underlying DSL to find a program $P$ whose abstract behavior satisfies the examples. However, since program $P$ may be spurious with respect to the concrete semantics, our approach iteratively refines the abstraction until we either find a program that satisfies the examples or prove that no such DSL program exists. Because many programs have the same input-output behavior in terms of their abstract semantics, this synthesis methodology significantly reduces the search space compared to existing techniques that use purely concrete semantics. While synthesis using abstraction refinement (SYNGAR) could be implemented in different settings, we propose a refinement-based synthesis algorithm that uses abstract finite tree automata (AFTA). Our technique uses a coarse initial program abstraction to construct an initial AFTA, which is iteratively refined by constructing a proof of incorrectness of any spurious program. In addition to ruling out the spurious program accepted by the previous AFTA, proofs of incorrectness are also useful for ruling out many other spurious programs. We implement these ideas in a framework called \tool. We have used the BLAZE framework to build synthesizers for string and matrix transformations, and we compare BLAZE with existing techniques. Our results for the string domain show that BLAZE compares favorably with FlashFill, a domain-specific synthesizer that is now deployed in Microsoft PowerShell. In the context of matrix manipulations, we compare BLAZE against Prose, a state-of-the-art general-purpose VSA-based synthesizer, and show that BLAZE results in a 90x speed-up over Prose.

This paper addresses the problem of verifying equivalence between a pair of programs that operate over databases with different schemas. This problem is particularly important in the context of web applications, which typically undergo database refactoring either for performance or maintainability reasons. While web applications should have the same externally observable behavior before and after schema migration, there are no existing tools for proving equivalence of such programs. This paper takes a first step towards solving this problem by formalizing the equivalence and refinement checking problems for database-driven applications. We also propose a proof methodology based on the notion of bisimulation invariants over relational algebra with updates and describe a technique for synthesizing such bisimulation invariants. We have implemented the proposed technique in a tool called Mediator for verifying equivalence between database-driven applications written in our intermediate language and evaluate our tool on 21 benchmarks extracted from textbooks and real-world web applications. Our results show that the proposed methodology can successfully verify 20 of these benchmarks.