We address the problem of synthesizing code completions for programs using APIs. Given a program with holes, we synthesize completions for holes with the most likely sequences of method calls. Our main idea is to reduce the problem of code completion to a natural-language processing problem of predicting probabilities of sentences. We design a simple and scalable static analysis that extracts sequences of method calls from a large codebase, and index these into a statistical language model. We then employ the language model to find the highest ranked sentences, and use them to synthesize a code completion. Our approach is able to synthesize sequences of calls across multiple objects together with their arguments. Experiments show that our approach is fast and effective. Virtually all computed completions typecheck, and the desired completion appears in the top 3 results in 90% of the cases.

We present a new approach for predicting program properties from massive codebases (aka "Big Code"). Our approach first learns a probabilistic model from existing data and then uses this model to predict properties of new, unseen programs. The key idea of our work is to transform the input program into a representation which allows us to phrase the problem of inferring program properties as structured prediction in machine learning. This formulation enables us to leverage powerful probabilistic graphical models such as conditional random fields (CRFs) in order to perform joint prediction of program properties. As an example of our approach, we built a scalable prediction engine called JSNice for solving two kinds of problems in the context of JavaScript: predicting (syntactic) names of identifiers and predicting (semantic) type annotations of variables. Experimentally, JSNice predicts correct names for 63% of name identifiers and its type annotation predictions are correct in 81% of the cases. In the first week since its release, JSNice was used by more than 30,000 developers and in only few months has become a popular tool in the JavaScript developer community. By formulating the problem of inferring program properties as structured prediction and showing how to perform both learning and inference in this context, our work opens up new possibilities for attacking a wide range of difficult problems in the context of "Big Code" including invariant generation, decompilation, synthesis and others.

In this paper we introduce a new approach for learning precise and general probabilistic models of code based on decision tree learning. Our approach directly benefits an emerging class of statistical programming tools which leverage probabilistic models of code learned over large codebases (e.g., GitHub) to make predictions about new programs (e.g., code completion, repair, etc). The key idea is to phrase the problem of learning a probabilistic model of code as learning a decision tree in a domain specific language over abstract syntax trees (called TGen). This allows us to condition the prediction of a program element on a dynamically computed context. Further, our problem formulation enables us to easily instantiate known decision tree learning algorithms such as ID3, but also to obtain new variants we refer to as ID3+ and E13, not previously explored and ones that outperform ID3 in prediction accuracy. Our approach is general and can be used to learn a probabilistic model of any programming language. We implemented our approach in a system called Deep3 and evaluated it for the challenging task of learning probabilistic models of JavaScript and Python. Our experimental results indicate that Deep3 predicts elements of JavaScript and Python code with precision above 82% and 69%, respectively. Further, Deep3 often significantly outperforms state-of-the-art approaches in overall prediction accuracy.

We present a new approach for learning programs from noisy datasets. Our approach is based on two new concepts: a regularized program generator which produces a candidate program based on a small sample of the entire dataset while avoiding overfitting, and a dataset sampler which carefully samples the dataset by leveraging the candidate program's score on that dataset. The two components are connected in a continuous feedback-directed loop. We show how to apply this approach to two settings: one where the dataset has a bound on the noise, and another without a noise bound. The second setting leads to a new way of performing approximate empirical risk minimization on hypotheses classes formed by a discrete search space. We then present two new kinds of program synthesizers which target the two noise settings. First, we introduce a novel regularized bitstream synthesizer that successfully generates programs even in the presence of incorrect examples. We show that the synthesizer can detect errors in the examples while combating overfitting -- a major problem in existing synthesis techniques. We also show how the approach can be used in a setting where the dataset grows dynamically via new examples (e.g., provided by a human). Second, we present a novel technique for constructing statistical code completion systems. These are systems trained on massive datasets of open source programs, also known as ``Big Code''. The key idea is to introduce a domain specific language (DSL) over trees and to learn functions in that DSL directly from the dataset. These learned functions then condition the predictions made by the system. This is a flexible and powerful technique which generalizes several existing works as we no longer need to decide a priori on what the prediction should be conditioned (another benefit is that the learned functions are a natural mechanism for explaining the prediction). As a result, our code completion system surpasses the prediction capabilities of existing, hard-wired systems.

Like shared-memory multi-threaded programs, event-driven programs such as client-side web applications are susceptible to data races that are hard to reproduce and debug. Race detection for such programs is hampered by their pervasive use of ad hoc synchronization, which can lead to a prohibitive number of false positives. Race detection also faces a scalability challenge, as a large number of short-running event handlers can quickly overwhelm standard vector-clock-based techniques. This paper presents several novel contributions that address both of these challenges. First, we introduce race coverage , a systematic method for exposing ad hoc synchronization and other (potentially harmful) races to the user, significantly reducing false positives. Second, we present an efficient connectivity algorithm for computing race coverage. The algorithm is based on chain decomposition and leverages the structure of event-driven programs to dramatically decrease the overhead of vector clocks. We implemented our techniques in a tool called EventRacer and evaluated it on a number of public web sites. The results indicate substantial performance and precision improvements of our approach over the state-of-the-art. Using EventRacer , we found many harmful races, most of which are beyond the reach of current techniques.