Probabilistic model for code with decision trees

By Veselin Raychev, Pavol Bielik, Martin T. Vechev. 2016

280 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems, Software, Code Generation & Program Synthesis, Program Analysis & Verification

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Percy Liang
2. Mohit Bansal
3. Graham Neubig
4. Jianfeng Gao
5. Luke S. Zettlemoyer
6. Caiming Xiong
7. Quoc V. Le
8. Yejin Choi

In Information Systems

1. Charles Sutton
2. Miltiadis Allamanis
3. Uri Alon
4. Michael Pradel
5. Eran Yahav
6. Koushik Sen
7. Rishabh Singh
8. Percy Liang

In Software

1. Michael Pradel
2. Charles Sutton
3. Koushik Sen
4. Miltiadis Allamanis
5. Rishabh Singh
6. Uri Alon
7. Vincent J. Hellendoorn
8. Yue Wang

In Code Generation & Program Synthesis

1. Charles Sutton
2. Uri Alon
3. Miltiadis Allamanis
4. Michael Pradel
5. Rishabh Singh
6. Eran Yahav
7. Ge Li
8. Koushik Sen

In Program Analysis & Verification

1. Petros Maniatis
2. Aditya Kanade
3. Charles Sutton
4. Eran Yahav
5. Uri Alon
6. Georgios Gousios
7. Marc Brockschmidt
8. Michael Pradel

Citations in the corpus

• Program Synthesis with Large Language Models (2021, methodology): “Code completion has been a particular focus of interest [Raychev et al., 2016, Karampatsis et al., 2020, Svyatkovskiy et al., 2020, Kim et al., 2020].”
• CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis (2022, background): “A popular input forms include preceding code in the same line (Raychev et al., 2016), pseudo-code (Kulal et al.”
• Aligning Language Models with Preferences through f-divergence Minimization (2023, methodology): “For that, we condition on Python function signatures in the Python150 dataset (Raychev et al., 2016) which consists of Python source code obtained from GitHub.”
• CodeXGLUE: A Machine Learning Benchmark Dataset for Code Understanding and Generation (2021, methodology): “We use two influential datasets, PY150[57] in python and Github Java Corpus[1] in Java.”
• Measuring Coding Challenge Competence With APPS (2021, neutral): “We compared to several datasets in the (Kulal et al., 2019; Yu et al., 2018; Raychev et al., 2016; Iyer et al., 2018; Lu et al., 2021) main paper.”
• code2seq: Generating Sequences from Structured Representations of Code (2018, background): “Code representation models that use syntactic information have usually been evaluated on relatively easier tasks, which mainly focus on “filling the blanks” in a given program (Alon et al., 2018b; Bielik et al., 2016; Raychev et al., 2016; 2015; Allamanis et al., 2018) or semantic classification of ”
• Learning to Represent Programs with Graphs (2017, methodology): “For example, Hindle et al. (2012); Bhoopchand et al. (2016) model the code as a sequence of tokens, while Maddison & Tarlow (2014); Raychev et al. (2016) model the syntax tree structure of code.”
• DeepBugs: a learning approach to name-based bug detection (2018, methodology): “As a corpus of code, we use 150,000 JavaScript files provided by the authors of earlier work [Raychev et al. 2016b].”
• Multi-lingual Evaluation of Code Generation Models (2022, methodology): “(2021) composed a token and line completion datasets in Java and Python based on existing benchmarks (Allamanis & Sutton, 2013; Raychev et al., 2016).”
• WILDS: A Benchmark of in-the-Wild Distribution Shifts (2020, positive): “We study a variant of the Py150 Dataset (Raychev et al., 2016; Lu et al., 2021), where the goal is to predict the next token given the context of previous tokens.”
• Neural Program Repair by Jointly Learning to Localize and Repair (2019, methodology): “Primarily, we use ETH-Py150 2 , a public corpus of GitHub Python files extensively used in the literature (Raychev et al., 2016; Vechev & Yahav, 2016).”
• Pre-trained Contextual Embedding of Source Code (2019, methodology): “We use the ETH Py150 corpus (Raychev et al., 2016) to generate datasets for the fine-tuning tasks.”
• UniXcoder: Unified Cross-Modal Pre-training for Code Representation (2022, methodology): “We use PY150 (Raychev et al., 2016) and Github Java Corpus (Allamanis and Sutton, 2013) datasets in CodeXGLUE (Lu et al., 2021) for line-level code completion tasks.”
• Learning to Fuzz: Application-Independent Fuzz Testing with Probabilistic, Generative Models of Input Data (2016, background): “In contrast, TreeFuzz supports models that cannot be synthesized by Deep3 and has linear time complexity without sampling.”
• CodeNet: A Large-Scale AI for Code Dataset for Learning a Diversity of Coding Tasks (2021, background): “Such tasks include clone detection, vulnerability detection [10, 11], cloze test [12], code completion [13, 14], code repair [15], code-to-code translation, natural language code search [16], text-to-code generation [17], and code summarization [16].”
• Learning Autocompletion from Real-World Datasets (2020, background): “[9] leveraged a collection of 150k committed Python files from GitHub.”