JupOtter: Cell-Level Bug Detection in Jupyter Notebooks
This program is tentative and subject to change.
Jupyter Notebooks are an increasingly popular coding environment used across many domains, especially in Python-based data science and scientific computing. Originally used for prototyping and interactive exploration, notebooks are increasingly used to develop more complex programs, leading to a rapid rise in buggy notebooks on platforms like GitHub. To address this trend, we present \textbf{JupOtter}, a bug detection system designed specifically for Jupyter Notebooks. JupOtter features three novel contributions: (1) a notebook-specific tokenization strategy that preserves cell structure, (2) a cell-level bug prediction technique, and (3) a new labeled dataset, \textbf{OtterDataset}, containing over 21,000 notebooks annotated for fine-grained cell-level bug detection. JupOtter achieves cell-level bug detection F1 scores that surpass static analyzers and large language models in two out of three evaluation datasets.
This program is tentative and subject to change.
Wed 16 SepDisplayed time zone: Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna change
11:00 - 12:30 | Session 1 - Code Whisperers: AI in the Developer’s ChairRegistered Reports / Research Papers Track / Tool Demonstration and Data Showcase Track / Visions and Emerging Results Track / Industry Track at Auditorium Chair(s): Masud Rahman Dalhousie University Theme: AI-Driven Software Development | ||
11:00 20mPaper | Integrating Crash Report Mining and LLMs for Bug Localization and Repair: An Industrial Report Industry Track Marcos Medeiros Federal University of Rio Grande do Norte, Uirá Kulesza Federal University of Rio Grande do Norte, Christoph Treude Singapore Management University, Daniel Lucena Federal University of Rio Grande do Norte, Rafael Gomes Federal University of Rio Grande do Norte, Roberta Coelho , Eiji Adachi Barbosa Federal University of Rio Grande do Norte, Rodrigo Bonifácio Informatics Center - CIn/UFPE and Computer Science Department / University of Brasília Pre-print | ||
11:20 20mPaper | JupOtter: Cell-Level Bug Detection in Jupyter Notebooks Research Papers Track | ||
11:40 20mPaper | Quantize with Confidence? An Empirical Study of Quantization for Code Generation Research Papers Track Saima Afrin William and Mary, USA, Md. Zahidul Haque William & Mary, Antonio Mastropaolo William and Mary, USA | ||
12:00 10mShort-paper | ATLAS: Multi-View Code Representation Tool for C and C++ Source Programs Tool Demonstration and Data Showcase Track Jaid Monwar Chowdhury University of Texas at Arlington, Ahmad Farhan Shahriar Chowdhury Bangladesh University of Engineering and Technology, Humayra Binte Monwar Bangladesh University of Engineering and Technology, Mahmuda Naznin Bangladesh University of Engineering and Technology Pre-print Media Attached | ||
12:10 10mShort-paper | Code Review is a Conversation: Toward Conversational AI Review Assistants Visions and Emerging Results Track Rosalia Tufano Università della Svizzera Italiana Pre-print | ||
12:20 10mShort-paper | Cleaning Logs for Downstream Tasks Registered Reports Zahra Ghavidel Yazdi University of Luxembourg, Van-Hoang Le University of Luxembourg, Nyyti Saarimäki University of Luxembourg, Donghwan Shin University of Sheffield, Domenico Bianculli University of Luxembourg, Lionel Briand University of Ottawa, Canada; Lero centre, University of Limerick, Ireland Pre-print | ||