TCSE logo 
 Sigsoft logo
Sustainability badge

This program is tentative and subject to change.

Fri 2 May 2025 12:00 - 12:15 at 212 - AI for Analysis 4

Automatic log analysis is essential for the efficient Operation and Maintenance (O&M) of software systems, providing critical insights into system behaviors. However, existing approaches mostly treat log analysis as training a model to perform an isolated task ( e.g., anomaly detection, log parsing, etc.) using task-specific log-label pairs. These task-based approaches are inflexible in generalizing to complex scenarios, depend on task-specific training data, and cost significantly when deploying multiple models. In this paper, we propose an instruction-based training approach that transforms log-label pairs from multiple tasks and domains into a unified format of instruction-response pairs. Our trained model, LogLM, can follow complex user instructions and generalize better across different tasks, thereby increasing flexibility and reducing the dependence on task-specific training data. By integrating major log analysis tasks into a single model, our approach also relieves model deployment burden. Experimentally, LogLM outperforms existing approaches across five log analysis capabilities, and exhibits strong generalization abilities on complex instructions and unseen tasks.

This program is tentative and subject to change.

Fri 2 May

Displayed time zone: Eastern Time (US & Canada) change

11:00 - 12:30
11:00
15m
Talk
RepairAgent: An Autonomous, LLM-Based Agent for Program Repair
Research Track
Islem BOUZENIA University of Stuttgart, Prem Devanbu University of California at Davis, Michael Pradel University of Stuttgart
11:15
15m
Talk
Evaluating Agent-based Program Repair at Google
SE In Practice (SEIP)
Patrick Rondon Google, Renyao Wei Google, José Pablo Cambronero Google, USA, Jürgen Cito TU Wien, Aaron Sun Google, Siddhant Sanyam Google, Michele Tufano Google, Satish Chandra Google, Inc
11:30
15m
Talk
Anomaly Detection in Large-Scale Cloud Systems: An Industry Case and Dataset
SE In Practice (SEIP)
Mohammad Saiful Islam Toronto Metropolitan University, Toronto, Canada, Mohamed Sami Rakha Toronto Metropolitan University, Toronto, Canada, William Pourmajidi Toronto Metropolitan University, Toronto, Canada, Janakan Sivaloganathan Toronto Metropolitan University, Toronto, Canada, John Steinbacher IBM, Andriy Miranskyy Toronto Metropolitan University (formerly Ryerson University)
Pre-print
11:45
15m
Talk
Crash Report Prioritization for Large-Scale Scheduled Launches
SE In Practice (SEIP)
Nimmi Rashinika Weeraddana University of Waterloo, Sarra Habchi Ubisoft Montréal, Shane McIntosh University of Waterloo
12:00
15m
Talk
LogLM: From Task-based to Instruction-based Automated Log Analysis
SE In Practice (SEIP)
Yilun Liu Huawei co. LTD, Yuhe Ji Huawei co. LTD, Shimin Tao University of Science and Technology of China; Huawei co. LTD, Minggui He Huawei co. LTD, Weibin Meng Huawei co. LTD, Shenglin Zhang Nankai University, Yongqian Sun Nankai University, Yuming Xie Huawei co. LTD, Boxing Chen Huawei Canada, Hao Yang Huawei co. LTD
Pre-print
12:15
7m
Talk
Using ML filters to help automated vulnerability repairs: when it helps and when it doesn’t
New Ideas and Emerging Results (NIER)
Maria Camporese University of Trento, Fabio Massacci University of Trento; Vrije Universiteit Amsterdam
:
:
:
: