June 09, 2023
| 8:30 – 9:00 | Breakfast (—) |
| 9:00 – 9:15 | SEMLA 2022 Opening Session Opening by Prof. Foutse Khomh Welcome message from TBD |
| 9:15 – 10:45 | Keynote Session: Operationalizing Trustworthy Large-Scale AI — Chair: TBD Lionel C. Briand (University of Ottawa, University of Lexembourg) Miryung Kim (University of California, Los Angeles) Walid Maalej (University of Hamburg): Tailoring Requirements Engineering for Responsible AI |
| 10:45 – 11:00 | Coffee break |
| 11:00 – 12:00 | Panel 1: Operationalizing Trustworthy Large-Scale AI — Moderator: TBD Lionel Briand (University of Ottawa, University of Lexembourg) Miryung Kim (University of California, Los Angeles) Walid Maalej (University of Hamburg) |
| 12:00 – 14:00 | Lunch break |
| 14:00 – 16:00 | Tutorials – Chair: TBD Debugging and Testing Deep Reinforcement Learning. Ahmed Haj Yahmed, Rached Bouchoucha, Tambon Florian, Vahid Majdinasab (Polytechnique Montréal) |
| 16:00 – 16:30 | Coffee break |
| 16:30 – 18:00 | Session: Trustworthy Large Language Models – Chair: TBD Sarath Chandar (Polytechnique Montréal, MILA) Su Lin Blodgett (Microsoft Research) Yuan Tian (Queen’s University): Optimizing Software Project Management with AI-powered Tracking Tools |
| 18:05-20:00 | Reception & Poster session — Poster chair: TBD |
June 10, 2023
| 8:30 – 9:00 | Breakfast |
| 9:00 – 10:30 | Keynote Session: Keynote session II: Trustworthy Large-Scale AI in Practice — Chair: TBD Sumit Gulwani (Microsoft Research) – AI-assisted Programming: Applications, Experiences, and Neuro-symbolic perspectives Maryam Ahmadi (Canadian National Railway, BrainStation) – Event Log Anomaly Detection Ahmed E. Hassan (Queen’s University) – Software Engineering and Foundation Models: Software 4.0 and AI teammates |
| 10:30 – 11:00 | Coffee break |
| 11:00 – 12:00 | Industry Panel Discussion: Trustworthy Large-Scale AI in Practicek — Moderator: TBD Ahmed E. Hassan (Queen’s University) Foutse Khomh (Polytechnique Montréal, MILA) Sumit Gulwani (Microsoft Research) Maryam Ahmadi (Canadian National Railway, BrainStation) |
| 12:00 – 14:00 | Lunch break |
| 14:00 – 15:30 | Special Panel: Ethical and Legal Implications of LLMs for Code – Moderator: TBD Sumit Gulwani (Microsoft Research) Fouse Khomh (Polytechnique Montréal) Valentin Callipel (Laboratoire de cyberjustice, Université de Montréal) |
| 15:30 – 16:00 | Coffee break |
| 16:00 – 17:30 | Session: Quality Assurance for AI — Chair: TBD Jin Guo (McGill University) Ettore Merlo (Polytechnique Montréal) – Out-of-distribution Analysis and Robustness of Deep Neural Networks Fatemeh Hendijani Fard (University of British Columbia) – Exploiting The Learned Knowledge of Language Models Using Adapters |
| 17:30 – 17:45 | SEMLA wrap-up — Foutse Khomh & Maxime Lamothe & Heng Li & Mohammad Hamdaqa & Zohreh Sharafi |
Abstract: Anomaly detection plays an important role in management of modern large-scale distributed systems. Logs, which record system runtime information, are widely used for anomaly detection. However, Unsupervised anomaly detection algorithms face challenges in addressing complex systems, which generate vast amounts of multivariate time series data. Timely anomaly detection is crucial for managing these systems effectively and minimizing downtime. This proactive approach minimizes system downtime and plays a vital role in incident management for large-scale systems. To address these challenges, a method called Multi-Scale Convolutional Recurrent Encoder-Decoder (MSCRED) has been developed for detecting anomalies in CN PTC system logs. MSCRED leverages the power of multivariate time series data to perform anomaly detection and diagnosis. It creates multi-scale signature matrices that capture different levels of system statuses across various time steps. The method utilizes a convolutional encoder to capture inter-sensor correlations and a Convolutional Long-Short Term Memory (ConvLSTM) network with attention mechanisms to capture temporal patterns.
Abstract
Abstract: Language models such as RoBERTa, CodeBERT, and GraphCodeBERT have gotten much attention in the past three years for various Software Engineering tasks. Though these models are proven to have state-of-the-art performance for many SE tasks, such as code summarization, they often require to be fully fine-tuned for the downstream task. Is there a better way for fine-tuning these models that require training fewer parameters? Can we impose new information on the current models without pre-training them again? How do these models perform for different programming languages, especially low-resource ones with less training data available? How can we use the knowledge learned from other programming languages to improve the performance of low-resource languages? This talk will review a series of experiments and our contributions to answering these questions.
