Jeremy Bradbury

🚀 Mark your calendars 🗓️! The IEEE CASCON 2025 call for contributions and proposals is open.

We look forward to your submissions!

#CASCON2025 will be held this year at vibrant Toronto, Second Student Centre, York University, 15 Library Lane, Toronto, ON, Canada
More details: cascon.ca

SEER Lab’s Jeremy Bradbury & Riddhi More are in Naples, Italy 🇮🇹 attending the 18th IEEE Int’l Conference on Software Testing, Verification and Validation (#ICST2025) where they are presenting two research papers. Dr. Bradbury is also a member of the faculty panel at the ICST Doctoral Symposium.

Full program details are available at: conf.researchr.org/home/icst-20...

Today we had a successful 8th IEEE Int’l Workshop on Next Level of Test Automation (#NEXTA2025). Highlights included:

✅ KEYNOTE by Serge Demeyer
✅ TALKS
- Influence of Pure & Unit-Like Tests on SBFL
- Witness Test Program Generation
- Adaptive Testing for LLM-Based Applications
✅ DISCUSSION!

So most actors are really ghost actors 🤔

Looking for evidence of data leakage in the #HumanEval code generation #LLM benchmark? Check out our analysis of a subset of HumanEval tasks vs comparable tasks on #ChatGPT, #Claude and #Llama. @riddhimore.bsky.social

Human Brain vs. LLM Agent
🧠 vs. 🤖

Hardware…
Consciousness…
Learning style…
Creativity…

- Bang Liu, #SANER2025 Keynote

The Large Language Model (#LLM) Lifecycle from #SANER2025 keynote by Dr. Bang Liu, University of Montreal

➡️ Pre-Training (e.g. data synthesis, model architecture)
➡️ Post-Training (e.g. instruction tuning, bias mitigation)
➡️ Serving (e.g. cloud deployment)
➡️ Utilization(e.g. prompt engineering, RAG)

Riddhi More presenting at FAIRNESS 2025

Yesterday Riddhi More presented our paper “Assessing Data Augmentation-Induced Bias in Training and Testing of Machine Learning Models” at the 1st Int’l Workshop on Fairness in Software Systems (#FAIRNESS2025) in Montreal, Canada. 🇨🇦
Paper preprint, slides & code/data available at: bit.ly/4brEBJm

“Code generation is just the tip of the iceberg towards automated software engineering!!” - Dr. Baishakhi Ray (Columbia University), #SANER2025 Keynote

MSc student Riddhi More & Prof. Jeremy Bradbury are in Montreal, PQ this week attending the IEEE Int’l Conference on Software Analysis, Evolution and Reengineering (#SANER2025) and the 1st Int’l Workshop on Fairness in Software Systems (#FAIRNESS2025).

A great keynote by Foutse Khomh (@swatlab.bsky.social) on “Navigating Fairness Challenges in Intelligent Software Systems” at #FAIRNESS2025. I appreciated the emphasis on #AI model transparency, auditing and governance as essential to the development of fair systems that mitigate bias.

Dr. Michael Miljanovic presenting at SIGCSE 2025.

Today Dr. Michael Miljanovic presented our latest paper at #SIGCSE2025. The paper titled “How Effective and Efficient Are Student-Written Software Tests?”, is co-authored by Amanda Showler, Michael Miljanovic, and Jeremy Bradbury. 📄: www.sqrlab.ca/publications...

Hello 🦋 world!

We're getting ready for IEEE SCAM 2025 in Auckland, New Zealand.

conf.researchr.org/home/scam-2025

Thank you to the organizers of the #AI Meets Software Quality Int’l Colloquium for inviting Riddhi More & I to speak on Monday. If you're interested in #LLMs & #flakytests the preprints of the papers presented are available at bit.ly/40OvYUa (FlakyXbert) & bit.ly/4jILzNT (augmentation bias).

Any job? I don’t think so. The job of a CEO? Probably. 😊

Tomorrow Prof. Jeremy Bradbury & MSc student Riddhi More will be presenting "FlakyXbert: A Few-Shot Learning Framework for Detecting and Classifying Flaky Tests" at the AI Meets Software Quality International Colloquium (bit.ly/42DAZ4M)
#AI #LLM #testing #flakytests #dataaugmentation #bias

Our paper "Addressing #DataLeakage in #HumanEval Using Combinatorial Test Design" has been accepted for publication at the Int’l Conference on Software Testing, Verification & Validation (#ICST2025), Short Papers, Vision & Emerging Results.
co-author: Riddhi More
arxiv.org/abs/2412.01526
#benchmark

She’s already in the starter pack 😊

I'm happy to share our paper "Assessing Data Augmentation-Induced Bias in Training and Testing of Machine Learning Models," has been accepted to the 1st International Workshop on Fairness in Software Systems (FAIRNESS 2025), co-located with #SANER2025! #ontariotech @riddhimore.bsky.social

Are you working on #softwaretesting & passionate about test #automation and #AI testing? If so, consider submitting to the 8th IEEE Workshop on the Next level of Test Automation #NEXTA2025, co-located w/ #ICST2025
Submission Date: Jan 10, 2025 [UPDATED!]
bit.ly/nexta2025

I’m excited to share that our paper "An Analysis of #LLM Fine-Tuning & Few-Shot Learning for Flaky Test Detection & Classification" has been accepted for publication at the Int’l Conf on Software Testing, Verification & Validation (#ICST2025). This paper is first-authored by @riddhimore.bsky.social

I had a great time today attending the #OntarioTech Cakes & #AI Winter Show-And-Tell event. A great opportunity to engage with researchers from across the university and establish interdisciplinary collaborations. Thanks to @petelewis.bsky.social for organizing! 🍰

I’m just saying that 200k might be the revenue from tuition but those students also need a classroom, labs, accessibility services, library resources, academic advising, etc. All of those costs need to be factored into any analysis.

First, I think sessional instructors are underpaid and deserve fair compensation. However, I think your analysis of profit per course needs to factor in the true cost of education which includes student services, facilities, materials, equipment, etc.

The latest episode of the Front Burner podcast is worth a listen for anyone with a connection to Canadian #highered. 👇
Featuring an interview with @alexusherhesa.bsky.social on the current financial crisis facing Canadian universities & colleges. bit.ly/3B1FjiW

Akramul Azim presenting

Thank you to Prof. Akramul Azim for giving a guest lecture today in my “Applications of #LLMs to Software Engineering” grad course. He talked about his #CASCON2024 paper “FaultLLAMA2: A Fine Tuned Large Language Model for Code Recommendations to Reduce Java Lock Contention” #OntarioTech #concurrency

Cristiano Politowski speaking

It was great yesterday to have my colleague Cristiano Politowski speak to our graduate students and share his research experience at the intersection of Software Engineering and Game Development. #OntarioTech #softwareengineering #gamedev @crispolito87.bsky.social

Interested in joining the Software Engineering & Education Research Lab? Find out more at bit.ly/3ARpCe8 #OntarioTech 🇨🇦

Dear Google Labs -- notebookLLM creates nice, sunny podcasts on STEM papers---but...

Science, 4real, is -ve & adversarial. Please train another model, with data from actual grad students or faculty, so it creates more grumpy pods, which also criticize scope, methods, results etc of papers? 🔭🐡🧪