From the 2 of 8 linked papers with an AI index.
8 papers
Deep4ge: DNN Training Trajectories for Fault Detection and Diagnosis
Sigma Jahan
The paper introduces Deep4ge, a benchmark dataset of over 14,000 deep neural network training runs—including faulty and correct variants—capturing per‑epoch metrics and features to…
Toward Localizing and Repairing Bias in Transformer Attention Heads
Sigma Jahan
The paper proposes ROBIN, a white‑box method that identifies and modifies specific transformer attention heads at inference time to reduce gender bias while preserving language mod…
Evaluation-Strategy Gap in Fault Diagnosis of Deep Learning Programs
Sigma Jahan
Deep Learning (DL) programs can fail during training for many reasons, and diagnosing the cause is a costly and time-consuming maintenance task. Techniques for diagnosing such fail…
Hierarchical Fault Detection and Diagnosis for Transformer Architectures
Sigma Jahan, Saurabh Singh Rajput, Tushar Sharma +1
Transformers now underpin critical AI systems across industry and research. Yet their faults can silently alter model behavior without runtime errors, and existing techniques offer…
Why Attention Fails: A Taxonomy of Faults in Attention-Based Neural Networks
Sigma Jahan, Saurabh Singh Rajput, Tushar Sharma +1
Attention mechanisms are at the core of modern neural architectures, powering systems ranging from ChatGPT to autonomous vehicles and driving a major economic impact. However, high…
Towards Understanding the Challenges of Bug Localization in Deep Learning Systems
Sigma Jahan, Mehil B. Shah, Mohammad Masudur Rahman
Software bugs cost the global economy billions of dollars annually and claim ~50\% of the programming time from software developers. Locating these bugs is crucial for their resolu…