Reza Karimzadeh

Reza Karimzadeh, Emad Fatemizadeh, Hossein Arabi

Automated medical image segmentation is an essential task to aid/speed up diagnosis and treatment procedures in clinical practices. Deep convolutional neural networks have exhibited promising performance in accurate and automatic seminal segmentation. For segmentation tasks, these methods normally rely on minimizing a cost/loss function that is designed to maximize the overlap between the estimated target and the ground-truth mask delineated by the experts. A simple loss function based on the degrees of overlap (i.e., Dice metric) would not take into account the underlying shape and morphology of the target subject, as well as its realistic/natural variations; therefore, suboptimal segmentation results would be observed in the form of islands of voxels, holes, and unrealistic shapes or deformations. In this light, many studies have been conducted to refine/post-process the segmentation outcome and consider an initial guess as prior knowledge to avoid outliers and/or unrealistic estimations. In this study, a novel shape-based cost function is proposed which encourages/constrains the network to learn/capture the underlying shape features in order to generate a valid/realistic estimation of the target structure. To this end, the Principal Component Analysis (PCA) was performed on a vectorized training dataset to extract eigenvalues and eigenvectors of the target subjects. The key idea was to use the reconstruction weights to discriminate valid outcomes from outliers/erroneous estimations.

Ehsan Yousefzadeh-Asl-Miandoab, Reza Karimzadeh, Danyal Yorulmaz et al.

Collocating deep learning training tasks improves GPU utilization but risks resource contention, severe slowdowns, and out-of-memory (OOM) failures. Accurate memory estimation is essential for robust collocation, and GPU utilization estimation -- a key proxy for contention -- enables interference-aware scheduling. Existing GPU memory estimators span three paradigms -- analytical models, CPU-side libraries, and ML-based estimators -- each with distinct limitations: dependence on detailed model specifications, intrusive integration, poor generalization, and varying latency overhead. GPU heterogeneity further complicates estimation, as identical tasks can exhibit different memory footprints across hardware generations. GPU utilization remains comparatively understudied, further complicated by non-additive utilization metrics and GPU heterogeneity. We conduct a systematic analysis of representative estimators from each paradigm -- Horus, PyTorch FakeTensor, and our lightweight ML-based estimator -- evaluating accuracy, generalizability, and overhead. We construct a synthetic dataset spanning MLPs, CNNs, and Transformers with controlled architectural variations, and train MLP- and Transformer-based estimators for memory prediction, and experiment with utilization estimation. Our evaluation reveals key tradeoffs and validates estimators against real-world unseen models. Significant challenges remain: analytical models lack generalization and cannot easily be extended to new GPU architectures or accurately reflect memory optimization savings; CPU-side libraries impose intrusive integration overhead; and both analytical and ML-based estimators rely on model specifications or computation graphs, limiting generalization across diverse architectures and hardware variants. We release all datasets, tools, and artifacts to support further research.

Reza Karimzadeh, Albert Alonso, Frans Zdyb et al.

Faithful yet compact explanations for vision models remain a challenge, as commonly used dense perturbation masks are often fragmented and overfitted, needing careful post-processing. Here, we present a training-free explanation method that replaces dense masks with smooth tunable contours. A star-convex region is parameterized by a truncated Fourier series and optimized under an extremal preserve/delete objective using the classifier gradients. The approach guarantees a single, simply connected mask, cuts the number of free parameters by orders of magnitude, and yields stable boundary updates without cleanup. Restricting solutions to low-dimensional, smooth contours makes the method robust to adversarial masking artifacts. On ImageNet classifiers, it matches the extremal fidelity of dense masks while producing compact, interpretable regions with improved run-to-run consistency. Explicit area control also enables importance contour maps, yielding a transparent fidelity-area profiles. Finally, we extend the approach to multi-contour and show how it can localize multiple objects within the same framework. Across benchmarks, the method achieves higher relevance mass and lower complexity than gradient and perturbation based baselines, with especially strong gains on self-supervised DINO models where it improves relevance mass by over 15% and maintains positive faithfulness correlations.

Ehsan Yousefzadeh-Asl-Miandoab, Reza Karimzadeh, Bulat Ibragimov et al.

GPUs running deep learning (DL) workloads are frequently underutilized. Collocating multiple DL training tasks on the same GPU can improve utilization but introduces two key risks: (1) out-of-memory (OOM) crashes for newly scheduled tasks, and (2) severe performance interference among co-running tasks, which can negate any throughput gains. These issues reduce system robustness, quality of service, and energy efficiency. We present CARMA, a task-level, collocation-aware resource management system for the server-scale. CARMA addresses collocation challenges via (1) fine-grained monitoring and bookkeeping of GPUs and a collocation risk analysis that filters out the high-risk GPUs; (2) task placement policies that cap GPU utilization to avoid OOMs and limit interference; (3) integration of GPU memory need estimators for DL tasks to minimize OOMs during collocation; and (4) a lightweight recovery method that relaunches jobs crashed due to OOMs. Our evaluation on a DL training workload derived from real-world traces shows that CARMA uses GPUs more efficiently by making more informed collocation decisions: for the best-performing collocation policy, CARMA increases GPU streaming multiprocessor (SM) utilization by 54%, the parallelism achieved per SM by 61%, and memory use by 62%. This results in a $\sim$35% and $\sim$15% reduction in the end-to-end execution time (makespan) and GPU energy consumption, respectively, for this workload.