14 July 2024 in Performance Evaluation
20 September 2024 in IEEE Transactions on Biomedical Engineering
Objective: Our aim is to determine if data collected with inertial measurement units (IMUs) during steady-state running could be used to estimate ground reaction forces (GRFs) and to derive biomechanical variables (e.g., contact time, impulse, change in velocity) using lightweight machine-learning approaches. In contrast, state-of-the-art estimation using LSTMs suffers from prohibitive inference times on edge devices, requires expensive training and hyperparameter optimization, and results in black box models. Methods: We proposed a novel lightweight solution, SVD Embedding Regression (SER), using linear regression between SVD embeddings of IMU data and GRF data. We also compared lightweight solutions including SER and k-Nearest-Neighbors (KNN) regression with state-of-the-art LSTMs. Results: We performed extensive experiments to evaluate these techniques under multiple scenarios and combinations of IMU signals and quantified estimation errors for predicting GRFs and biomechanical variables. We did this using training data from different athletes, from the same athlete, or both, and we explored the use of acceleration and angular velocity data from sensors at different locations (sacrum and shanks). Conclusion: Our results illustrated that lightweight solutions such as SER and KNN can be similarly accurate or more accurate than LSTMs. The use of personal data reduced estimation errors of all methods, particularly for most biomechanical variables (as compared to GRFs); moreover, this gain was more pronounced in the lightweight methods. Significance: The study of GRFs is used to characterize the mechanical loading experienced by individuals in movements such as running, which is clinically applicable to identify athletes at risk for stress-related injuries.
16 February 2022 in IEEE Transactions on Parallel and Distributed Systems
Training jobs of deep neural networks (DNNs) can be accelerated through distributed variants of stochastic gradient descent (SGD), where multiple nodes process training examples and exchange updates. The total throughput of the nodes depends not only on their computing power, but also on their networking speeds and coordination mechanism (synchronous or asynchronous, centralized or decentralized), since communication bottlenecks and stragglers can result in sublinear scaling when additional nodes are provisioned. In this paper, we propose two classes of performance models to predict throughput of distributed SGD: fine-grained models , representing many elementary computation/communication operations and their dependencies; and coarse-grained models , where SGD steps at each node are represented as a sequence of high-level phases without parallelism between computation and communication. Using a PyTorch implementation, real-world DNN models and different cloud environments, our experimental evaluation illustrates that, while fine-grained models are more accurate and can be easily adapted to new variants of distributed SGD, coarse-grained models can provide similarly accurate predictions when augmented with ad hoc heuristics, and their parameters can be estimated with profiling information that is easier to collect.
26 October 2023 in ACM Transactions on Modeling and Computer Simulation
We evaluate a stochastic upper bound on the response time Probability Density Function (PDF) of complex workflows through an efficient and accurate compositional approach. Workflows consist of activities having generally distributed stochastic durations with bounded supports, composed through sequence, choice/merge, and balanced/unbalanced split/join operators, possibly breaking the structure of well-formed nesting. Workflows are specified using a formalism defined in terms of Stochastic Time Petri Nets that permits decomposition into a hierarchy of subworkflows with positively correlated response times, guaranteeing that a stochastically larger end-to-end response time PDF is obtained when intermediate results are approximated by stochastically larger PDFs and when dependencies are simplified by replicating activities appearing in multiple subworkflows. In particular, an accurate stochastically larger PDF is obtained by combining shifted truncated Exponential terms with positive or negative rates. Experiments are performed on sets of manually and randomly generated models with increasing complexity, illustrating under which conditions different decomposition heuristics work well in terms of accuracy and complexity and showing that the proposed approach outperforms simulation having the same execution time.