High-Energy Physics experiments are facing a multi-fold data increase with every new iteration. This is certainly the case for the upcoming High-Luminosity LHC upgrade. Such increased data processing requirements forces revisions to almost every step of the data processing pipeline. One such step in need of an overhaul is the task of particle track reconstruction, a.k.a., tracking. A Machine Learning-assisted solution is expected to provide significant improvements, since the most time-consuming step in tracking is the assignment of hits to particles or track candidates. This is the topic of this paper. We take inspiration from large language models. As such, we consider two approaches: the prediction of the next word in a sentence (next hit point in a track), as well as the one-shot prediction of all hits within an event. In an extensive design effort, we have experimented with three models based on the Transformer architecture and one model based on the U-Net architecture, performing track association predictions for collision event hit points. In our evaluation, we consider a spectrum of simple to complex representations of the problem, eliminating designs with lower metrics early on. We report extensive results, covering both prediction accuracy (score) and computational performance. We have made use of the REDVID simulation framework, as well as reductions applied to the TrackML data set, to compose five data sets from simple to complex, for our experiments. The results highlight distinct advantages among different designs in terms of prediction accuracy and computational performance, demonstrating the efficiency of our methodology. Most importantly, the results show the viability of a one-shot encoder-classifier based Transformer solution as a practical approach for the task of tracking.
Serverless offers a scalable and cost-effective service model for users to run applications without focusing on underlying infrastructure or physical servers. While the Serverless architecture is not designed to address the unique challenges posed by resource-intensive workloads, e.g., Machine Learning (ML) tasks, it is highly scalable. Due to the limitations of Serverless function deployment and resource provisioning, the combination of ML and Serverless is a complex undertaking. We tackle this problem through decomposition of large ML models into smaller sub-models, referred to as slices. We set up ML inference tasks using these slices as a Serverless workflow, i.e., sequence of functions. Our experimental evaluations are performed on the Serverless offering by AWS for demonstration purposes, considering an open-source format for ML model representation, Open Neural Network Exchange. Achieved results portray that our decomposition method enables the execution of ML inference tasks on Serverless, regardless of the model size, benefiting from the high scalability of this architecture while lowering the strain on computing resources, such as required run-time memory.
We present how electrical Extra-Functional Behavioural (EFB) metrics can be used in generating accurate representations of executional units for industrial Cyber-Physical Systems (CPS). We achieve this by employing our concept of power passports (the representation), created per metric and per execution phase (executional unit). We employ these representations alongside supervised classification algorithms, i.e., Decision Tree and Random Forest, in an effective data analytical pipeline. Our approach is capable of detecting anomalous operational conditions and predicting the type of anomaly, out of different known types, with significant overall accuracies, as high as 99% in certain set-ups. We consider anomalous operational conditions as nonreference conditions, resulting in loss of performance or unreliable operation of a system. Our experiments are designed to reflect real-world conditions as much as possible and all of our collected raw data comes from real executions, normal and anomalous, with no synthetic manipulation. Our results show that a black box approach towards systems under scrutiny for anomaly detection and classification, given its accuracy and considering the limitations applicable to low-power industrial CPS, can be the preferred one.
Neural networks have become a cornerstone of machine learning. As the trend for these to get more and more complex continues, so does the underlying hardware and software infrastructure for training and deployment. In this survey we answer three research questions: "What types of model parallelism exist?", "What are the challenges of model parallelism?", and "What is a modern use-case of model parallelism?" We answer the first question by looking at how neural networks can be parallelised and expressing these as operator graphs while exploring the available dimensions. The dimensions along which neural networks can be parallelised are intra-operator and inter-operator. We answer the second question by collecting and listing both implementation challenges for the types of parallelism, as well as the problem of optimally partitioning the operator graph. We answer the last question by collecting and listing how parallelism is applied in modern multi-billion parameter transformer networks, to the extend that this is possible with the limited information shared about these networks.
Serverless computing has emerged as an attractive paradigm due to the efficiency of development and the ease of deployment without managing any underlying infrastructure. Nevertheless, serverless computing approaches face numerous challenges to unlock their full potential in hybrid environments. To gain a deeper understanding and firsthand knowledge of serverless computing in edge-cloud deployments, we review the current state of open-source serverless platforms and compare them based on predefined requirements. We then design and implement a serverless computing platform with a novel edge orchestration technique that seamlessly deploys serverless functions across the edge and cloud environments on top of the Knative serverless platform. Moreover, we propose an offloading strategy for edge environments and four different functions for experimentation and showcase the performance benefits of our solution. Our results demonstrate that such an approach can efficiently utilize both cloud and edge resources by dynamically offloading functions from the edge to the cloud during high activity, while reducing the overall application latency and increasing request throughput compared to an edge-only deployment.