PowerTrain: Fast, Generalizable Time and Power Prediction Models to Optimize DNN Training on Accelerated Edges

Accelerated edge devices, like Nvidia's Jetson with 1000+ CUDA cores, are increasingly used for DNN training and federated learning, rather than just for inferencing workloads. A unique feature of these compact devices is their fine-grained control over CPU, GPU, memory frequencies, and active CPU cores, which can limit their power envelope in a constrained setting while throttling the compute performance. Given this vast 10k+ parameter space, selecting a power mode for dynamically arriving training workloads to exploit power-performance trade-offs requires costly profiling for each new workload, or is done \textit{ad hoc}. We propose \textit{PowerTrain}, a transfer-learning approach to accurately predict the power and time consumed when training a given DNN workload (model + dataset) using any specified power mode (CPU/GPU/memory frequencies, core-count). It requires a one-time offline profiling of $1000$s of power modes for a reference DNN workload on a single Jetson device (Orin AGX) to build Neural Network (NN) based prediction models for time and power. These NN models are subsequently transferred (retrained) for a new DNN workload, or even a different Jetson device, with minimal additional profiling of just $50$ power modes to make accurate time and power predictions. These are then used to rapidly construct the Pareto front and select the optimal power mode for the new workload. PowerTrain's predictions are robust to new workloads, exhibiting a low MAPE of $<6\%$ for power and $<15\%$ for time on six new training workloads for up to $4400$ power modes, when transferred from a ResNet reference workload on Orin AGX. It is also resilient when transferred to two entirely new Jetson devices with prediction errors of $<14.5\%$ and $<11\%$. These outperform baseline predictions by more than $10\%$ and baseline optimizations by up to $45\%$ on time and $88\%$ on power.

Building AI Agents for Autonomous Clouds: Challenges and Design Principles

The rapid growth in the use of Large Language Models (LLMs) and AI Agents as part of software development and deployment is revolutionizing the information technology landscape. While code generation receives significant attention, a higher-impact application lies in using AI agents for operational resilience of cloud services, which currently require significant human effort and domain knowledge. There is a growing interest in AI for IT Operations (AIOps) which aims to automate complex operational tasks, like fault localization and root cause analysis, thereby reducing human intervention and customer impact. However, achieving the vision of autonomous and self-healing clouds though AIOps is hampered by the lack of standardized frameworks for building, evaluating, and improving AIOps agents. This vision paper lays the groundwork for such a framework by first framing the requirements and then discussing design decisions that satisfy them. We also propose AIOpsLab, a prototype implementation leveraging agent-cloud-interface that orchestrates an application, injects real-time faults using chaos engineering, and interfaces with an agent to localize and resolve the faults. We report promising results and lay the groundwork to build a modular and robust framework for building, evaluating, and improving agents for autonomous clouds.

CORNET 2.0: A Co-Simulation Middleware forRobot Networks

We present a networked co-simulation framework for multi-robot systems applications. We require a simulation framework that captures both physical interactions and communications aspects to effectively design such complex systems. This is necessary to co-design the multi-robots' autonomy logic and the communication protocols. The proposed framework extends existing tools to simulate the robot's autonomy and network-related aspects. We have used Gazebo with ROS/ROS2 to develop the autonomy logic for robots and mininet-WiFi as the network simulator to capture the cyber-physical systems properties of the multi-robot system. This framework addresses the need to seamlessly integrate the two simulation environments by synchronizing mobility and time, allowing for easy migration of the algorithms to real platforms.

Heuristic Algorithms for Co-scheduling of Edge Analytics and Routes for UAV Fleet Missions

Unmanned Aerial Vehicles (UAVs) or drones are increasingly used for urban applications like traffic monitoring and construction surveys. Autonomous navigation allows drones to visit waypoints and accomplish activities as part of their mission. A common activity is to hover and observe a location using on-board cameras. Advances in Deep Neural Networks (DNNs) allow such videos to be analyzed for automated decision making. UAVs also host edge computing capability for on-board inferencing by such DNNs. To this end, for a fleet of drones, we propose a novel Mission Scheduling Problem (MSP) that co-schedules the flight routes to visit and record video at waypoints, and their subsequent on-board edge analytics. The proposed schedule maximizes the utility from the activities while meeting activity deadlines as well as energy and computing constraints. We first prove that MSP is NP-hard and then optimally solve it by formulating a mixed integer linear programming (MILP) problem. Next, we design two efficient heuristic algorithms, JSC and VRC, that provide fast sub-optimal solutions. Evaluation of these three schedulers using real drone traces demonstrate utility-runtime trade-offs under diverse workloads.

Holistic Measures for Evaluating Prediction Models in Smart Grids

The performance of prediction models is often based on "abstract metrics" that estimate the model's ability to limit residual errors between the observed and predicted values. However, meaningful evaluation and selection of prediction models for end-user domains requires holistic and application-sensitive performance measures. Inspired by energy consumption prediction models used in the emerging "big data" domain of Smart Power Grids, we propose a suite of performance measures to rationally compare models along the dimensions of scale independence, reliability, volatility and cost. We include both application independent and dependent measures, the latter parameterized to allow customization by domain experts to fit their scenario. While our measures are generalizable to other domains, we offer an empirical analysis using real energy use data for three Smart Grid applications: planning, customer education and demand response, which are relevant for energy sustainability. Our results underscore the value of the proposed measures to offer a deeper insight into models' behavior and their impact on real applications, which benefit both data mining researchers and practitioners.