arXiv subjects

Odej Kao

Publications and source records attributed to Odej Kao.

At least 37 records · Page 2Linked to original sources

arXiv2025

Efficiency Will Not Lead to Sustainable Reasoning AI

AI research is increasingly moving toward complex problem solving, where models are optimized not only for pattern recognition but for multi-step reasoning. Historically, computing's global energy footprint has been stabilized by sustained efficiency gains and natural saturation thresholds in demand. But as efficiency improvements are approaching physical limits, emerging reasoning AI lacks comparable saturation points: performance is no longer limited by the amount of available training data but continues to scale with exponential compute investments in both training and inference. This paper argues that efficiency alone will not lead to sustainable reasoning AI and discusses research and policy directions to embed explicit limits into the optimization and governance of such systems.

cs.AI↗

arXiv2025

What happens when nanochat meets DiLoCo?

Although LLM training is typically centralized with high-bandwidth interconnects and large compute budgets, emerging methods target communication-constrained training in distributed environments. The model trade-offs introduced by this shift remain underexplored, and our goal is to study them. We use the open-source nanochat project, a compact 8K-line full-stack ChatGPT-like implementation containing tokenization, pretraining, fine-tuning, and serving, as a controlled baseline. We implement the DiLoCo algorithm as a lightweight wrapper over nanochat's training loop, performing multiple local steps per worker before synchronization with an outer optimizer, effectively reducing communication by orders of magnitude. This inner-outer training is compared against a standard data-parallel (DDP) setup. Because nanochat is small and inspectable, it enables controlled pipeline adaptations and allows direct comparison with the conventional centralized baseline. DiLoCo achieves stable convergence and competitive loss in pretraining but yields worse MMLU, GSM8K, and HumanEval scores after mid-training and SFT. We discover that using DiLoCo-pretrained weights and running mid- and post-training with DDP fails to recover performance, revealing irreversible representation drift from asynchronous updates that impairs downstream alignment. We provide this implementation as an official fork of nanochat on GitHub.

cs.DC↗

arXiv2025

Comparative Analysis of Large Language Models for the Machine-Assisted Resolution of User Intentions

Large Language Models (LLMs) have emerged as transformative tools for natural language understanding and user intent resolution, enabling tasks such as translation, summarization, and, increasingly, the orchestration of complex workflows. This development signifies a paradigm shift from conventional, GUI-driven user interfaces toward intuitive, language-first interaction paradigms. Rather than manually navigating applications, users can articulate their objectives in natural language, enabling LLMs to orchestrate actions across multiple applications in a dynamic and contextual manner. However, extant implementations frequently rely on cloud-based proprietary models, which introduce limitations in terms of privacy, autonomy, and scalability. For language-first interaction to become a truly robust and trusted interface paradigm, local deployment is not merely a convenience; it is an imperative. This limitation underscores the importance of evaluating the feasibility of locally deployable, open-source, and open-access LLMs as foundational components for future intent-based operating systems. In this study, we examine the capabilities of several open-source and open-access models in facilitating user intention resolution through machine assistance. A comparative analysis is conducted against OpenAI's proprietary GPT-4-based systems to assess performance in generating workflows for various user intentions. The present study offers empirical insights into the practical viability, performance trade-offs, and potential of open LLMs as autonomous, locally operable components in next-generation operating systems. The results of this study inform the broader discussion on the decentralization and democratization of AI infrastructure and point toward a future where user-device interaction becomes more seamless, adaptive, and privacy-conscious through locally embedded intelligence.

cs.SE↗

arXiv2025

QONNECT: A QoS-Aware Orchestration System for Distributed Kubernetes Clusters

Modern applications increasingly span across cloud, fog, and edge environments, demanding orchestration systems that can adapt to diverse deployment contexts while meeting Quality-of-Service (QoS) requirements. Standard Kubernetes schedulers do not account for user-defined objectives such as energy efficiency, cost optimization, and global performance, often leaving operators to make manual, cluster-by-cluster placement decisions. To address this need, we present QONNECT, a vendor-agnostic orchestration framework that enables declarative, QoS-driven application deployment across heterogeneous Kubernetes and K3s clusters. QONNECT introduces a distributed architecture composed of a central Knowledge Base, Raft-replicated Resource Lead Agents, and lightweight Resource Agents in each cluster. Through a minimal YAML-based interface, users specify high-level QoS goals, which the system translates into concrete placement and migration actions. Our implementation is evaluated on a federated testbed of up to nine cloud-fog-edge clusters using the Istio Bookinfo microservice application. The system demonstrates dynamic, policy-driven microservice placement, automated failover, QoS-compliant rescheduling, and leader re-election after node failure, all without manual intervention. By bridging the gap between declarative deployment models and operational QoS goals, QONNECT transforms the cloud-edge continuum into a unified, self-optimizing platform.

cs.DC↗

arXiv2025

Distributed Low-Communication Training with Decoupled Momentum Optimization

The training of large models demands substantial computational resources, typically available only in data centers with high-bandwidth interconnects. However, reducing the reliance on high-bandwidth interconnects between nodes enables the use of distributed compute resources as an alternative to centralized data center training. Building on recent advances in distributed model training, we propose an approach that further reduces communication by combining infrequent synchronizations across distributed model replicas with gradient momentum compression. In particular, we treat the optimizer momentum as a signal and decompose the Nesterov momentum into high- and low-frequency components via the discrete cosine transform (DCT). Only the high-frequency components are synchronized across model replicas every $H$ steps. Empirically, our method achieves up to a $16\times$ reduction in communication compared to the baseline DiLoCo, and it generalizes across architectures, including transformer-based language models and convolutional neural networks for images. Overall, this work advances the feasibility of training large models on distributed nodes with low-bandwidth interconnects.

cs.LG↗

arXiv2025

Optimizing Microgrid Composition for Sustainable Data Centers

As computing energy demand continues to grow and electrical grid infrastructure struggles to keep pace, an increasing number of data centers are being planned with colocated microgrids that integrate on-site renewable generation and energy storage. However, while existing research has examined the tradeoffs between operational and embodied carbon emissions in the context of renewable energy certificates, there is a lack of tools to assess how the sizing and composition of microgrid components affects long-term sustainability and power reliability. In this paper, we present a novel optimization framework that extends the computing and energy system co-simulator Vessim with detailed renewable energy generation models from the National Renewable Energy Laboratory's (NREL) System Advisor Model (SAM). Our framework simulates the interaction between computing workloads, on-site renewable production, and energy storage, capturing both operational and embodied emissions. We use a multi-horizon black-box optimization to explore efficient microgrid compositions and enable operators to make more informed decisions when planning energy systems for data centers.

cs.DC↗

arXiv2025

Empirical Evidences for the Effects of Feature Diversity in Open Set Recognition and Continual Learning

Open set recognition (OSR) and continual learning are two critical challenges in machine learning, focusing respectively on detecting novel classes at inference time and updating models to incorporate the new classes. While many recent approaches have addressed these problems, particularly OSR, by heuristically promoting feature diversity, few studies have directly examined the role that feature diversity plays in tackling them. In this work, we provide empirical evidence that enhancing feature diversity improves the recognition of open set samples. Moreover, increased feature diversity also facilitates both the retention of previously learned data and the integration of new data in continual learning. We hope our findings can inspire further research into both practical methods and theoretical understanding in these domains.

cs.CV↗

arXiv2025

Moving Beyond Marginal Carbon Intensity: A Poor Metric for Both Carbon Accounting and Grid Flexibility

Marginal Carbon Intensity (MCI) has been promoted as an effective metric for carbon-aware computing. Although it is already considered as impractical for carbon accounting purposes, many still view it as valuable when optimizing for grid flexibility by incentivizing electricity usage during curtailment periods. In this statement paper, we argue that MCI is neither reliable nor actionable for either purpose. We outline its fundamental limitations, including non-observability, reliance on opaque predictive models, and the lack of verifiability. Moreover, MCI fails to reflect curtailment caused by high-carbon sources and offers no insight into the quantity of available excess power. We advocate moving beyond MCI and instead call for research on more actionable metrics, such as direct reporting of excess power, explicit modeling of energy storage and grid stability, and integration with emerging granular renewable energy certificate markets.

eess.SY↗

arXiv2025

Quantifying the Energy Consumption and Carbon Emissions of LLM Inference via Simulations

The environmental impact of Large Language Models (LLMs) is rising significantly, with inference now accounting for more than half of their total lifecycle carbon emissions. However, existing simulation frameworks, which are increasingly used to determine efficient LLM deployments, lack any concept of power and, therefore, cannot accurately estimate inference-related emissions. We present a simulation framework to assess the energy and carbon implications of LLM inference under varying deployment setups. First, we extend a high-fidelity LLM inference simulator with a GPU power model that estimates power consumption based on utilization metrics, enabling analysis across configurations like batch size, sequence length, and model parallelism. Second, we integrate simulation outputs into an energy system co-simulation environment to quantify carbon emissions under specific grid conditions and explore the potential of carbon-aware scheduling. Through scenario-based analysis, our framework reveals how inference parameters affect energy demand and carbon footprint, demonstrates a renewable offset potential of up to 69.2% in an illustrative deployment case, and provides a foundation for future carbon-aware inference infrastructure design.

cs.DC↗

arXiv2025

Choosing the Right Battery Model for Data Center Simulations

As demand for computing resources continues to rise, the increasing cost of electricity and anticipated regulations on carbon emissions are prompting changes in data center power systems. Many providers are now operating compute nodes in microgrids, close to renewable power generators and energy storage, to maintain full control over the cost and origin of consumed electricity. Recently, new co-simulation testbeds have emerged that integrate domain-specific simulators to support research, development, and testing of such systems in a controlled environment. Yet, choosing an appropriate battery model for data center simulations remains challenging, as it requires balancing simulation speed, realism, and ease of configuration. In this paper, we implement four different battery models for data center scenarios within the co-simulation framework Vessim and analyze their behavior. The results show that linear models, which consider inefficiencies and power limits, closely match the behavior of complex physics-based models in short-term experiments while offering faster execution, and not requiring knowledge on electrochemical reactions and circuit-level dynamics. In contrast, simple, lossless models fail to accurately represent complex behavior and provide no further runtime advantage.

cs.DC↗

arXiv2025

Towards Machine-Generated Code for the Resolution of User Intentions

The growing capabilities of Artificial Intelligence (AI), particularly Large Language Models (LLMs), prompt a reassessment of the interaction mechanisms between users and their devices. Currently, users are required to use a set of high-level applications to achieve their desired results. However, the advent of AI may signal a shift in this regard, as its capabilities have generated novel prospects for user-provided intent resolution through the deployment of model-generated code. This development represents a significant progression in the realm of hybrid workflows, where human and artificial intelligence collaborate to address user intentions, with the former responsible for defining these intentions and the latter for implementing the solutions to address them. In this paper, we investigate the feasibility of generating and executing workflows through code generation that results from prompting an LLM with a concrete user intention, and a simplified application programming interface for a GUI-less operating system. We provide an in-depth analysis and comparison of various user intentions, the resulting code, and its execution. The findings demonstrate the general feasibility of our approach and that the employed LLM, GPT-4o-mini, exhibits remarkable proficiency in the generation of code-oriented workflows in accordance with provided user intentions.

cs.AI↗

arXiv2025

Informed Mixing -- Improving Open Set Recognition via Attribution-based Augmentation

Open set recognition (OSR) is devised to address the problem of detecting novel classes during model inference. Even in recent vision models, this remains an open issue which is receiving increasing attention. Thereby, a crucial challenge is to learn features that are relevant for unseen categories from given data, for which these features might not be discriminative. To facilitate this process and "optimize to learn" more diverse features, we propose GradMix, a data augmentation method that dynamically leverages gradient-based attribution maps of the model during training to mask out already learned concepts. Thus GradMix encourages the model to learn a more complete set of representative features from the same data source. Extensive experiments on open set recognition, close set classification, and out-of-distribution detection reveal that our method can often outperform the state-of-the-art. GradMix can further increase model robustness to corruptions as well as downstream classification performance for self-supervised learning, indicating its benefit for model generalization.

cs.CV↗

arXiv2025

Predicting the Performance of Scientific Workflow Tasks for Cluster Resource Management: An Overview of the State of the Art

Scientific workflow management systems support large-scale data analysis on cluster infrastructures. For this, they interact with resource managers which schedule workflow tasks onto cluster nodes. In addition to workflow task descriptions, resource managers rely on task performance estimates such as main memory consumption and runtime to efficiently manage cluster resources. Such performance estimates should be automated, as user-based task performance estimates are error-prone. In this book chapter, we describe key characteristics of methods for workflow task runtime and memory prediction, provide an overview and a detailed comparison of state-of-the-art methods from the literature, and discuss how workflow task performance prediction is useful for scheduling, energy-efficient and carbon-aware computing, and cost prediction.

cs.DC↗

arXiv2025

$β$-GNN: A Robust Ensemble Approach Against Graph Structure Perturbation

Graph Neural Networks (GNNs) are playing an increasingly important role in the efficient operation and security of computing systems, with applications in workload scheduling, anomaly detection, and resource management. However, their vulnerability to network perturbations poses a significant challenge. We propose $β$-GNN, a model enhancing GNN robustness without sacrificing clean data performance. $β$-GNN uses a weighted ensemble, combining any GNN with a multi-layer perceptron. A learned dynamic weight, $β$, modulates the GNN's contribution. This $β$ not only weights GNN influence but also indicates data perturbation levels, enabling proactive mitigation. Experimental results on diverse datasets show $β$-GNN's superior adversarial accuracy and attack severity quantification. Crucially, $β$-GNN avoids perturbation assumptions, preserving clean data structure and performance.

cs.LG↗

arXiv2025

Flora: Efficient Cloud Resource Selection for Big Data Processing via Job Classification

Distributed dataflow systems like Spark and Flink enable data-parallel processing of large datasets on clusters of cloud resources. Yet, selecting appropriate computational resources for dataflow jobs is often challenging. For efficient execution, individual resource allocations, such as memory and CPU cores, must meet the specific resource demands of the job. Meanwhile, the choices of cloud configurations are often plentiful, especially in public clouds, and the current cost of the available resource options can fluctuate. Addressing this challenge, we present Flora, a low-overhead approach to cost-optimizing cloud cluster configurations for big data processing. Flora lets users categorize jobs according to their data access patterns and derives suitable cluster resource configurations from executions of test jobs of the same category, considering current resource costs. In our evaluation on a new dataset comprising 180 Spark job executions on Google Cloud, Flora's cluster resource selections exhibit an average deviation below 6% from the most cost-optimal solution, with a maximum deviation below 24%.

cs.DC↗

arXiv2025

Sizey: Memory-Efficient Execution of Scientific Workflow Tasks

As the amount of available data continues to grow in fields as diverse as bioinformatics, physics, and remote sensing, the importance of scientific workflows in the design and implementation of reproducible data analysis pipelines increases. When developing workflows, resource requirements must be defined for each type of task in the workflow. Typically, task types vary widely in their computational demands because they are simply wrappers for arbitrary black-box analysis tools. Furthermore, the resource consumption for the same task type can vary considerably as well due to different inputs. Since underestimating memory resources leads to bottlenecks and task failures, workflow developers tend to overestimate memory resources. However, overprovisioning of memory wastes resources and limits cluster throughput. Addressing this problem, we propose Sizey, a novel online memory prediction method for workflow tasks. During workflow execution, Sizey simultaneously trains multiple machine learning models and then dynamically selects the best model for each workflow task. To evaluate the quality of the model, we introduce a novel resource allocation quality (RAQ) score based on memory prediction accuracy and efficiency. Sizey's prediction models are retrained and re-evaluated online during workflow execution, continuously incorporating metrics from completed tasks. Our evaluation with a prototype implementation of Sizey uses metrics from six real-world scientific workflows from the popular nf-core framework and shows a median reduction in memory waste over time of 24.68% compared to the respective best-performing state-of-the-art baseline.

cs.DC↗

arXiv2025

Experimentally Evaluating the Resource Efficiency of Big Data Autoscaling

Distributed dataflow systems like Spark and Flink enable data-parallel processing of large datasets on clusters. Yet, selecting appropriate computational resources for dataflow jobs is often challenging. For efficient execution, individual resource allocations, such as memory and CPU cores, must meet the specific resource requirements of the job. An alternative to selecting a static resource allocation for a job execution is autoscaling as implemented for example by Spark. In this paper, we evaluate the resource efficiency of autoscaling batch data processing jobs based on resource demand both conceptually and experimentally by analyzing a new dataset of Spark job executions on Google Dataproc Serverless. In our experimental evaluation, we show that there is no significant resource efficiency gain over static resource allocations. We found that the inherent conceptual limitations of such autoscaling approaches are the inelasticity of node size as well as the inelasticity of the ratio of memory to CPU cores.

cs.DC↗

arXiv2024

KS+: Predicting Workflow Task Memory Usage Over Time

Scientific workflow management systems enable the reproducible execution of data analysis pipelines on cluster infrastructures managed by resource managers such as Kubernetes, Slurm, or HTCondor. These resource managers require resource estimates for each workflow task to be executed on one of the cluster nodes. However, task resource consumption varies significantly between different tasks and for the same task with different inputs. Furthermore, resource consumption also fluctuates during a task's execution. As a result, manually configuring static memory allocations is error-prone, often leading users to overestimate memory usage to avoid costly failures from under-provisioning, which results in significant memory wastage. We propose KS+, a method that predicts a task's memory consumption over time depending on its inputs. For this, KS+ dynamically segments the task execution and predicts the memory required for each segment. Our experimental evaluation shows an average reduction in memory wastage of 38% compared to the best-performing state-of-the-art baseline for two real-world workflows from the popular nf-core repository.

cs.DC↗