Energy-Efficient ML

As machine learning (ML) continues its rapid expansion, the environmental cost of model training and inference has become a critical societal concern. Existing benchmarks overwhelmingly focus on standard performance metrics such as accuracy, BLEU, or mAP, while largely ignoring energy consumption and carbon emissions. This single-objective evaluation paradigm is increasingly misaligned with the practical requirements of large-scale deployment, particularly in energy-constrained environments such as mobile devices, developing regions, and climate-aware enterprises. In this paper, we propose AI-CARE, an evaluation tool for reporting energy consumption, and carbon emissions of ML models. In addition, we introduce the carbon-performance tradeoff curve, an interpretable tool that visualizes the Pareto frontier between performance and carbon cost. We demonstrate, through theoretical analysis and empirical validation on representative ML workloads, that carbon-aware benchmarking changes the relative ranking of models and encourages architectures that are simultaneously accurate and environmentally responsible. Our proposal aims to shift the research community toward transparent, multi-objective evaluation and align ML progress with global sustainability goals. The tool and documentation are available at https://github.com/USD-AI-ResearchLab/ai-care.

Early-exit deep neural networks enable adaptive inference by terminating computation when sufficient confidence is achieved, reducing cost for edge AI accelerators in resource-constrained settings. Existing methods, however, rely on suboptimal exit policies, ignore input difficulty, and optimize thresholds independently. This paper introduces DART (Input-Difficulty-Aware Adaptive Threshold), a framework that overcomes these limitations. DART introduces three key innovations: (1) a lightweight difficulty estimation module that quantifies input complexity with minimal computational overhead, (2) a joint exit policy optimization algorithm based on dynamic programming, and (3) an adaptive coefficient management system. Experiments on diverse DNN benchmarks (AlexNet, ResNet-18, VGG-16) demonstrate that DART achieves up to \textbf{3.3$\times$} speedup, \textbf{5.1$\times$} lower energy, and up to 42% lower average power compared to static networks, while preserving competitive accuracy. Extending DART to Vision Transformers (LeViT) yields power (5.0$\times$) and execution-time (3.6$\times$) gains but also accuracy loss (up to 17 percent), underscoring the need for transformer-specific early-exit mechanisms. We further introduce the Difficulty-Aware Efficiency Score (DAES), a novel multi-objective metric, under which DART achieves up to a 14.8 improvement over baselines, highlighting superior accuracy, efficiency, and robustness trade-offs.

As the burgeoning power requirements of sophisticated neural architectures escalate, the information retrieval community has recognized ecological sustainability as a pivotal priority that necessitates a fundamental paradigm shift in model design. While contemporary neural rankers have attained unprecedented accuracy, the substantial environmental externalities associated with their computational intensity often remain overlooked in large-scale deployments. We present GaiaFlow, an innovative framework engineered to facilitate carbon-frugal search by operationalizing semantic-guided diffusion tuning. Our methodology orchestrates the convergence of retrieval-guided Langevin dynamics and a hardware-independent performance modeling strategy to optimize the trade-off between search precision and environmental preservation. By incorporating adaptive early exit protocols and precision-aware quantized inference, the proposed architecture significantly mitigates operational carbon footprints while maintaining robust retrieval quality across heterogeneous computing infrastructures. Extensive experimental evaluations demonstrate that GaiaFlow achieves a superior equilibrium between effectiveness and energy efficiency, offering a scalable and sustainable pathway for next-generation neural search systems.

Knowledge distillation (KD) is a technique to compress a larger teacher system into a smaller student. In machine translation, KD is commonly evaluated through translation quality and inference efficiency, without jointly accounting for the environmental costs of producing and deploying the distilled system. We evaluate representative KD methods both on bespoke MT models and LLMs, by considering both translation quality and computational cost, using the Machine Learning Life Cycle Assessment tool, which accounts for costs throughout the KD model life cycle. Our key finding is that the deployment volume required to amortize KD is serving-dependent and can shift by several orders of magnitude under batching. We include actionable guidance for selecting, developing, and evaluating KD methods under quality and compute-induced constraints.

The training and deployment of machine learning (ML) models have become extremely energy-intensive. While existing optimization efforts focus primarily on hardware energy efficiency, a significant but overlooked source of inefficiency is software energy waste caused by poor software design. This often includes redundant or poorly designed operations that consume more energy without improving performance. These inefficiencies arise in widely used ML frameworks and applications, yet developers often lack the visibility and tools to detect and diagnose them. We propose differential energy debugging, a novel approach that leverages the observation that competing ML systems often implement similar functionality with vastly different energy consumption. Building on this insight, we design and implement Magneton, an energy profiler that compares energy consumption between similar ML systems at the operator level and automatically pinpoints code regions and configuration choices responsible for excessive energy use. Applied to 9 popular ML systems spanning LLM inference, general ML frameworks, and image generation, Magneton detects and diagnoses 16 known cases of software energy inefficiency and further discovers 8 previously unknown cases, 7 of which have been confirmed by developers.

Federated Learning (FL) holds the potential to advance equality in health by enabling diverse institutions to collaboratively train deep learning (DL) models, even with limited data. However, the significant resource requirements of FL often exclude centres with limited computational infrastructure, further widening existing healthcare disparities. To address this issue, we propose a Green AI-oriented adaptive layer-freezing strategy designed to reduce energy consumption and computational load while maintaining model performance. We tested our approach using different federated architectures for Magnetic Resonance Imaging (MRI)-to-Computed Tomography (CT) conversion. The proposed adaptive strategy optimises the federated training by selectively freezing the encoder weights based on the monitored relative difference of the encoder weights from round to round. A patience-based mechanism ensures that freezing only occurs when updates remain consistently minimal. The energy consumption and CO2eq emissions of the federation were tracked using the CodeCarbon library. Compared to equivalent non-frozen counterparts, our approach reduced training time, total energy consumption and CO2eq emissions by up to 23%. At the same time, the MRI-to-CT conversion performance was maintained, with only small variations in the Mean Absolute Error (MAE). Notably, for three out of the five evaluated architectures, no statistically significant differences were observed, while two architectures exhibited statistically significant improvements. Our work aligns with a research paradigm that promotes DL-based frameworks meeting clinical requirements while ensuring climatic, social, and economic sustainability. It lays the groundwork for novel FL evaluation frameworks, advancing privacy, equity and, more broadly, justice in AI-driven healthcare.

Federated learning (FL) is a powerful paradigm for distributed learning, but increasing model complexity leads to significant energy consumption from client-side computations for local training. This challenge is critical in energy-harvesting FL (EHFL) systems, where the participation availability of each device fluctuates because of limited energy. To address this, we propose PipeCycle, a battery-aware distributed learning framework that organizes clients into pipelined cyclic groups. When a group completes its intra-group aggregation, its aggregated model is relayed directly to a newly formed group as a reference for local training, allowing multiple groups to coexist in the pipeline while overlapping client recharging periods with active training in other pipeline stages. We provide a convergence analysis of PipeCycle under a realistic energy consumption model in which local training spans multiple time slots, and show that the cyclic structure of the pipeline imposes a finite-horizon staleness bound that avoids the exponential factors typical of asynchronous FL analyses. Numerical experiments across both IID and non-IID data and various battery charging probabilities show that PipeCycle reaches a target accuracy with substantially lower cumulative energy than existing FL baselines, particularly under severe label skew where competing cyclic schemes collapse to near-chance accuracy.

Backpropagation (BP) powers modern deep learning, but it must hold every activation in memory until a global backward sweep consumes it, and that sweep locks each layer until the one above has finished. Forward-only algorithms, from Hinton's Forward-Forward (FF) through Cascaded Forward (CaFo) to Mono-Forward (MF), replace the global gradient with local learning signals and are promoted as more efficient on the basis of FLOP counts rather than measurements. We test that claim on hardware. Across almost 1000 runs on four vision benchmarks (MNIST, Fashion-MNIST, CIFAR-10 and CIFAR-100), we read energy, power and memory from an NVIDIA A100 GPU and compare each algorithm against a tuned BP baseline. Removing the backward pass does not by itself save energy: FF costs 4.1x to 6.3x BP's energy and CaFo 3.0x to 7.1x. Mono-Forward is the exception. Given the same number of data passes as BP, it uses 12.5% to 19.9% less energy and 24.6% to 27.5% less memory on the 3x2000 multilayer perceptrons, the largest configurations tested, at lower power and near-equal wall-clock time; on CIFAR-100 its test accuracy is statistically equivalent to BP's. A pre-registered ablation identifies gradient detachment, not MF's auxiliary losses, as the dominant factor tested. Trained to convergence, MF exceeds BP's test accuracy on CIFAR-10 and CIFAR-100 by 3.90 and 5.04 percentage points, but needs 3.5x to 4.5x as many passes and 3.2x to 4.0x the energy. Faster convergence is therefore the remaining target for an end-to-end gain.

Image- and audio-based remote IoT applications offer a substantial opportunity to address ecological changes. However, to be effective, they must operate for long periods on limited power while also providing data via LPWAN. Previous studies have investigated the energy efficiency of isolated components such as radio transmission or data processing methods within an IoT system. However, a holistic evaluation is required to determine the main drivers of energy consumption in sensor nodes. To address this, we systematically analyze energy consumption across the entire sensor-node pipeline, from data acquisition to optional on-device CNN classification and wireless network transmission, using an ESP32-S3 microcontroller. We compare on-device CNN classification followed by result transmission with transmitting raw sensor data across three environmental monitoring scenarios. Thereby, we vary developer-controllable variables, notably the model architecture and the LPWAN technologies (LTE-M, NB-IoT, LoRaWAN). Our results indicate that on-device inference can reduce sensor node energy consumption by 1.7x to 6.9x compared to raw data transmission, depending on transmission frequency, with an acceptable loss in accuracy. Further, our results narrow down the energy tradeoff to variables IoT developers can control. Therefore, our findings provide empirical guidance for designing energy-efficient, long-range sensing systems and demonstrate that TinyML enables sustainable, autonomous deployments.

The field of artificial intelligence (AI) has taken a tight hold on broad aspects of society, industry, business, and governance in ways that dictate the prosperity and might of the world's economies. The AI market size is projected to grow from {$}189 billion in 2023 to {$}4.8 trillion by 2033. Currently, AI is dominated by large language models (LLMs) that exhibit linguistic and visual intelligence. However, training these models requires a massive amount of data scraped from the web as well as large amounts of energy (50-60 GWh to train GPT-4). Despite these costs, these models often hallucinate, a characteristic that prevents them from being deployed in critical application domains. In contrast, the human brain consumes only 20W of power. What is needed is the next level of AI evolution in which lightweight domain-specific multimodal models, especially compact models with 10--20B parameters for bounded domains, with higher levels of intelligence can reason, plan, and make decisions in dynamic environments with real-time data and prior knowledge, while learning continuously and evolving in ways that enhance future decision-making capability. This will define the next wave of AI, progressing from today's large models, trained with vast amounts of data, to nimble energy-efficient domain-specific agents that can reason and think in a world full of uncertainty. To support such agents, hardware will need to be reimagined to allow system-level energy efficiencies $\geq {1000X}$ over the state of the art for targeted domain tasks, subject to accuracy, latency, and coverage constraints. Such a vision of future AI systems is developed in this work.

Recent Machine Learning (ML) approaches have shown increased performance on benchmarks at the cost of escalating compute demands. Hardware, algorithmic and carbon optimizations have been proposed to curb energy use and environmental impacts. We estimate the environmental impacts associated with training models documented in the Epoch AI database over the last decade, with a particular focus on impacts associated with Large Language Models and the hardware used to train them. We find that energy use and environmental impacts associated with training ML models have increased exponentially, even when considering impact reduction strategies such as using less carbon intensive electricity mixes or more efficient hardware. Optimization strategies do not mitigate the impacts induced by model training, suggesting rebound effect. We show that the impacts of hardware must be considered over the entire life cycle rather than the sole use phase in order to avoid impact shifting. Our study demonstrates that increasing efficiency alone does not ensure sustainability. There is an urgent need to systematically integrate environmental impacts in NLP evaluation practices to better inform the community and support the use of impact as a feature in research planning and decision making.

The energy cost of Large Language Model (LLM) inference is rapidly becoming a barrier to sustainable and scalable deployment. Although modern serving architectures expose distinct prefill and decode behaviors, existing systems fail to exploit these phase differences for energy-efficient serving under strict latency SLOs. This paper introduces VoltanaLLM, the first system that explicitly targets and reduces the energy bloat in modern prefill-decode (P/D) disaggregated LLM serving. Guided by a control-theory perspective, VoltanaLLM separates two levers: per-instance operating-point selection (GPU frequency per iteration) and system-level state-space routing of requests. We empirically observe that LLM inference exhibits a U-shaped energy-frequency curve creating "sweet spots" that depend on phase behavior and load. VoltanaLLM exploits this by combining phase-specific, iteration-level frequency selection driven by a lightweight, online-adaptive latency predictor, with a decode state-space guided router that avoids architectural granularity-induced inefficiencies, all while meeting desired SLOs. We implement VoltanaLLM using SGLang and evaluate it across multiple models and real-world workloads. Our results show VoltanaLLM reduces end-to-end energy by up to 36.3% versus a static max-frequency baseline while maintaining high SLO attainment, and generalizes to newer GPUs. These results point to sustainable LLM serving via phase-aware, iteration-level frequency selection coupled with architecture-aware routing. Source code is available in https://github.com/Supercomputing-System-AI-Lab/VoltanaLLM.

Sleep stage classification is critical for diagnosing and managing disorders like sleep apnea and insomnia. However, conventional methods like polysomnography are costly and impractical for long-term, home-based monitoring. This study presents an energy-efficient approach for detecting four sleep stages (wake, rapid eye movement (REM), light sleep, deep sleep) using a single-lead electrocardiogram (ECG) signal. We evaluate various machine learning and deep learning models, introducing two windowing strategies: (1) a 5-minute window with 30-second steps for machine learning and (2) a 30-second window with 10-second steps for deep learning, enabling 10-second temporal resolution for real-time predictions. While deep learning models like MobileNet-v1 achieve high accuracy (92%) and F1-score (91%), their energy demands make them unsuitable for wearables. To address this, we design SleepLiteCNN, optimized for ECG-based sleep staging, achieving 89% accuracy and 89% F1-score while minimizing energy use. Applying 8-bit quantization further reduces energy consumption to 5.48 microJ per inference, with 90% accuracy and F1-score. Additionally, field-programmable gate array (FPGA) deployment shows significant reductions in resource usage. This approach provides a practical, energy-efficient solution for continuous ECG-based sleep monitoring in resource-constrained wearable devices.

Predicting future observations plays a central role in machine learning, biology, economics, and many other fields. It lies at the heart of organizational principles such as the variational free energy principle and, based on the second law of thermodynamics, has even been shown to be necessary for reaching the fundamental energetic limits of information processing on a tape. While the usefulness of the predictive paradigm is undisputed, complex adaptive systems that interact with their environment are more than just predictive machines: they have the power to act upon their environment and cause change. In this work, we develop a framework to analyze the thermodynamics of information processing in percept-action loops, a model of agent-environment interaction, allowing us to investigate the thermodynamic implications of actions and percepts on equal footing. To this end, we introduce the concept of work capacity, defined as the maximum rate at which an agent can expect to extract work from its environment. Our results reveal that work-efficient agents must balance prediction and forgetting. This highlights a fundamental departure from the thermodynamics of passive observation, suggesting that prediction and energy efficiency may be at odds in active learning systems.

Modern applications such as autonomous vehicles, intelligent surveillance, and smart city systems increasingly require object detection on resource-constrained edge devices. Yet, there is still limited understanding of how different object detection models behave across heterogeneous edge devices and under varying scene complexity. In this paper, we benchmark YOLOv8 (Nano, Small, Medium), EfficientDet Lite (Lite0, Lite1, Lite2), and SSD (SSD MobileNet V1, SSDLite MobileDet) on Raspberry Pi 3, 4, 5 with/without Coral TPU accelerators, Raspberry Pi 5 with AI HAT+, Jetson Nano, and Jetson Orin Nano. We evaluate energy consumption, inference time, and accuracy, and further examine how accuracy changes with the number of objects in the input image. The results reveal clear trade-offs among accuracy, latency, and energy efficiency across model-device combinations. SSD MobileNet V1 achieves the lowest latency and energy consumption but the lowest accuracy, whereas YOLOv8 Medium achieves the highest accuracy at higher computational cost. TPU-based Raspberry Pi devices improve the efficiency of SSD and EfficientDet Lite while reducing YOLOv8 accuracy. Orin Nano offers the most favorable overall balance across most model families. The object-count-based analysis further shows that models achieve more similar accuracy on simpler images, while the accuracy gap widens as scene complexity increases.

Spike-based encoders represent information as sequences of spikes or pulses, which are transmitted between neurons. A prevailing consensus suggests that spike-based approaches demonstrate exceptional capabilities in capturing the temporal dynamics of neural activity and have the potential to provide energy-efficient solutions for low-power applications. The Spiketrum encoder efficiently compresses input data using spike trains or code sets (for non-spiking applications) and is adaptable to both hardware and software implementations, with lossless signal reconstruction capability. The paper proposes and assesses Spiketrum's hardware, evaluating its output under varying spike rates and its classification performance with popular spiking and non-spiking classifiers, and also assessing the quality of information compression and hardware resource utilization. The paper extensively benchmarks both Spiketrum hardware and its software counterpart against state-of-the-art, biologically-plausible encoders. The evaluations encompass benchmarking criteria, including classification accuracy, training speed, and sparsity when using encoder outputs in pattern recognition and classification with both spiking and non-spiking classifiers. Additionally, they consider encoded output entropy and hardware resource utilization and power consumption of the hardware version of the encoders. Results demonstrate Spiketrum's superiority in most benchmarking criteria, making it a promising choice for various applications. It efficiently utilizes hardware resources with low power consumption, achieving high classification accuracy. This work also emphasizes the potential of encoders in spike-based processing to improve the efficiency and performance of neural computing systems.

The emergence of Transformer-based Large Language Models (LLMs) has substantially augmented the capabilities of Natural Language Processing (NLP), thereby intensifying the demand for computational resources. Therefore, enhancing efficiency based on factors like computational requirements, energy consumption, carbon footprint and financial cost has become a vital area of research. This motivates us to conduct a survey on Transformer-based LLMs in NLP from the perspective of efficiency. In this survey of 312 articles, the efficiency-improvement endeavors have been systematically discussed targeting various aspects such as data curation, model design, model downsizing, and dynamic inferencing. This has been augmented with efficiency considerations in model adaptation strategies like pre-training, fine-tuning, prompt-engineering and Retrieval-Augmented Generation (RAG). Furthermore, a statistical analysis followed by an in-depth evaluation of the efficiency and efficacy of more than 30 renowned NLP models has been performed on 13 evaluation benchmarks. This paper offers valuable insights for researchers, professionals as well as scholars, and explores the trend of research toward sustainable practices in NLP.