Search Results (70)

Electric vehicle (EV) adoption is rising rapidly, creating growing challenges for charging infrastructure planning, energy demand management, and grid stability. However, most existing studies rely on single-domain data, such as behavioral charging sessions or station metadata, which limits their ability to capture the joint effects of user behavior, charger characteristics, and market context. To address this gap, this study proposes a multi-domain machine learning framework for EV charger-type prediction by integrating behavioral, infrastructure, and market-level data. Behavioral charging logs are transformed into structured event-token sequences and modeled using XLM-RoBERTa (Cross-lingual Language Model–RoBERTa), which is used here as a transformer-based sequence encoder to capture long-range dependencies in charging behavior. Structured infrastructure and market features are modeled using LightGBM and TabNet. The study contributes a unified multi-domain framework, a systematic comparison of transformer and tabular-learning models, and a broader evaluation through ablation analysis, cross-validation, confusion matrix analysis, and confidence calibration. The results show that multi-domain fusion consistently improves performance over single-domain learning. XLM-RoBERTa achieved the best overall performance on the fused dataset, with 98.76% accuracy and 97.86% weighted F1-score, while TabNet demonstrated stronger calibration and deployment reliability. Full article

Large language model (LLM) serving has emerged as a major source of energy consumption in modern AI infrastructure. In current deployments, graphics processing units (GPUs) are typically operated at default high-frequency settings to maximize performance. However, under practical service-level objectives (SLOs), peak performance is often unnecessary, especially during the memory-bound decode stage, resulting in substantial power redundancy and avoidable energy waste. Existing studies that apply GPU dynamic voltage and frequency scaling (DVFS) to improve the energy efficiency of LLM serving have shown promising results. However, they generally rely on coarse-grained control, accurate output length prediction, or request-level resource management, which limits their effectiveness under highly dynamic workloads and strict SLO constraints. We present EcoInfer, a fine-grained DVFS framework for energy-efficient LLM serving. EcoInfer performs iteration-level, workload-aware GPU frequency control that adapts to the current inference phase and system state while preserving latency guarantees. It comprises three tightly integrated modules: a machine-learning-based frequency–latency predictor that estimates iteration latency across candidate GPU frequencies using lightweight iteration-level features; an SLO-aware frequency controller that selects the minimum feasible frequency within a sweet-spot-guided candidate range; and a low-overhead runtime optimization layer that combines adaptive decision caching with asynchronous execution to reduce and hide the overhead of online control. Implemented on top of vLLM, EcoInfer achieves up to 25.4% energy savings and 21.5% average energy savings and improves energy efficiency by 1.28× on average in terms of Tokens/J while maintaining a nearly unchanged SLO attainment rate compared with the default vLLM baseline. Full article

The transition toward carbon-neutral cities and sustainable infrastructure requires massive capital mobilization, yet traditional static valuation models like discounted cash flow (DCF) systematically undervalue green projects due to high initial capital expenditures and long-term uncertainty. To address this critical gap in sustainable finance, this study proposes a novel Artificial Intelligence–Blockchain–Multiple Real Options (AI-MRO) integrated framework. This model aligns infrastructure profitability with Environmental, Social, and Governance (ESG) criteria and United Nations Sustainable Development Goals (SDGs), specifically SDG 11 (Sustainable Cities), SDG 13 (Climate Action), and SDG 9 (Industry, Innovation, and Infrastructure). The core approach integrates AI-based probabilistic forecasting for carbon footprint optimization and cash flow prediction, MRO-based operational flexibility assessment, and blockchain-based smart contracts (Security Token Offerings, STOs) to ensure transparent green finance governance and social inclusion. Through empirical validation at Singapore’s Punggol Digital District (PDD)—a flagship smart city project featuring a district-level smart grid reducing 1700 tonnes of CO₂ and generating 3000 MWh of solar energy annually—this model successfully captured investment resilience (Extended Net Present Value, ENPV > 0) even in crisis scenarios where conventional DCF models failed. The results demonstrate that integrating digital twins and AI-driven ESG metrics structurally reduces the risk premium and amplifies the strategic value of sustainable investments. This study represents a substantial methodological contribution toward data-driven, automated, and transparent governance, offering a scalable financial framework for global net-zero infrastructure development. Full article

Environmental, Social, and Governance (ESG) text classification is important for applications in sustainable finance. However, it remains a challenging task due to domain terminology and regulatory constraints. While transformer-based models achieve strong predictive performance, they often lead to high energy costs and provide limited interpretability. To address these limitations, we introduce ESG-Graph, a lightweight and interpretable graph-based framework for modeling ESG disclosures. In our approach, each sentence is represented as a token-level dependency graph augmented with virtual nodes initialized from a European Sustainability Reporting Standards (ESRS)-based taxonomy, enabling the addition of new ESG concepts without retraining. A multi-layer Graph Attention Network is used instead of transformer encoders, allowing grammatical structure and domain semantics to be modeled jointly. Experiments on three ESG benchmark datasets show that ESG-Graph achieves performance comparable to efficient transformer baselines while consuming up to 60× less energy and using 10× fewer parameters. Additional attribution and ablation studies suggest the method’s policy alignment, interpretability, and robustness. Full article

Cloud cover significantly reduces the electrical power output of photovoltaic systems, making accurate short-term cloud movement predictions essential for reliable solar energy production planning. This article presents a deep learning framework that directly estimates cloud movement from ground-based all-sky camera images, rather than predicting future production from past power data. The system is based on a three-step process: First, a lightweight Convolutional Neural Network segments cloud regions and produces probabilistic masks that represent the spatial distribution of clouds in a compact and computationally efficient manner. This allows subsequent models to focus on the geometry of clouds rather than irrelevant visual features such as illumination changes. Second, a Vector Quantized Variational Autoencoder compresses these masks into discrete latent token sequences, reducing dimensionality while preserving fundamental cloud structure patterns. Third, a GPT-style autoregressive transformer learns temporal dependencies in this token space and predicts future sequences based on past observations, enabling iterative multi-step predictions, where each prediction serves as the input for subsequent time steps. Our evaluations show an average intersection-over-union ratio of 0.92 and a pixel accuracy of 0.96 for single-step (5 s ahead) predictions, while performance smoothly decreases to an intersection-over-union ratio of 0.65 and an accuracy of 0.80 in 10 min autoregressive propagation. The framework also provides prediction uncertainty estimates through token-level entropy measurement, which shows positive correlation with prediction error and serves as a confidence indicator for downstream decision-making in solar energy forecasting applications. Full article

Internet of Things (IoT) applications often require only minimal necessary information—such as threshold judgments, binning, or prefixes—yet they must control privacy leakage arising from multi-round and cross-entity access without exposing raw values. Existing solutions, however, frequently rely on ciphertext structures and server-side states, making it difficult to define a leakage upper bound for restricted answers in the sense of Differential Privacy (DP), or they lack unified information budgeting and k-use control. To address these challenges, this paper proposes a verifiable differential privacy partial disclosure scheme for IoT. We employ DP accounting to uniformly constrain the leakage of three types of operators: threshold, binning, and prefix. Furthermore, we design stateless k-use tokens based on Verifiable Random Functions (VRFs) and chained receipts to generate publicly verifiable compliance evidence for each response. We implemented an end-edge-cloud prototype system and evaluated its performance on two use cases: smart meter threshold alarms and industrial sensor out-of-bound detection. Experimental results demonstrate that compared with a baseline relying on server-state counting for k-use control, our stateless k-use mechanism improves throughput by approximately 25–37% under concurrency scales of 1, 8, and 16, and reduces p95 latency by an average of 15%. Meanwhile, in multi-party splicing attack experiments, the re-identification accuracy remains stable in the 0.50–0.52 range, approximating random guessing. These results validate that the proposed scheme possesses low-energy engineering feasibility and audit-friendliness while effectively suppressing splicing risks. Full article

The transition of Large Language Models (LLMs) from centralized clouds to edge environments is critical for addressing privacy concerns, latency bottlenecks, and operational costs. However, existing edge benchmarking frameworks remain tailored to discriminative Deep Learning tasks (e.g., object detection), failing to capture the multidimensional challenges of generative AI, specifically the trade-offs between token generation speed, semantic accuracy, and hardware sustainability. To address this gap, we introduce LEAF (LLM Edge Assessment Framework), a novel evaluation methodology that integrates Circular Economy principles directly into performance metrics. LEAF assesses edge deployments across five synergistic pillars: Circular Economy Score, Energy Efficiency (Joules/Token), Performance Speed (Tokens/Second), semantic accuracy (BERTScore), and End-to-End Latency. We validate LEAF through an extensive experimental analysis of five distinct hardware classes, ranging from embedded IoT devices (Raspberry Pi 4 and 5, NVIDIA Jetson Nano) to professional edge servers (NVIDIA T400) and repurposed legacy workstations (NVIDIA GTX 1050 Ti). Utilizing 4-bit quantized models via the Ollama runtime, our results reveal a counterintuitive insight: repurposed consumer hardware significantly outperforms modern purpose-built edge SoCs. The legacy GTX 1050 Ti achieved a 20× speedup over the Raspberry Pi 4 and maintained superior energy-per-task efficiency compared to low-power ARM architectures by minimizing active runtime. These findings challenge the prevailing narrative that newer silicon is essential for Edge AI, demonstrating that sustainable, high-performance inference can be achieved by extending the lifecycle of existing hardware. LEAF thus provides a blueprint for a “Green Edge” ecosystem that balances computational capability with environmental responsibility. Full article

The accelerating adoption of electric vehicles (EVs) is intensifying pressure on urban power grids, particularly during evening peak hours. Existing smart-charging frameworks remain constrained by centralized control, static pricing, and limited integration of predictive intelligence. This study presents SMARGE, a hybrid AI–Blockchain smart charging platform that combines load forecasting, dynamic pricing, and cryptocurrency-based incentives to enhance decentralized EV energy management in Gaziantep Province. An ensemble of forecasting models (SARIMA, LightGBM, N-BEATS, and TFT) predicts 2026 hourly electricity demand, while an adaptive inverse-sigmoid pricing mechanism generates real-time incentives and disincentives for EV charging behavior. A fuzzy logic-based behavioral model simulates both unmanaged and managed charging across three scenarios. Results show that managed charging reduces peak load by 22.43%, shifts 67.45% of energy demand to off-peak periods, and achieves 94.86% charging fulfillment under constrained grid conditions. The blockchain layer—implemented through a custom ERC-20 token (SMARGE) on the Ethereum Sepolia testnet—enables secure, transparent, and low-cost microtransactions with an average confirmation time of 0.63 s. These findings demonstrate that tightly coupling AI forecasting with tokenized blockchain incentives can improve grid stability, lower operational costs, and enhance user autonomy in a scalable and decentralized manner. While promising, the study is limited by assumptions of synthetic user behavior and ideal communication conditions; future work will validate the platform in real-world pilot deployments and across different urban regions. Full article

Digital transformation in water conservancy and urban water utilities demands perception systems that are accurate, fast, and energy-efficient and maintainable over long service lifecycles at the edge. We present HydroSNN, a neuromorphic computer-vision framework that couples an event-driven sensing pipeline with a spiking-transformer backbone to support monitoring of canals, reservoirs, treatment plants, and buried pipeline networks. By reducing always-on compute and unnecessary data movement, HydroSNN targets sustainability goals in smart water infrastructure: lower operational energy use, fewer site visits, and improved resilience under harsh illumination and weather. HydroSNN introduces three novel components: (i) spiking temporal tokenization (STT), which converts asynchronous events and optional frames into latency-aware spike tokens while preserving motion cues relevant to hydraulics; (ii) physics-guided spiking attention (PGSA), which injects lightweight mass-conservation/continuity constraints into attention weights via a differentiable regularizer to suppress physically implausible interactions; and (iii) cross-modal self-supervision (CM-SSL), which aligns RGB frames, event streams, and low-cost acoustic/vibration traces using masked prediction to reduce annotation requirements. We evaluate HydroSNN on public water-surface and event-vision benchmarks (MaSTr1325, SeaDronesSee, DSEC, MVSEC, DAVIS, and DDD20) and report accuracy, latency, and an operation-based energy proxy. HydroSNN improves mIoU/F1 over strong CNN/ViT baselines while reducing end-to-end latency and the estimated energy proxy in event-driven settings. These efficiency gains are practically relevant for off-grid or power-constrained deployments and support sustainable development by enabling continuous, low-power monitoring and timely anomaly response. These results demonstrate that event-driven spiking vision, augmented with simple physics guidance, offers a practical and efficient solution for resilient perception in smart water infrastructure. Full article

This paper introduces AdamN, a nested-momentum adaptive optimizer that replaces the single Exponential Moving Average (EMA) numerator in Adam/AdamW with a compounded EMA of gradients plus an EMA of that EMA, paired with an exact double-EMA bias correction. This yields a smoother, curvature-aware search direction at essentially first-order cost, with longer, more faithful gradient-history memory and a stable, warmup-free start. Under comparable wall-clock time per epoch, AdamN matches AdamW’s final accuracy on ResNet-18/CIFAR-100, while reaching 80% and 90% training-accuracy milestones ~127 s and ~165 s earlier, respectively. On pre-benchmarking workloads (toy problems and CIFAR-10), AdamN shows the same pattern: faster early-phase convergence with similar or slightly better final accuracy. On language modeling with token-frequency imbalance—Wikitext-2-style data with training-only token corruption and a 10% low-resource variant—AdamN lowers rare-token perplexity versus AdamW without warmup while matching head and mid-frequency performance. In full fine-tuning of Llama 3.1–8B on a small dataset, AdamN reaches AdamW’s final perplexity in roughly half the steps (≈$2.25\times$ faster time-to-quality). Finally, on a ViT-Base/16 transferred to CIFAR-100 (batch size 256), AdamN achieves 88.8% test accuracy vs. 84.2% for AdamW and reaches 40–80% validation-accuracy milestones in the first epoch (AdamW reaches 80% by epoch 59), reducing epochs, energy use, and cost. Full article

We examine whether aggregate “music mood” derived from globally popular songs can help forecast primary equity issuance. We build a Friday-anchored weekly panel that merges SEC EDGAR counts of priced Initial Public Offerings (IPOs) with features from the Spotify Daily Top 200 (audio descriptors such as valence, energy, danceability, tempo, loudness, etc.) and Genius-scraped lyrics. We extract lyric sentiment by tokenizing Genius-scraped lyrics and aggregating lexicon-based affect scores (valence and arousal) into popularity-weighted weekly indices. To address sparsity and regime shifts in issuance, we train a leakage-safe Long Short-Term Memory (LSTM) network on a smoothed target—the forward 4-week sum of IPOs—and obtain next-week forecasts by dividing the predicted sum by 4. On a chronological holdout, a single LSTM with look-back K = 8 outperforms strong baselines—reducing MAE by 13.9%, RMSE by 15.9%, and mean Poisson deviance by 27.6% relative to the best baseline in each metric. Furthermore, we adopt SHapley Additive exPlanations (SHAP) to explain our LSTM model, showing that IPO persistence remains the dominant driver, but music and lyrics covariates contribute incremental and robust signal. These results suggest that aggregate music sentiment contains economically meaningful information about near-term IPO activity. Full article

With the large-scale integration of power electronic converters, non-linear loads, and renewable energy generation, voltage and current waveform distortion in modern power systems has become increasingly severe, making harmonic resonance amplification and non-stationary distortion more prominent. Accurate and robust harmonic-level prediction and detection have become essential foundations for power quality monitoring and operational protection. However, traditional harmonic analysis methods remain highly dependent on pre-designed time–frequency transformations and manual feature extraction. They are sensitive to noise interference and operational variations, often exhibiting performance degradation under complex operating conditions. To address these challenges, a Unified Physics-Transformer-based harmonic detection scheme is proposed to accurately forecast harmonic levels in offshore wind farms (OWFs). This framework utilizes real-world wind speed data from Bozcaada, Turkey, to drive a high-fidelity electromagnetic transient simulation, constructing a benchmark dataset without reliance on generative data expansion. The proposed model features a Feature Tokenizer to project continuous physical quantities (e.g., wind speed, active power) into high-dimensional latent spaces and employs a Multi-Head Self-Attention mechanism to explicitly capture the complex, non-linear couplings between meteorological inputs and electrical states. Crucially, a Multi-Task Learning (MTL) strategy is implemented to simultaneously regress the Total Harmonic Distortion (THD) and the characteristic 5th Harmonic (H5), effectively leveraging shared representations to improve generalization. Comparative experiments with Random Forest, LSTM, and GRU systematically evaluate the predictive performance using metrics such as root mean square error (RMSE) and mean absolute percentage error (MAPE). Results demonstrate that the Physics-Transformer significantly outperforms baseline methods in prediction accuracy, robustness to operational variations, and the ability to capture transient resonance events. This study provides a data-efficient, high-precision approach for harmonic forecasting, offering valuable insights for future renewable grid integration and stability analysis. Full article

Seconds count differently for people in danger. We present a real-time streaming pipeline for audio-based detection of hazardous life events affecting life and property. The system operates online rather than as a retrospective analysis tool. Its objective is to reduce the latency between the occurrence of a crime, conflict, or accident and the corresponding response by authorities. The key idea is to map reality as perceived by audio into a written story and question the text via a large language model. The method integrates streaming, zero-shot algorithms in an online decoding mode that convert sound into short, interpretable tokens, which are processed by a lightweight language model. CLAP text–audio prompting identifies agitation, panic, and distress cues, combined with conversational dynamics derived from speaker diarization. Lexical information is obtained through streaming automatic speech recognition, while general audio events are detected by a streaming version of Audio Spectrogram Transformer tagger. Prosodic features are incorporated using pitch- and energy-based rules derived from robust F0 tracking and periodicity measures. The system uses a large language model configured for online decoding and outputs binary (YES/NO) life-threatening risk decisions every two seconds, along with a brief justification and a final session-level verdict. The system emphasizes interpretability and accountability. We evaluate it on a subset of the X-Violence dataset, comprising only real-world videos. We release code, prompts, decision policies, evaluation splits, and example logs to enable the community to replicate, critique, and extend our blueprint. Full article

The article presents a blockchain-based architecture for decentralized electricity trading that tokenizes energy delivery rights and cash-flows. Energy Attribute Certificates (EACs) are implemented as NFTs, while buy/sell orders are encoded as ERC-1155 tokens whose tokenId packs a time slot and price, enabling precise matching across hours. A clearing smart contract (Matcher) burns filled orders, mints an NFT option, and issues two ERC-20 assets: PT, the right to consume kWh within a specified interval, and YT, the producer’s claim on revenue. We propose a simple, linearly increasing discounted buyback for YT within the slot and introduce an aggregating token, IndexYT, which accumulates YTs across slots, redeems them at par at maturity, and gradually builds on-chain reserves—turning IndexYT into a liquid, yield-bearing instrument. We outline the PT/YY lifecycle, oracle-driven policy controls for DSO (e.g., transfer/splitting constraints), and discuss transparency, resilience, and capital efficiency. The contribution is a Pendle-inspired split of electricity into Principal/Yield tokens combined with a time-stamped on-chain order book and IndexYT, forming a programmable market for short-term delivery rights and yield derivatives with deterministic settlement. Full article

Spiking Neural Networks (SNNs) have drawn increasing attention for their event-driven and energy-efficient characteristics. However, achieving accurate and efficient inference within limited time-steps remains a major challenge. This paper proposes an efficient spiking Transformer framework that integrates cross-time-step knowledge distillation, multi-scale resolution processing, and attention-based token pruning to enhance both temporal modeling and energy efficiency. The cross-time-step distillation mechanism enables earlier time steps to learn from later ones, which improves inference consistency and accuracy, leading to better performance. Meanwhile, the multi-scale processing module dynamically adjusts input resolution and reuses features across scales, while the attention-based token pruning adaptively removes redundant tokens to reduce computational overhead. Extensive experimental results on static datasets (CIFAR-10/100 and ImageNet-1K) and dynamic event-based datasets (DVS128-Gesture and CIFAR10-DVS) demonstrate that the proposed method achieves higher accuracy and more than 1.4× inference speedup compared to baseline SNN–Transformer models. This framework provides a promising solution for developing energy-efficient and high-performance neuromorphic vision systems. Full article