Search Results (143)

Resource-constrained traffic classification requires mapping heterogeneous numerical features into a fixed-length binary input. Equal-width encoding ignores differences in discriminability, whereas range-only allocation may spend precision on weak predictors. We propose discriminability-aware symmetric bit-wise feature encoding (DA-SBFE), which combines training-fold mutual information with a robust numerical range to allocate tiered feature precision under a fixed budget. P4-programmable SmartNICs serve as a deployment-oriented case study because constrained per-packet processing motivates compact, deterministic representations. We do not claim hardware deployment. On 83,868 packets from the official UNSW IoT traces, DA-SBFE with BNN-Large achieved 0.9330 ± 0.0019 macro-F1 at 256 bits versus 0.9185 ± 0.0020 for Range-SBFE, with positive differences in all five folds. The gain increased from 0.0144 at 256 bits to 0.0348 at 192 bits and 0.0556 at 128 bits, and remained at 0.0176 after removing port features. Cross-day transfer showed only a small positive mean difference (0.0032) with mixed seed-level directions, so stable temporal robustness was not established. Deterministic integer XNOR–popcount inference matched PyTorch for all 16,774 held-out packets. The evidence supports DA-SBFE as a software-validated encoding method. Compilation, resource-use measurement, throughput evaluation, and latency evaluation remain future work. Full article

Generating a virtual anchor from speech has to satisfy three demands at once: the face must stay recognizable as the same person, the motion has to be temporally coherent, and the lips must follow the audio. Existing methods often fall short on the first two points because identity and expression share one entangled representation, and temporal dynamics are modeled only implicitly. We propose DynaID-VAE to address these problems. At its core is an identity–expression disentangled conditional VAE (DC-VAE) that splits the latent space into a time-varying expression subspace and a static identity subspace, held apart by mutual-information minimization and orthogonality regularization. A temporal memory module (TMM) then regularizes the expression trajectory: a GRU propagates sequential state, attention retrieves from a learnable key–value prototype memory, and residual fusion combines the two. Multiscale adversarial supervision and lip–audio synchronization losses complete the training objective. We evaluate on VirtualAnchor-100, a benchmark we recorded ourselves (100 h, 10 anchors), under two complementary protocols. Cross-identity driving is scored only with non-paired measures, namely lip synchronization, distributional video quality, and identity preservation; full-reference image metrics are confined to a self-reenactment protocol, where a genuine paired ground truth exists. DynaID-VAE outperforms the one-reference baselines Wav2Lip, PC-AVS, SadTalker, and DiffTalk under both protocols and on unseen VoxCeleb2 identities. The margins are stable across five identity-disjoint, nested cross-validation folds and are confirmed by an external SyncNet evaluator that never takes part in training, while the model runs at 41.2 FPS with 14.3 M parameters. Ablations separate the contribution of each regularizer and each TMM component. Linear and capacity-matched non-linear probes quantify the factorization as a large reduction of decodable reference identity; full independence is not claimed. A user study confirms the perceptual gains. Full article

Machine-learning intrusion detection is challenged by attacks resembling legitimate traffic and by class imbalance. This study evaluates Random Forest detection on the UNSW-NB15 dataset for five binary attack-versus-normal tasks: DoS, Exploit, Backdoor, Analysis, and Reconnaissance. Categorical attributes were label-encoded, features selected using mutual information, hyperparameters optimized by randomized search, decision thresholds tuned on validation data, and SMOTE applied to training data at a fixed target of 56,000 minority samples. SVM and KNN results are descriptive because the shared label-encoded representation precludes model-independent comparison. Without SMOTE, Random Forest achieved F1-scores of 0.9468 for DoS, 0.9504 for Exploit, 0.9839 for Backdoor, 0.9317 for Analysis, and 0.9720 for Reconnaissance. Fixed-SMOTE pipelines achieved 0.9539, 0.9516, 0.9668, 0.8370, and 0.9651, respectively. Thus, SMOTE improved DoS, marginally improved Exploit, and reduced performance for Backdoor, Analysis, and Reconnaissance. Confusion matrices showed fewer false positives and false negatives for DoS, while Analysis gained recall but produced substantially more false positives. These results compare separately optimized complete pipelines and do not isolate SMOTE’s effect. They characterize filtered target-attack-versus-normal streams, not operational mixed-attack traffic. Under this protocol, performance and precision–recall trade-offs were attack-dependent under the evaluated fixed-SMOTE experimental configuration, indicating that oversampling should be evaluated separately for each attack category. Full article

This paper investigates distributed optimization-assisted output consensus for nonlinear multi-agent systems with mutually incommensurate Caputo orders, unavailable velocity-like states, bounded disturbances, measurement noise, and actuator saturation. A mixed-power exact penalty flow generates practical optimal references from local costs and intermittent neighbor broadcasts. The penalty gain and a smoothing bias bound are determined from a public interval, topology information, and certified local gradient data without prior knowledge of the aggregate optimizer. An autonomous decaying threshold provides event-triggered communication, an initial condition-independent fixed-time practical certificate for the integer-order optimizer, and exclusion of finite-time event accumulation. The physical layer is analyzed with established Caputo quadratic inequalities and agentwise Mittag–Leffler comparison. Fractional reference and command filters, a composite observer, and two-gain anti-saturation compensation form the output feedback controller, while the physical result is formulated as a finite-horizon regional verification certificate. Numerical studies include same-model and communication budget comparisons, a recent method-inspired optimizer benchmark, certificate tightening, and robustness tests for initialization, the fractional order, measurement noise, and the integration step size. Full article

The Northeast black soil region is a major grain-producing area in China, where county-scale assessment of topsoil quality is essential for soil conservation and provides a potential indicator framework for regional soil quality monitoring. In this study, 500 composite cultivated-layer samples were collected from dryland croplands in Keshan County, Heilongjiang Province. Soil physical properties, basic chemical properties, and macro-, secondary, and micronutrient contents were measured to establish a total data set (TDS). Three hierarchical total data sets (TDS1, TDS2, and TDS3) were established from the original TDS according to the progressive incorporation of soil physical properties, basic chemical properties, macronutrients, secondary nutrients, and micronutrients. Within each hierarchical TDS, key indicators were selected using random forest (RF), mutual information (MI), principal component analysis (PCA), and minimum spanning tree (Tree) methods to construct algorithm-specific minimum data sets (MDSs). TDS1 included soil physical properties, basic chemical properties, and macronutrients; TDS2 further incorporated secondary nutrients; and TDS3 additionally included micronutrients to represent progressively comprehensive soil nutrient information. The soils exhibited considerable soil organic matter and cation exchange capacity, with mean values of 51.45 g kg⁻¹ and 34.85 cmolc kg⁻¹, respectively, and a mean pH of 6.04. Across the four algorithms, commonly retained indicators included SOM, TN, pH, CEC, and Silt, while nutrient-related indicators such as AP, AK, S, Zn, and Fe were additionally selected under different hierarchical MDSs, reflecting the importance of multi-nutrient information in soil quality characterization. Available phosphorus, available potassium, sulfur, and zinc showed greater spatial variability than basic physicochemical properties. The four algorithms differed in indicator selection, reflecting their distinct sensitivities to linear variation, information gain, multilevel contributions, and network structure. RF_MDS1 showed the highest agreement with the TDS (R² = 0.924), indicating its strong capability in preserving overall soil quality information using a simplified indicator set. RF_MDS3 maintained a high consistency with the TDS (R² = 0.836) while incorporating additional secondary nutrients and micronutrients. Therefore, RF_MDS3 was considered a more comprehensive MDS framework when multi-nutrient representation and potential nutrient constraint identification were prioritized, whereas RF_MDS1 remained an efficient option for simplified soil quality assessment. This framework may provide a transferable approach for soil quality assessment and nutrient management in comparable black-soil regions and dryland farming systems. Full article

Feature selection is a critical step in regression problems where a large number of continuous explanatory variables explain the same target through different dependency structures. Classical filters may remain sensitive to a single form of dependence, a single scale, or a specific discretization scheme; generalized entropy measures, on the other hand, typically require the parameters to be fixed at a single point. This study proposes a framework that evaluates the Sharma–Mittal entropy volumetrically across a two-dimensional parameter region rather than for a single parameter pair. For the continuous target and explanatory variables, the marginal, joint, and conditional densities are obtained using a Gaussian kernel density estimation; the conditional entropy and information gain surfaces are integrated across the region Ω = [0.05, 0.95]² in the α-β plane to define three indices: PICSME, which measures the conditional uncertainty volume; PIGSME, which measures the gain volume; and NIGSME, which is the ratio of this gain to the total entropy volume of the target. The method is supported by bandwidth consistency and the renormalization of conditional densities; thus, the issue of negative gain that can occur in the continuous variables is resolved, yielding positive and interpretable scores across all six datasets. It is formally demonstrated that the fact that the three indices produce the same ranking is not an empirical observation but rather the result of a monotonicity relationship valid under a fixed target entropy volume. The method is compared with Pearson and Spearman correlations, the Shannon information gain, mutual information, and random forest variable importance across six regression datasets (Airfoil Self-Noise, AirQualityUCI, BodyFat, Meteorology, Concrete, and WineQualityWhite) that differ in their sample size, dimensions, and application domain. The evaluation is not limited to ranking consistency; the out-of-sample prediction performance is measured using least-squares models on the top-k subsets, with rankings calculated from the training partition. The findings show that NIGSME exhibits a performance comparable to that of built-in filters, outperforms them on the Concrete and Meteorology datasets, and never ranks as the weakest method on any dataset. The results demonstrate that volumetric entropy metrics defined across the entire parameter space provide a feature-ranking tool that is independent of parameter selection for continuous variables. Full article

Stock markets are complex adaptive systems whose feature dependencies evolve over time, and identifying which signals genuinely drive returns is fundamentally a problem of information. Traditional feature selection methods and linear models rely on statistical correlations and overlook causality, which weakens their predictive performance in non-linear, evolving markets. This paper develops an information-theoretic approach to causal feature selection based on Markov Blanket discovery. We apply the Iterative Parent–Child-based search of Markov Blanket (IPCMB), whose conditional-independence tests are conditional mutual information measures, to recover the Markov Blanket of next-month returns—the minimal feature set that carries all Shannon information about the target. We then pair it with a Classification and Regression Tree (CART), whose impurity-based splitting admits an information-gain interpretation, to predict the direction of Chinese A-share returns non-linearly. Using data on 2760 stocks, IPCMB selects 13 causal features from 72 candidates, and CART forecasts whether the next month’s return is positive. The empirical results show an accuracy of 58.0%, 7.7 percentage points above the all-features CART benchmark, and a 13-month cumulative return 35.88% higher than that of the CSI 300 index. The findings indicate that selecting features according to their causal information content and combining them with interpretable tree-based prediction can support more reliable investment decisions in emerging markets. Full article

This paper, based on a keynote given at the Nordic Environmental and Sustainability Educators Conference, explores a set of principles that might inform the responses of environmental educators, in particular those working in Europe, to the growth in political authoritarianism allied with climate denial and ecological extractivism. Taking the form of a manifesto and intended to prompt debate and mutual support, the piece proposes 11 principles: Don’t panic; Acknowledge your own emotions; Explore Possibilities Beyond the Classroom; Work with hope as praxis; Proliferate stories of change; Nurture alternatives; Name and protect gains; Diversify resources and tactics; Practice simple sabotage; See the bigger picture; and Don’t forget the dancing. Full article

The current trajectory of Artificial Intelligence (AI) development represents a critical phase transition from a tenable academic pursuit to an untenable industrial behemoth, and ultimately toward an unsustainable environmental burden. This conceptual review and perspective article synthesizes evidence across sustainability science, information theory, AI governance, and regulatory studies to argue that Brute Force AI constitutes a systemic sustainability threat whose resolution requires a return to Algorithmic Parsimony. We formally redefine sustainability in sensu lato through a contrapositive logical criterion applicable to any system. We introduce an operationalized Intelligence-per-Joule ($\mathcal{I}/J$) index and Sustainability Index S, defined in terms of mutual information gain relative to thermodynamic and computational resource expenditure, and demonstrate their interpretive application across landmark AI systems. A quantitative synthesis of published empirical data across six landmark models (BERT through GPT-4) documents the Intelligence–Cost Divergence: the growing chasm between logarithmic capability gain and exponential resource cost. We further distinguish warranted from unwarranted scale—acknowledging that some large-scale capabilities are qualitatively irreplaceable—while arguing that the current default toward scale without efficiency justification is ecologically and socially indefensible. We introduce the Symbiotic Policy Covenant—a three-pillar governance framework encompassing Algorithmic Parsimony Mandates, Expanded Waste Taxonomy, and an AI Equity Safeguard addressing both access and benefit inequity—operationalized through a proposed ISO/IEC 42001-Plus standard addendum. We conclude that genuine intelligence and genuine sustainability are not in tension but are, at their mathematical foundations, the same aspiration. Full article

Class imbalance remains one of the most challenging problems in sentiment analysis, particularly in tourism review datasets where positive opinions substantially outnumber neutral and negative comments. This issue is especially critical because minority classes often contain the most valuable information regarding customer dissatisfaction, service failures, and opportunities for improvement. In this work, we propose a hybrid balancing methodology for sentiment analysis in Mexican Spanish tourism reviews that combines undersampling and Large Language Model (LLM)-based oversampling. The proposed framework first extracts representative words from each sentiment class using Mutual Information, then enriches them through dictionary-based or embedding-based lexical substitutions, and finally generates synthetic reviews using GPT-4o-mini guided by these representative terms. Experiments were conducted on a corpus that contains approximately 300,000 tourism reviews collected from TripAdvisor, exhibiting severe sentiment imbalance. Three undersampling strategies and multiple oversampling configurations were evaluated across six traditional machine learning classifiers and one Transformer-based model (BETO). Results show that random undersampling consistently outperformed centroid-based and K-means-based alternatives while also requiring the lowest computational cost. The best overall performance was obtained by BETO, achieving a Macro-F1 score of 0.57 compared to 0.51 on the original imbalanced dataset, representing an improvement of 11.8%. Significant gains were also observed for minority classes, with improvements exceeding 16% for the most underrepresented category. Furthermore, the proposed methodology consistently outperformed direct prompt-based generation using GPT-4o-mini, Gemini 2.5 Flash, and Llama 3.3 70B. These findings suggest that guiding synthetic review generation through representative words effectively preserves domain-specific lexical and semantic patterns of Mexican Spanish tourism reviews, resulting in more balanced datasets and improved sentiment classification performance. Full article

Machine learning-based network intrusion detection systems are often optimized using aggregate accuracy, although operational security depends on the reliable detection of rare, high-impact attacks. This paper proposes a data-preserving intrusion detection framework that combines hybrid feature selection, heterogeneous ensemble learning and cost-sensitive optimization for imbalanced multi-class attack detection. The method first applies Mutual Information filtering and Recursive Feature Elimination to reduce the NSL-KDD feature space from 122 one-hot encoded attributes to 25 discriminative features. Four classifiers, Random Forest, XGBoost, Support Vector Machine and K-Nearest Neighbors, are evaluated individually, and a stacking ensemble is constructed using Logistic Regression as a meta-learner. Class imbalance is addressed by balanced class weighting rather than by synthetic oversampling, preserving the original minority-class observations. Experiments on the NSL-KDD benchmark show that the proposed cost-sensitive configuration improves rare attack recognition, most notably increasing U2R recall from 0.00% to 35.82% (24 of 67 test instances) for the stacking ensemble; this improvement, together with the accompanying weighted F1-score change from 0.7120 to 0.7214, is statistically significant under the Wilcoxon signed-rank test across repeated random seeds, and both values are reported with their variability rather than as single point estimates. SVM obtains the largest global gain, with a 7.06 percentage point improvement in weighted F1-score. The results show that cost-sensitive learning is a simple and practical mechanism for improving rare-attack visibility, but also reveal a remaining limitation for R2L detection, where feature overlap with Normal traffic remains substantial. The revised validation design explicitly includes direct resampling baselines, repeated-seed evaluation, statistical significance testing, feature-subset sensitivity analysis, and absolute true-positive counts for R2L and U2R in order to avoid overinterpreting marginal point-estimate gains. All experiments, including the resampling comparison, the component ablation, the feature-subset sensitivity analysis and the repeated-seed statistical evaluation, are executed on the complete KDDTrain+ training set of 125,973 instances under a single unified protocol, so that every reported per-class value refers to the same experimental setting. The revised study additionally reports probability-level evaluation for the primary model, including class-level PR-AUC, precision-recall curves and a U2R threshold and alert-budget analysis, and validates the framework externally on the UNSW-NB15 benchmark, where balanced class weighting raises the recall of the rarest categories (Worms, Shellcode, Backdoor) from near-zero baseline levels to 69–96% under an identical protocol. Full article

Complex diseases often involve distributed interactions among biological regions, physiological systems, imaging phenotypes, and clinical variables that are not fully captured by anatomical proximity, isolated biomarkers, or conventional feature-based representations. In oncology, neuroimaging, critical care, and systems medicine, distant or apparently separate biomedical sectors may show strong statistical or functional coupling associated with multimodal imaging signatures, inflammatory responses, metabolic constraints, treatment-induced changes, or shared disease-state organization. In this work, we introduce a proof-of-concept relational graph framework for representing such candidate hidden connectivity in terms of correlation-induced accessibility bridges. The novelty of the framework is that it does not treat biomedical correlation, graph distance, and network connectivity as separate descriptors but explicitly couples non-factorizable inter-sector correlation to localized accessibility compression in an emergent disease-state geometry. The proposed framework represents a biomedical system as a weighted relational graph in which nodes correspond to clinically relevant entities, such as tissue regions, imaging-derived features, biomarker modules, physiological variables, or disease states, while weighted edges encode constraints on functional, statistical, or pathological accessibility. Within this structure, coarse-grained biomedical sectors are defined as organized subsystems, and non-factorizable coupling between sectors is quantified using mutual-information-type measures. Candidate biomedical bridges are then defined operationally as localized, high-gain reductions in effective inter-sector accessibility distance. We introduce explicit coupling rules linking sector-level correlation to bridge-specific accessibility compression, including an effective distance-compression model and an ensemble-based formulation. Numerical proof-of-concept simulations on randomized modular graph ensembles show that increasing correlation strength systematically reduces effective inter-sector distance and increases bridge gain. The strongest compression occurs when correlation modulates a designated bridge architecture, exceeding the effects observed under random non-bridge or generic inter-sector modulation. These simulations are not intended to validate a disease-specific biological mechanism but to test whether the proposed correlation–compression rule produces bridge-specific effects distinguishable from null graph perturbations. The resulting structures should not be interpreted as physical anatomical tunnels or direct causal pathways unless supported by additional biological evidence. Rather, they represent correlation-induced accessibility bridges: localized, high-gain routes in a patient- or disease-specific relational geometry. The framework may therefore provide a theoretical and computational basis for prioritizing candidate hidden connectivity patterns in radiomics, multimodal prognosis, physiological deterioration, recurrence modeling, and systems-level disease networks. Full article

Intrusion detection on in-vehicle networks requires high accuracy, which is reported by many papers so far, but also computational efficiency to make it suitable for real-world scenarios. The achievement of both requirements at the same time becomes harder to achieve, especially as the number of attacks diversifies. An approach to leverage computational costs is the use of sliding windows, i.e., batch processing, which extends the detection over multiple frames, but the use of multitask learning is also advantageous because a number of layers are shared between classes to extract common relevant features. While indeed the greatest computational gains are from the use of a sliding window, multitask learning has benefits too and is in fact necessary as multiple attack types can coexist in the same window. We explore the benefits of this approach on three existing attack datasets and we also build our own dataset that garners more attack complexity so that we can concretely measure the benefits of multitask learning both in terms of detection rate and computational savings. Our approach considers the feature-level similarity between attack types and legitimate frames, extracted from the mutual information between the two, and extends detection over windows of multiple frames, which justify multitask learning as frames belonging to different classes can co-exist in the same window. Full article

Induction motors are critical in modern industry, powering over 70% of industrial processes. Reliable operation is essential to minimize downtime and ensure production continuity. This paper proposes an integrated multimodal methodology for fault diagnosis and prognosis in induction motors, based on an extended Pearson and Gain feature fusion framework. The approach preprocesses vibration, current, voltage, torque, and speed signals through denoising, normalization, synchronization, and sliding-window segmentation. Over 200 features per window are extracted across time, frequency, envelope, wavelet, harmonic, slip-based, and MCSA domains. A key innovation is correlation-driven multimodal fusion, combining Pearson correlation, spectral coherence, cross-spectral energy, and mutual information to produce Gain-enhanced features with improved discriminative capability. Fault diagnosis is performed using RF, SVM, XGBoost, and MLP models, with time-aware data splitting to avoid temporal leakage. Prognosis employs a continuous Degradation Index (DI) modeled via Gaussian Process Regression for uncertainty-aware prediction, with failure probability and Remaining Useful Life (RUL) estimated from DI thresholds. Experimental results demonstrate that the proposed methodology achieves diagnostic accuracy above 97%, enhances feature relevance, and provides stable long-term prognostic performance, offering a robust framework for predictive maintenance of induction motors. Full article

The Transformer architecture drives modern Artificial Intelligence (AI), yet the physical principles that may constrain self-attention training remain poorly characterized. We develop a thermodynamic framework for attention training, drawing on the established Boltzmann correspondence between softmax attention and equilibrium statistical mechanics, and we propose a First Law analogue that decomposes the training energy budget into a heat term (the entropic cost of ordering attention) and a work term (the gain in mutual information about the target). From this framework we derive a Landauer-type bound on learning, which states that the loss reduction during training is bounded below by the entropic cost of structuring attention against thermal noise. The bound is satisfied across all configurations tested: 625 grid points spanning three datasets on a compact Vision Transformer trained from scratch (MNIST, CIFAR-10, and OrganAMNIST), and ten temperatures on a pretrained ViT-Small fine-tuned on Food-101. Reusing the same physical principles at inference time, we show that the thermodynamic work performed by each input patch provides a quantitative, energy-based measure of feature importance that outperforms standard attention weights and Integrated Gradients on ImageNet across pretrained ViT-Small, ViT-Base, and ViT-Large (22M to 304M parameters). The result is an integrated diagnostic framework that links phase structure, training-time bounds, and inference-time attribution within a single empirically falsifiable thermodynamic apparatus. Full article