Kanad Ray

Remaining useful life (RUL) prediction is central to prognostics and health management of turbomachinery, but deep models are often large and opaque. We re-evaluate a compact quantum-classical hybrid (a 1D convolutional encoder, a four-qubit variational circuit with eight trainable angles and a small regression head; 8237 parameters) on all four NASA C-MAPSS sub-datasets. Validation uses held-out engines instead of held-out windows, which we show leaks information, and results are averaged over several seeds. Controls without the circuit, a frozen random circuit, parameter-matched LSTM and GRU models and ridge regression are included. The hybrid reaches mean test RMSE values of 19.05, 20.78, 18.50 and 30.82 cycles on FD001–FD004 (19.16 and 25.88 on FD002 and FD004 with condition-wise normalisation). It is indistinguishable from the classical controls on FD001–FD003 and worse on FD004, and matched-size recurrent models or ridge regression are better on single-condition data. Simulated finite-shot sampling with 256 shots costs about 0.3 cycles. SHAP attribution points to compressor-path sensors, with moderate rank stability across seeds. The study provides a controlled testbed rather than evidence of quantum advantage.

Background: This paper aims to systematically review, based on a set of 125 papers, smart-home automation with IoT from the perspective of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The main evidence synthesis was conducted on empirical studies published from 2024 to 2026, with background studies as a supplement. Methods: IEEE Xplore, Scopus, SpringerLink, ACM Digital Library, and other databases were used in a systematic search. There were 81 empirical studies published from January 2024 to March 2026 that were included in the systematic evidence synthesis; 44 additional foundational and background works were added separately to provide a theoretical and contextual base, resulting in a corpus of 125 works. Due to significant methodological variation, results were summarized qualitatively rather than by meta-analysis, employing a structured framework of evidence tiers. Results: Real-world validation was limited across the reviewed corpus; reported savings were typically in the 15–25% range, with the majority of reported savings from simulations, testbeds, and/or short-term evaluations. Based on our review, six major gaps are identified: seamless interoperability, lack of real-world validation, weak security and privacy, generalizability, standardized benchmarks, and user trust. Discussion: Overall, the research roadmap and recommendations are given in an action-oriented, future-looking way for both industry and policymakers.

Photocatalytic materials form an important class of quantum materials in which the electronic structure governs charge-carrier behaviour and reaction pathways through quantum-mechanical processes. To accelerate the discovery of superior photocatalytic quantum materials, we present a quantum-enhanced machine learning framework that integrates many-body electronic-structure descriptors with ensemble machine-learning algorithms. The framework explicitly incorporates quantum coherence lengths, electron–phonon coupling strengths, exciton binding energies, and spin–orbit coupling parameters obtained from density functional theory (DFT) and many-body perturbation theory GW approximation and the Bethe–Salpeter equation (GW-BSE) calculations. The database is based on a screened library of 8642 photocatalytic materials that have been experimentally characterised in H 2 evolution reactions and contains 152 descriptors relating to electronic correlations, quantum confinement effects, and spin-dependent properties. Five-fold cross-validation yields R 2 = 0.82 and MAE = 0.28 log units, n 18% relative improvement in prediction accuracy ( R 2 : 0.82 vs 0.69) over a baseline model using only classical descriptors ( R 2 = 0.69 , MAE = 0.38 log units). High-throughput screening of 45 000 hypothetical compositions yielded 214 candidate quantum materials; DFT and GW-BSE validation on representative subset of 23 of these candidates showed 72% agreement within quantum uncertainty bounds. quantum coherence shows a statistically significant association with photocatalytic charge separation, with optimal coherence lengths of 15–35 nm being strongly correlated with increased hydrogen-evolution production. Our findings indicate that many-body interactions, including exciton dissociation dynamics and quantum tunnelling contributions to interfacial charge transfer, are predictive of photocatalytic activity in emerging quantum materials.

Healthcare environments present uniquely demanding constraints for the deployment of artificial intelligence (AI). Clinical decisions of significant consequence are rarely the product of a single isolated computation; rather, safe and effective patient care relies on the continuous collaboration of distributed entities, including clinicians, hospital networks, laboratory systems, monitoring devices, and patients themselves. As generative models and large language models are increasingly deployed in clinical settings, the transition from monolithic AI models to Multi‐Agent Systems (MAS) offers unprecedented capabilities for distributed monitoring, coordinated decision support, and service integration. However, multi‐agent collaboration simultaneously introduces severe novel vulnerabilities, including topological contamination, privacy leakage, cascading systemic failures, and the socio‐psychological risks of undue algorithmic persuasion. This work introduces a comprehensive, privacy‐preserving, explainable, and human‐centred MAS framework explicitly designed for collaborative clinical decision support. Rejecting the paradigm of a single, omniscient AI model with unrestricted data access, the proposed architecture operationalises clinical AI as a heavily regulated team of specialised, interacting agents. The framework defines six distinct agent classes: Data Stewardship, Modality, Explanation, Consensus, Safety Monitor, and Clinician Interface agents. Furthermore, the implementation is governed by six foundational design principles grounded in classic Autonomous Agents and MAS theory, adapted for modern generative systems. By enforcing minimum‐necessary data exchange, verifiable team‐level explainability, explicit hierarchical escalation, secure communication topologies, and human‐centred safety protocols, this framework provides both a domain‐grounded blueprint and a transferable set of robust design principles for deploying trustworthy MAS in high‐stakes environments.

Diagnosis of cancer is one of the hardest problems faced in modern medicine and involves integrating different data sources such as medical images, genomic profiles and clinical records. Traditional machine learning methods have difficulty handling the high-dimensional and complex correlation properties of multimodal medical data. In view of this, we propose a new Quantum-Enhanced Multimodal Fusion Network (QEMFN) framework to break through traditional image–text matching based on quantum computing principles for CT imaging with genomic sequencing data and EHR information. Our approach utilizes variational quantum circuits for feature encoding, quantum kernel methods for crossmodal attention, and hybrid quantum–classical architectures for final classification. We realize the framework using Google Cirq quantum computing library and validate it on publicly available datasets including TCIA (The Cancer Imaging Archive), TCGA (The Cancer Genome Atlas), and MIMIC-III clinical database. The matched multimodal cohort comprises 847 lung cancer patients, 623 colorectal cancer patients, and 401 liver cancer patients with complete imaging, genomic, and clinical records, assembled via de-identified patient ID linkage across the three archives. The experiment takes steps toward the realization of quantum-enhanced diagnostic systems and offers a path for subsequent experimental confirmation. We theoretically analyze the potential quantum advantage, present detailed implementation details using Cirq, and describe a roadmap to clinical translation for quantum-enhanced diagnostic tools.

Despite significant progress in general-purpose diffusion-based models capable of producing high-quality media, this approach is still too difficult to implement on consumer/gamer hardware. We present here a memory-optimized DreamBooth framework designed for consumer-grade GPUs with 16 GB of VRAM, that allows for end-to-end image personalization and addresses some of the limitations of existing solutions. Our system reduces peak GPU memory from 22 GB (baseline DreamBooth) to 14.2 GB through novel hierarchical memory management, including attention slicing, Variational Autoencoder (VAE) tiling, gradient accumulation, and gradient checkpointing integrated within the Hugging Face Accelerate ecosystem. The framework further incorporates state-of-the-art techniques for preserving facial features and a comprehensive automated quality management system. The result is a complete end-to-end pipeline achieving a peak memory of 14.2 GB, with quantitative performance (LPIPS: 0.139, SSIM: 0.879, identity: 0.852, and FID: 23.1) competitive with methods requiring significantly more hardware resources.