Philip S. Yu and Osmar R. Zaïane succeed at a pivotal stage of Intelligent Computing's growth as an international open-access journal. The journal publishes innovative research on artificial intelligence, machine learning, and emerging interdisciplinary areas.
The new framework reduces processing time by an average of 87.7% compared to the state-of-the-art approach, achieving cache hit rates between 70% and 90%. It demonstrates linear scalability, handling high query loads efficiently.
A new security framework based on blockchain technology and distributed reinforcement learning ensures secure data storage and transmission while adapting to evolving threats. The framework demonstrated improved memory consumption and transaction latency compared to existing approaches.
Autograph, a new framework, uses graph neural networks and deep reinforcement learning to achieve higher accuracy and faster execution of compute-intensive programs. It outperformed other approaches across various datasets, with notable improvements on Polybench, NPB, and SPEC 2006 benchmarks.
A research team developed an optimization method for path planning in uncertain environments using deep reinforcement learning and action curiosity module. The algorithm showed remarkable improvements in convergence speed, training duration, and path planning success rate compared to baseline algorithms.
Researchers developed an innovative AI approach called GraSSCoL to predict complex astrochemical reactions. The model achieved outstanding Top-k accuracy scores, outperforming earlier state-of-the-art models by a significant margin.
The survey highlights a dual-path framework for harnessing large language models' intrinsic reasoning and integrating them with external methodologies. Large language models are employed for step-by-step reasoning methods, enabling the decomposition of complex tasks and exploration of multiple reasoning paths.
Researchers developed a regional explanation method to capture nonlinear relationships between molecular features and properties, offering fine-grained insights into chemical stability. The method was validated on two datasets, demonstrating broad applicability across different chemical domains.
Silicon spin qubits boast long coherence times and high gate fidelities, enabling universal quantum computers. Recent studies demonstrate gate fidelities required for fault-tolerant operations at temperatures above 1 Kelvin.
A comprehensive review explores DNA computing circuits operating within living cells, leveraging dynamic nanodevices powered by DNA strand displacement reactions. Key findings include the integration of computational principles with random biochemical processes and chemical reactions in biological systems.
Researchers developed an AI model that classifies variable stars from light curves with high accuracy, outperforming traditional approaches. The StarWhisper LightCurve series achieves near 90% accuracy with minimal manual intervention, paving the way for parallel data analysis and multi-modal AI applications in astronomy.
A new quantum-classical approach has been developed for designing photochromic materials, accelerating the discovery of novel compounds. The method identified five promising candidates with key properties essential for photopharmacology applications.
Researchers at Peking University demonstrate potential of nuclear electric resonance to control nitrogen atom spins in DNA, encoding genetic information. The study reveals intricate relationships between electric field gradients, nitrogen orientations and DNA base structures.
Current energy-hungry transformer-based systems contrast with Turing's idea of machines that develop intelligence naturally, like human children. AI systems can now perform tasks exclusive to human intellect, such as generating coherent text and discussing abstract ideas, but with limitations on sustainability and societal impact
Quantum walks utilize quantum phenomena to design algorithms for applications such as database search, network analysis, and navigation. These models offer unique features and computational advantages, including faster diffusion and improved sampling efficiency.
Researchers achieved near-perfect accuracy in detecting Parkinson's disease by analyzing brain responses to emotional situations. The study identified distinct patterns in how patients processed emotions, enabling accurate differentiation between patients and healthy controls.
A comprehensive framework for developing embodied agents is proposed to improve human-agent interactions. The integration of nine components aims to facilitate longer interactions by assessing users' goals, mental states, and interrelationships.
Researchers developed innovative encoding methods that simplified quantum circuits for data encoding, reducing circuit depth by a factor of 100 while maintaining accuracy. These methods showed improved resilience against adversarial attacks, paving the way for practical application of quantum machine learning on current devices.
Researchers introduced a novel illumination beam design based on deep learning, eliminating the need for sophisticated optics tools. The approach enhances image quality by optimizing both the deep learning network and the illumination beam simultaneously.
Researchers have developed a novel approach using deep learning to accelerate the solution of Navier-Stokes equations, a set of classical equations that describe fluid dynamics. The team's method achieved inference latencies of just 7 milliseconds per input, outperforming traditional finite difference methods.
Speech emotion recognition models are susceptible to adversarial attacks, which can significantly reduce their performance. The study found that black-box attacks outperformed white-box attacks and achieved impressive results despite limited access to the model's internal workings.
A Chinese research team introduced a novel two-stage framework using stacked transformers for multimodal sentiment analysis, improving the analysis of emotions expressed through modality combinations. The framework was tested on three open datasets and performed better than or as well as benchmark models.
A new deep-physics-informed sparsity framework significantly enhances structural fidelity and universality in fluorescence microscopy. It integrates physical imaging models, prior knowledge, and deep learning to resolve finer details and outperform existing methods.
Researchers introduce novel deep learning methods for channel extrapolation and beamforming in terahertz communication systems. These methods offer significant improvements in efficiency and robustness, addressing challenges in imperfect channel state information conditions.
Researchers developed a structure called multiplexed neuron sets to reduce crosstalk in optical neural networks. The new backpropagation training algorithm achieved comparable performance while improving energy efficiency by a factor of 10.
A new framework mitigates bias in machine classification by evaluating under different fairness metrics and inferring specific bias terms from data. The framework substantially reduces bias in classification outcomes while preserving accuracy across seven datasets and 21 classifiers.
Researchers developed composite and adiabatic pulses to improve single-qubit gate robustness, reducing control field error by nearly an order of magnitude. Their designs mitigated leakage and seepage, essential factors in assessing quantum operation fidelity.
Researchers focused on measuring information transfer in biological neurons, simulated neurons, and electronic neuromorphic systems. The team demonstrated that it is possible to transform biological circuits into electronic circuits while maintaining the amount of information transferred.
Molecular quantum computing may connect quantum biology and cognitive science through shared concepts like quantum degrees of freedom. Researchers explore potential links between charge movement, spin states, and biological processes in neurons and photosynthesis.
Researchers developed an automated protocol-design approach to determine optimal random quantum circuits for quantum computational advantage experiments. The new method uses the Schrödinger-Feynman algorithm to evaluate complexity, reducing estimation time and increasing the gap between quantum computing and classical simulation.
A comprehensive survey published in Intelligent Computing explores deep learning techniques for cellular traffic prediction, enhancing intelligent 5G network construction and resource management. The review highlights three main applications of cellular traffic prediction, including temporal and spatial-temporal prediction methods.
A review published in Intelligent Computing outlines the strengths of automatic approaches to designing metaheuristics, which can lead to more successful outcomes and reduce redundant, metaphor-based algorithms. The authors encourage research that relies on automatic design, utilizing modular software frameworks and configuration tools.
TaskMatrix.AI uses APIs to connect general-purpose foundation models with specialized models for specific tasks. The tool can perform digital and physical tasks, provide interpretable responses, and learn continuously.
The review of affective computing highlights recent advancements and future trends in emotion perception, recognition, and response. Key findings include the growing publication volume in the field, with China leading the way, as well as emerging trends like multimodal fusion and knowledge-driven approaches.
Explainable AI methods have been developed to make audio models more interpretable and transparent. Researchers categorize existing audio XAI methods into two groups: general methods and audio-specific methods, offering new possibilities for improving the trustworthiness of AI decision-making in audio tasks.
Researchers developed artificial neural networks using silicon microresonators, offering a promising platform for efficient AI systems. The devices mimic biological neurons' nonlinear behavior, enabling precise control of light properties.
A new algorithm integrates deep learning and federated learning for accurate channel estimation, outperforming state-of-the-art models in sparse and dense scenarios. The algorithm's use of a user motivation scheme and federated learning framework provides robustness, adaptability, and scalability.
Researchers used quantum support vector machines to classify flow separation and angle of attack with increased accuracy, solving complex problems faster and more accurately than classical methods.
Researchers introduce a cycle-consistency-based uncertainty quantification technique to enhance the reliability of deep neural networks in solving inverse imaging problems. The method uses forward-backward cycles to estimate network uncertainty, demonstrating improved accuracy in detecting image corruption and out-of-distribution images.
This review article surveys existing deep active learning approaches, applications, and challenges in the context of foundation models. Effective query strategies and model training methods are essential for optimizing joint performance.
Researchers have developed a novel approach using tensor networks to bridge quantum concepts with machine learning, enabling efficient construction of probabilistic models from quantum states. This framework offers enhanced interpretability comparable to classical probabilistic machine learning.
Researchers suggest a new evaluation framework to assess AI reasoning abilities, comprising psychological experiments, self-reflection, and source code analysis. This approach aims to determine if AI systems genuinely reason like humans.
A new parallel hybrid quantum neural network demonstrates improved performance by combining the strengths of both quantum and classical layers. The model outperforms traditional machine learning methods in processing complicated patterns and relationships from data inputs.
The Digital Twin Brain platform combines intricate brain atlases and dynamic neural models to simulate human brain functions. This innovative approach holds promise for advancing artificial general intelligence and precision mental healthcare.
Researchers have developed a noise-resistant method for detecting object edges without prior imaging. The new method, called edge-sensitive single-pixel imaging, proves highly effective in accurately detecting object edges despite the presence of noise.
A joint research team published a review on in-sensor visual computing, a three-in-one hardware solution that overcomes high latency, power consumption, and privacy risks. The SCAMP chip is a key device, enabling general-purpose, programmable, and massively parallel systems for robotics and computer vision.
Researchers developed a novel optimization method combining natural evolutionary strategy with gradient descent to overcome the barren plateau problem in parametric quantum circuits. The new method exhibited superior performance in achieving higher accuracy, showcasing its potential for revolutionizing quantum algorithm optimization.
Researchers developed an entanglement witness circuit to detect qubit entanglement in cloud-based services, overcoming limitations and enabling users to test for entangled qubits. The new framework EW 2.0 is twice as efficient at detecting entanglement.
Researchers designed a simplified Mach-Zehnder interferometer mesh for real-valued matrix-vector multiplication, reducing hardware requirements and energy consumption. The new mesh detects incoherent light and is scalable, making it suitable for large-scale optical neural networks.
Researchers proposed a novel self-organizing approach for robot swarms to achieve consensus, combining aspects of centralized and decentralized control. This method reduces uncertainty sources and enhances collective perception accuracy, enabling robots to fuse sensor information without global or static communication networks.