Entropy最新文献

The lack of explicit negative labels issue is a prevalent challenge in numerous domains, including CV, NLP, and Recommender Systems (RSs). To address this challenge, many negative sample completion methods are proposed, such as optimizing sample distribution through pseudo-negative sampling and confidence screening in CV, constructing reliable negative examples by leveraging textual semantics in NLP, and supplementing negative samples via sparsity analysis of user interaction behaviors and preference inference in RS for handling implicit feedback. However, most existing methods fail to adequately address the Missing-Not-At-Random (MNAR) nature of the data and the potential presence of unmeasured confounders, which compromise model robustness in practice. In this paper, we first formulate the prediction task in RS with implicit feedback as a positive-unlabeled (PU) learning problem. We then propose a two-phase debiasing framework consisting of exposure status imputation, followed by debiasing through the proposed doubly robust estimator. Moreover, our theoretical analysis shows that existing propensity-based approaches are biased in the presence of unmeasured confounders. To overcome this, we incorporate a robust deconfounding method in the debiasing phase to effectively mitigate the impact of unmeasured confounders. We conduct extensive experiments on three widely used real-world datasets to demonstrate the effectiveness and potential of the proposed methods.

We propose a unified mathematical framework showing that the representational universality of modern foundational models arises from a shared finite latent domain. Building on the Finite Ring Continuum (FRC) framework, we model all modalities as epistemic projections of a common latent set Z⊂Ut, where Ut is a symmetry-complete finite-field shell. Using the uniqueness of minimal adequate representations, we prove the Universal Subspace Theorem, establishing that independently trained embeddings coincide, up to bijection, as coordinate charts on the same latent structure. This result explains cross-modal alignment, transferability, and semantic coherence as consequences of finite relational geometry rather than architectural similarity. The framework links representation learning, sufficiency theory, and FRC algebra, providing a principled foundation for universal latent structure in multimodal models.

Simultaneous ascending auctions find extensive applications in spectrum licensing and advertising space allocation. However, existing quantum sealed-bid auction protocols suffer from dual limitations: they cannot support multi-item simultaneous bidding scenarios, and their reliance on complex quantum resources along with requiring full quantum operational capabilities from bidders fails to accommodate practical constraints of quantum resource-limited users. To address these challenges, this paper proposes a multi-party semi-quantum simultaneous ascending auction protocol based on single-particle states. The protocol employs a trusted honest third party (HTP) responsible for quantum state generation, distribution, and security verification. Bidders determine their groups through quantum measurements and privately encode their bid vectors. Upon successful HTP authentication, each bidder obtains a unique identity code. During the bidding phase, HTP dynamically updates quantum sequences, allowing bidders to submit bids for multiple items by performing only simple unitary operations. HTP announces the highest bid for each item in real time and iteratively generates auction sequences until no new highest bid emerges, thereby achieving simultaneous ascending auctions for multiple items. It acts as a quantum-secured signaling layer, ensuring unconditional security for bid transmission and identity verification while maintaining classical auction logic. Quantum circuit simulations validate the protocol's feasibility with current technology while satisfying critical security requirements, including anonymity, verifiability, non-repudiation, and privacy preservation. It provides a scalable semi-quantum auction solution for resource-constrained scenarios.

The link between information and energy introduces the observer and their knowledge into the understanding of a fundamental quantity in physics. Two approaches compete to account for this link-Brillouin's negentropy law of information and Landauer's law on data erasure-which are often confused. The first, based on Clausius' inequality and Shannon's mathematical results, is very robust, whereas the second, based on the simple idea that information requires a material embodiment (data bits), is now perceived as more physical and therefore prevails. In this paper, we show that Landauer's idea results from a confusion between information (a global emergent concept) and data (a local material object). This confusion leads to many inconsistencies and is incompatible with thermodynamics and information theory. The reason it prevails is interpreted as being due to a frequent tendency of materialism towards reductionism, neglecting emergence and seeking to eliminate the role of the observer. A paradoxical trend, considering that it is often accompanied by the materialist idea that all scientific knowledge, nevertheless, originates from observation. Information and entropy are actually emergent quantities introduced in the theory by convention.

In natural systems, astrophysics, biological physics, and social physics, 1/f fluctuations are observed across a wide range of systems. Focusing on the case of music, we propose and verify a physical mechanism for generating these fluctuations. This mechanism is based on amplitude modulation (AM) and demodulation (DM), where the 1/f spectral law appears not in the raw waveform but in its demodulated amplitude envelope. Two distinct yet complementary processes generate the required AM: (i) stochastic synchronization among oscillators, modeled via an extended Kuramoto framework that captures perpetual synchronization-desynchronization cycles, and (ii) frequency-selective resonance, modeled by spectral accumulation of eigenmodes in acoustic or structural environments. Numerical simulations demonstrate that both mechanisms, acting alone or in combination, robustly generate 1/f spectra spanning several digits when demodulation is applied and that the classical Kuramoto critical point is not essential for its emergence. While this analysis focuses on 1/f fluctuations in musical performance and acoustics, we also note that 1/f fluctuations inherent in musical scores may be similarly described by the AM/DM mechanism.

Aside from languages having no form of written expression, it is usually the case with every language on this planet that texts are written in a single character. But every rule has its exceptions. A very rare exception is Japanese, the texts of which are written in the three kinds of characters. In European languages, no one can find a text written in a mixture of the Latin, Cyrillic, and Greek alphabets. For several Japanese texts currently available, we conduct a quantitative analysis of how the three characters are mixed using a methodology based on a binary pattern approach to the sequence that has been generated by a procedure. Specifically, we consider two different texts in the former and present constitutions as well as a famous American story that has been translated at least 13 times into Japanese. For the latter, a comparison is made among the human translations and four machine translations by DeepL and Google Translate. As metrics of divergence and diversity, the Hellinger distance, chi-square value, normalized Shannon entropy, and Simpson's diversity index are employed. Numerical results suggest that in terms of the entropy, the 17 translations consist of three clusters, and that overall, the machine-translated texts exhibit entropy higher than the human translations. The finding suggests that the present method can provide a tool useful for stylometry and author attribution. Finally, through comparison with the diversity index, capabilities of the entropic measure are confirmed. Lastly, in addition to the abovementioned texts, applicability to the Japanese version of the periodic table of elements is investigated.

The rapid growth of Internet of Things (IoT) devices and the emergence of 5G/6G networks have created major challenges in secure and reliable data transmission. Traditional cryptographic algorithms, while robust, often suffer from high computational complexity and latency, making them less suitable for large-scale, real-time applications. This paper proposes a chaos-based encryption framework that uses the Sprott chaotic oscillator to generate secure and unpredictable signals for encryption. To achieve accurate synchronization between the transmitter and the receiver, two bio-inspired metaheuristic algorithms-the Pachycondyla Apicalis Algorithm (API) and the Penguin Search Optimization Algorithm (PeSOA)-are employed to identify the optimal control parameters of the Sprott system. This optimization improves synchronization accuracy and reduces computational overhead. Simulation results show that PeSOA-based synchronization outperforms API in convergence speed and Root Mean Square Error (RMSE). The proposed framework provides robust, scalable, and low-latency encryption for IoT and 5G/6G networks, where massive connectivity and real-time data protection are essential.

We investigate the fundamental trade-off between entropy and the Gini index within income distributions, employing a stochastic framework to expose deficiencies in conventional inequality metrics. Anchored in the principle of maximum entropy (ME), we position entropy as a key marker of societal robustness, while the Gini index, identical to the (second-order) K-spread coefficient, captures spread but neglects dynamics in distribution tails. We recommend supplanting Lorenz profiles with simpler graphs such as the odds and probability density functions, and a core set of numerical indicators (K-spread K2/μ, standardized entropy Φμ, and upper and lower tail indices, ξ, ζ) for deeper diagnostics. This approach fuses ME into disparity evaluation, highlighting a path to harmonize fairness with structural endurance. Drawing from percentile records in the World Income Inequality Database from 1947 to 2023, we fit flexible models (Pareto-Burr-Feller, Dagum) and extract K-moments and tail indices. The results unveil a concave frontier: moderate Gini reductions have little effect on entropy, but aggressive equalization incurs steep stability costs. Country-level analyses (Argentina, Brazil, South Africa, Bulgaria) link entropy declines to political ruptures, positioning low entropy as a precursor to instability. On the other hand, analyses based on the core set of indicators for present-day geopolitical powers show that they are positioned in a high stability area.

The entropy measure of residual cumulative Sharma-Taneja-Mittal is an alternative measure of uncertainty for residual cumulative entropy. This study investigates further theoretical properties and develops nonparametric estimation procedures for the proposed measure. The performance of the estimator is evaluated through simulation experiments, and its practical relevance is illustrated using a real-world dataset on malignant tumor cases. Moreover, we investigate the properties of its dynamic version, including stochastic comparisons and its connections with the hazard rate function, mean residual function, and equilibrium random variables. Moreover, we introduce an alternative version of dynamic residual cumulative Sharma-Taneja-Mittal entropy and examine its monotonic properties. Additionally, we discuss this alternative version and its conditional form in the circumstances of record values. We introduce this alternative expression for the residual lifespan of upper record quantities in general distributions, characterizing it as a measure of upper record quantities derived from a distribution of uniform. Since Sharma-Taneja-Mittal entropy measures uncertainty, we also investigate its use in determining the entropy of the lifespan of mixed and coherent mechanisms, in which the lives of its constituent components are identically distributed and independent.

Proof of sequential work (PoSW), as an emerging cryptographic primitive, is designed to provide a verifiable method for proving that a computational process has incurred a real and continuous expenditure of time. This characteristic demonstrates its significant application potential in decentralized systems, time-stamping services, and trusted computing. This paper systematically reviews and discusses the developmental trajectory, typical variants, potential attacks, and diverse applications of PoSW. Concurrently, it places a special emphasis on analyzing the evolutionary path and application scenarios of its important special case-the verifiable delay function (VDF) aiming to provide a comprehensive reference for research and practice in related fields.