Simultaneously Improve Transferability and Discriminability for Adversarial Domain Adaptation

Although adversarial domain adaptation enhances feature transferability, the feature discriminability will be degraded in the process of adversarial learning. Moreover, most domain adaptation methods only focus on distribution matching in the feature space; however, shifts in the joint distributions of input features and output labels linger in the network, and thus, the transferability is not fully exploited. In this paper, we propose a matrix rank embedding (MRE) method to enhance feature discriminability and transferability simultaneously. MRE restores a low-rank structure for data in the same class and enforces a maximum separation structure for data in different classes. In this manner, the variations within the subspace are reduced, and the separation between the subspaces is increased, resulting in improved discriminability. In addition to statistically aligning the class-conditional distribution in the feature space, MRE forces the data of the same class in different domains to exhibit an approximate low-rank structure, thereby aligning the class-conditional distribution in the label space, resulting in improved transferability. MRE is computationally efficient and can be used as a plug-and-play term for other adversarial domain adaptation networks. Comprehensive experiments demonstrate that MRE can advance state-of-the-art domain adaptation methods.

Comparative study of flood coincidence risk estimation methods in the mainstream and its tributaries

The coincidence of floods in the mainstream and its tributaries may lead to a large flooding in the downstream confluence area, and the flood coincidence risk analysis is very important for flood prevention and disaster reduction. In this study, the multiple regression model was used to establish the functional relationship among flood magnitudes in the mainstream and its tributaries. The mixed von Mises distribution and Pearson Type III distribution were selected to fit the probability distribution of the annual maximum flood occurrence dates and magnitudes, respectively. The joint distributions of the annual maximum flood occurrence dates and magnitudes were established using copula function, respectively. Fuhe River in the Poyang Lake region was selected as a study case. The joint probability, co-occurrence probability and conditional probability of flood magnitudes were quantitatively estimated and compared with the predicted flood coincidence risks. The results show that the selected marginal and joint distributions can fit observed flood dataset very well. The coincidence probabilities of flood occurrence dates in the upper mainstream and its tributaries mainly occur from May to early July. It is found that the conditional probability is the most consistent with the predicted flood coincidence risks in the mainstream and its tributaries, and is more reliable and rational in practice.

Probabilistic interval estimation of design floods under non-stationary conditions by an integrated approach

Quantifying the uncertainty of non-stationary flood frequency analysis is very crucial and beneficial for planning and design of water engineering projects, which is fundamentally challenging especially in the presence of high climate variability and reservoir regulation. This study proposed an integrated approach that combined the Generalized Additive Model for Location, Scale and Shape parameters (GAMLSS) method, the Copula function and the Bayesian Uncertainty Processor (BUP) technique to make reliable probabilistic interval estimations of design floods. The reliability and applicability of the proposed approach were assessed by flood datasets collected from two hydrological monitoring stations located in the Hanjiang River of China. The precipitation and the reservoir index were selected as the explanatory variables for modeling the time-varying parameters of marginal and joint distributions using long-term (1954–2018) observed datasets. First, the GAMLSS method was employed to model and fit the time-varying characteristics of parameters in marginal and joint distributions. Second, the Copula function was employed to execute the point estimations of non-stationary design floods. Finally, the BUP technique was employed to perform the interval estimations of design floods based on the point estimations obtained from the Copula function. The results demonstrated that the proposed approach can provide reliable probabilistic interval estimations of design floods meanwhile reducing the uncertainty of non-stationary flood frequency analysis. Consequently, the integrated approach is a promising way to offer an indication on how design values can be estimated in a high-dimensional problem.

The Moment Properties of Order, Reversed Order and Upper Record Statistics for the Power Ailamujia Distribution

The power Ailamujia distribution has been successfully developed in statistics, both theoretically and practically, performing well in the fitting of various types of data. This paper investigates the moment properties of the associated order, reversed order and upper record statistics, which are indeed unexplored aspects of this distribution. In particular, the exact expressions for the single moments of the order and reversed order statistics are provided. Some recurrence relationships for both single and product moments for the order and upper record statistics are proved. For additional goals, certain joint distributions are also given.

Contextuality in Neurobehavioural and Collective Intelligence Systems

Contextuality is often described as a unique feature of the quantum realm, which distinguishes it fundamentally from the classical realm. This is not strictly true, and stems from decades of the misapplication of Kolmogorov probability. Contextuality appears in Kolmogorov theory (observed in the inability to form joint distributions) and in non-Kolmogorov theory (observed in the violation of inequalities of correlations). Both forms of contextuality have been observed in psychological experiments, although the first form has been known for decades but mostly ignored. The complex dynamics of neural systems (neurobehavioural regulatory systems) and of collective intelligence systems (social insect colonies) are described. These systems are contextual in the first sense and possibly in the second as well. Process algebra, based on the Process Theory of Whitehead, describes systems that are generated, transient, open, interactive, and primarily information-driven, and seems ideally suited to modeling these systems. It is argued that these dynamical characteristics give rise to contextuality and non-Kolmogorov probability in spite of these being entirely classical systems.

On asymptotic joint distributions of cherries and pitchforks for random phylogenetic trees

Tree shape statistics provide valuable quantitative insights into evolutionary mechanisms underpinning phylogenetic trees, a commonly used graph representation of evolutionary relationships among taxonomic units ranging from viruses to species. We study two subtree counting statistics, the number of cherries and the number of pitchforks, for random phylogenetic trees generated by two widely used null tree models: the proportional to distinguishable arrangements (PDA) and the Yule-Harding-Kingman (YHK) models. By developing limit theorems for a version of extended Pólya urn models in which negative entries are permitted for their replacement matrices, we deduce the strong laws of large numbers and the central limit theorems for the joint distributions of these two counting statistics for the PDA and the YHK models. Our results indicate that the limiting behaviour of these two statistics, when appropriately scaled using the number of leaves in the underlying trees, is independent of the initial tree used in the tree generating process.

EXPLICIT DISTRIBUTION OF SELECTED TWO-DIMENSIONAL AND THREE-DIMENSIONAL STATISTICS OF THE (0,1)-SEQUENCE

The joint distributions of the given number of 2-chains and the given number of 3-chains of a fixed form of a random bit sequence are considered, which allow performing a statistical analysis of local sections of this sequence. All configurations consisting of two consecutive zeros or ones of a bit sequence of a given length act as 2-chains. In turn, 3-chains are all configurations consisting of three consecutive either ones (provided that the 2-chains are zero) or zeros (provided that the 2-chains are one), as well as 3-chains all configurations are considered that consist either of three consecutive digits: one, zero and one (provided that the 2- chains are zero), or of three consecutive digits: zero, one and zero (provided that the 2- chains are one). The paper establishes explicit expressions for two-dimensional and three-dimensional joint distributions of events, reflecting the number of some combinations of the indicated chains in a finite random bit sequence. One of the basic assumptions is that zeros and ones in a bit sequence are independent, equally distributed random variables. The proofs of the formulas for the distributions of these events are based on counting the number of corresponding favorable events, provided that the bit sequence contains a fixed number of zeros and ones. As examples of using explicit expressions of joint distributions, tables are given in which the values of the probabilities of the events listed above for a random bit sequence of length 40 (tables 1–3) and length 24 (table 4) are given for some fixed values of the number of 2-chains and the number 3-chains under the assumption that zeros and ones appear independently and uniformly. For clarity, tables 1‑3 are illustrated with bubble charts. The established formulas may be of interest for the problems of testing local sections formed at the output of pseudo-random number generators, for some problems of protecting information from unauthorized access, as well as in other areas where it becomes necessary to analyze bit sequences.

Age of Information in a Cooperative Slotted Aloha Network: Marginal and Joint Distributions

In this paper, we investigate a slotted Aloha cooperative network where a source node and a relay node send status updates of two underlying stochastic processes to a common destination. Additionally, the relay node cooperates with the source by accepting its packets for further retransmissions, where the cooperation policy comprises acceptance and relaying probabilistic policies. Exact marginal steady state distributions of the source and relay Age of Information (AoI) and Peak AoI (PAoI) sequences are obtained using Quasi-Birth-Death (QBD) Markov chain models. Extending this approach, we also obtain the joint distribution of the source and relay AoI sequences out of which one can obtain the steady state distribution of the Squared Difference of the two AoI sequences (SDAoI), which finds applications in network scenarios where not only the timeliness of status updates of each process is desired but also their simultaneity is of crucial importance. In this regard, we numerically obtain the optimal cooperation policy in order to minimize the expected value of SDAoI subject to a constraint on the average PAoI of the relay. Finally, our proposed analytical approach is verified by simulations and the performance of the optimal policy is discussed based on the numerical results.

Age of Information in a Cooperative Slotted Aloha Network: Marginal and Joint Distributions

In this paper, we investigate a slotted Aloha cooperative network where a source node and a relay node send status updates of two underlying stochastic processes to a common destination. Additionally, the relay node cooperates with the source by accepting its packets for further retransmissions, where the cooperation policy comprises acceptance and relaying probabilistic policies. Exact marginal steady state distributions of the source and relay Age of Information (AoI) and Peak AoI (PAoI) sequences are obtained using Quasi-Birth-Death (QBD) Markov chain models. Extending this approach, we also obtain the joint distribution of the source and relay AoI sequences out of which one can obtain the steady state distribution of the Squared Difference of the two AoI sequences (SDAoI), which finds applications in network scenarios where not only the timeliness of status updates of each process is desired but also their simultaneity is of crucial importance. In this regard, we numerically obtain the optimal cooperation policy in order to minimize the expected value of SDAoI subject to a constraint on the average PAoI of the relay. Finally, our proposed analytical approach is verified by simulations and the performance of the optimal policy is discussed based on the numerical results.

Copula-based software metrics aggregation

A quality model is a conceptual decomposition of an abstract notion of quality into relevant, possibly conflicting characteristics and further into measurable metrics. For quality assessment and decision making, metrics values are aggregated to characteristics and ultimately to quality scores. Aggregation has often been problematic as quality models do not provide the semantics of aggregation. This makes it hard to formally reason about metrics, characteristics, and quality. We argue that aggregation needs to be interpretable and mathematically well defined in order to assess, to compare, and to improve quality. To address this challenge, we propose a probabilistic approach to aggregation and define quality scores based on joint distributions of absolute metrics values. To evaluate the proposed approach and its implementation under realistic conditions, we conduct empirical studies on bug prediction of ca. 5000 software classes, maintainability of ca. 15000 open-source software systems, and on the information quality of ca. 100000 real-world technical documents. We found that our approach is feasible, accurate, and scalable in performance.