Learning to sense from events via semantic variational autoencoder

In this paper, we introduce the concept of learning to sense, which aims to emulate a complex characteristic of human reasoning: the ability to monitor and understand a set of interdependent events for decision-making processes. Event datasets are composed of textual data and spatio-temporal features that determine where and when a given phenomenon occurred. In learning to sense, related events are mapped closely to each other in a semantic vector space, thereby identifying that they contain similar contextual meaning. However, learning a semantic vector space that satisfies both textual similarities and spatio-temporal constraints is a crucial challenge for event analysis and sensing. This paper investigates a Semantic Variational Autoencoder (SVAE) to fine-tune pre-trained embeddings according to both textual and spatio-temporal events of the class of interest. Experiments involving more than one hundred sensors show that our SVAE outperforms a competitive one-class classification baseline. Moreover, our proposal provides desirable learning requirements to sense scenarios, such as visualization of the sensor decision function and heat maps with the sensor’s geographic impact.

The problem of no hands: responsibility voids in collective decisions

The problem of no hands concerns the existence of so-called responsibility voids: cases where a group makes a certain decision, yet no individual member of the group can be held responsible for this decision. Criteria-based collective decision procedures play a central role in philosophical debates on responsibility voids. In particular, the well-known discursive dilemma has been used to argue for the existence of these voids. But there is no consensus: others argue that no such voids exist in the discursive dilemma under the assumption that casting an untruthful opinion is eligible. We argue that, under this assumption, the procedure used in the discursive dilemma is indeed immune to responsibility voids, yet such voids can still arise for other criteria-based procedures. We provide two general characterizations of the conditions under which criteria-based collective decision procedures are immune to these voids. Our general characterizations are used to prove that responsibility voids are ruled out by criteria-based procedures involving an atomistic or monotonic decision function. In addition, we show that our results imply various other insights concerning the logic of responsibility voids.

An Empirical Study on Intelligent Rural Tourism Service by Neural Network Algorithm Models

With the continuous development of social economy, tourism has become one of the many choices and is becoming more and more popular. However, it should be noted that how to provide high-quality and efficient tourism services is extremely important. This paper introduces the neural network algorithm and the optimal classification decision function, through unified combing, classification, and coding of scenic spots, to achieve the subclass classification of scenic spots, based on the optimal distribution function of random intelligent selection, and the formation of the corresponding scenic spots traversal clear tourism routes. The corresponding motivation iteration is obtained by using the corresponding travel route transmission, the best travel route is defined, the corresponding auxiliary decision support is provided, and the simulation experiment is carried out. The experimental results show that the neural network algorithm and the optimal classification decision function are effective and can support the intelligent decision assistance of rural tourism service.

MATHEMATICAL MODEL OF A CORRELATION RECEIVER WITH A PIECEWISE LINEAR APPROXIMATION OF THE DECISION FUNCTION OF THE THRESHOLD DEVICE

Разработана математическая модель двухканального корреляционного приемника радиосигналов с кусочно-линейной аппроксимацией, решающей функции порогового устройства. Приемник рассчитан на прием наиболее часто встречающихся на практике сигналов со случайной начальной фазой в условиях белого гауссовского шума. В синтезированной математической модели применяется кусочно-линейная аппроксимация решающей функции порогового устройства. Проведен сравнительный анализ характеристик обнаружения радиосигналов со случайной начальной фазой от отношения сигнал/шум, посчитанных с использованием разработанной математической модели корреляционного приемника с кусочно-линейной решающей функциeй порогового устройства. Представлены полученные в результате математического моделирования процесса функционирования корреляционного приемника при обнаружении сигналов со случайной начальной фазой в условиях шума зависимости вероятности правильного обнаружения от отношения сигнал/шум. Показано, что результаты имитационного моделирования согласуются с теоретическими расчетами. Выявлено, что представление решающих функций пороговых устройств в классической теории обнаружения сигналов в виде идеализированных (оптимальных), которые не учитывают их нелинейность, приводят либо к увеличению вероятности ложной тревоги, либо к уменьшению вероятности правильного обнаружения, что приводит к ошибкам первого рода We developed a mathematical model of a two-channel correlation receiver of radio signals with piecewise linear approximation of the decision function of the threshold device. The receiver is designed to receive the most commonly encountered signals in practice with a random initial phase in a white Gaussian noise environment. In the synthesized mathematical model, a piecewise linear approximation of the decision function of the threshold device is used. We carried out a comparative analysis of the characteristics of detecting radio signals with a random initial phase from the signal-to-noise ratio, calculated using the developed mathematical model of a correlation receiver with a piecewise linear decision function of the threshold device and known. The paper presents the dependences of the probability of correct detection on the signal-to-noise ratio obtained as a result of mathematical modeling of the process of functioning of the correlation receiver when detecting signals with a random initial phase under noise conditions. We show that the results of simulation are consistent with theoretical calculations. We found that the representation of the decision functions of threshold devices in the classical theory of signal detection in the form of idealized (optimal) ones, which do not take into account their nonlinearity, lead either to an increase in the probability of a false alarm, or to a decrease in the probability of correct detection, which leads to errors of the first kind

Robust diagnostic classification via Q-learning

Machine learning (ML) models have demonstrated the power of utilizing clinical instruments to provide tools for domain experts in gaining additional insights toward complex clinical diagnoses. In this context these tools desire two additional properties: interpretability, being able to audit and understand the decision function, and robustness, being able to assign the correct label in spite of missing or noisy inputs. This work formulates diagnostic classification as a decision-making process and utilizes Q-learning to build classifiers that meet the aforementioned desired criteria. As an exemplary task, we simulate the process of differentiating Autism Spectrum Disorder from Attention Deficit-Hyperactivity Disorder in verbal school aged children. This application highlights how reinforcement learning frameworks can be utilized to train more robust classifiers by jointly learning to maximize diagnostic accuracy while minimizing the amount of information required.

Determination of the complexity of the algorithm for the dynamic characteristics of the controlled process of the marine object

Своевременность решений характеризуется соотношением динамических характеристик управляемого процесса и расходов времени на принятие решения. Для получения такого соотношения сначала обоснуем утверждение: время принятия решения системой с нелинейной решающей функцией равняется времени принятия решений системой с линейной решающей функцией, если соответствующие алгоритмы имеют одинаковую сложность. При изучении организационных объектов важно выделять класс целенаправленных или целеустремленных систем. В этом классе, в свою очередь, можно выделить системы, в которых цели задаются извне (обычно это имеет место в закрытых системах), и системы, в которых цели формируются внутри системы (что характерно для открытых, самоорганизующихся систем). Целеустремленные системы являются системами с явно выраженными целевыми устремлениями по достижения цели. Как правило это системы управления движущимися морскими объектами: судами и кораблями в условиях информационно-сложных ситуациях и деятельность таких систем нацелена на результат, а результат может быть достигнут если есть обратная связь. Цель исследований: проведение интроспективного анализа и синтеза целеустремленных систем в условиях информационно – сложной ситуации. Методы исследований – интроспективный анализ и синтез целеустремленных систем. При этом подходе возможно решение принципиальных проблем разработки информационной системы (ИС) методом выделения инвариантных блоков, реализация которых позволяет рассматривать структуру ИС как набор однотипных элементов и возможность построения управляемых объектов. Важной особенностью ИС является обеспечение возможности описания метода организации данных и алгоритмов оперирования данными в терминах используемого информационного языка. Целью данной работы является не перечисление или выбор возможных инвариантных блоков, (этот выбор зависит от характера решаемых задач системой и от свойств управляемых объектов), а построение такой модели блоков ИС и их описания на языке, понятном человеку и машине, которая позволяет создать набор алгоритмов и программ, обслуживающих систему, независящих от содержимого блоков конкретных ИС. Timeliness of decisions is characterized by the ratio of the dynamic characteristics of the controlled process and the time spent on making a decision. To obtain such a relation, we first substantiate the statement: the decision time by a system with a nonlinear decision function is equal to the decision time by a system with a linear decision function, if the corresponding algorithms have the same complexity. When studying organizational objects, it is important to distinguish a class of purposeful or purposeful systems. In this class, in turn, it is possible to distinguish systems in which goals are set from the outside (usually this is the case in closed systems), and systems in which goals are formed within the system (which is typical for open, self-organizing systems). Purposeful systems are systems with clearly expressed purposeful aspirations to achieve a goal. As a rule, these are control systems for moving marine objects: ships and ships in information-complex situations and the activity of such systems is aimed at the result, and the result can be achieved if there is feedback. difficult situation. Research methods - introspective analysis and synthesis of purposeful systems. With this approach, it is possible to solve the fundamental problems of developing an information system (IS) by the method of identifying invariant blocks, the implementation of which allows us to consider the IS structure as a set of elements of the same type and the possibility of constructing controlled objects. An important feature of the IS is the ability to describe the method of organizing data and algorithms for operating with data in terms of the information language used. The purpose of this work is not to enumerate or select possible invariant blocks (this choice depends on the nature of the tasks being solved by the system and on the properties of controlled objects), but to build such a model of IS blocks and their description in a language understandable to humans and machines, which allows creating a set of algorithms and programs serving the system that do not depend on the contents of specific IS blocks.

R-optimal designs for second-order Scheffé model with qualitative factors

<abstract><p>Considering a mixture model with qualitative factors, the $ R $-optimal design problem is investigated when the levels of the qualitative factor interact with the mixture factors. In this paper, the conditions for $ R $-optimality of designs with mixture and qualitative factors are presented. General analytical expressions are also derived for the decision function under the $ R $-optimal designs, in order to verify that the resulting designs satisfy the general equivalence theorem. In addition, the relative efficiency of the $ R $-optimal design is discussed.</p></abstract>

Perceptron Ranking Using Interval Labels with Ramp Loss for Online Ordinal Regression

Due to its wide applications and learning efficiency, online ordinal regression using perceptron algorithms with interval labels (PRIL) has been increasingly applied to solve ordinal ranking problems. However, it is still a challenge for the PRIL method to handle noise labels, in which case the ranking results may change dramatically. To tackle this problem, in this paper, we propose noise-resilient online learning algorithms using ramp loss function, called PRIL-RAMP, and its nonlinear variant K-PRIL-RAMP, to improve the performance of PRIL method for noisy data streams. The proposed algorithms iteratively optimize the decision function under the framework of online gradient descent (OGD), and we justify the algorithms by showing the order preservation of thresholds. It is validated in the experiments that both approaches are more robust and efficient to noise labels than state-of-the-art online ordinal regression algorithms on real-world datasets.