Sound Source Separation Mechanisms of Different Deep Networks Explained from the Perspective of Auditory Perception

Thanks to the development of deep learning, various sound source separation networks have been proposed and made significant progress. However, the study on the underlying separation mechanisms is still in its infancy. In this study, deep networks are explained from the perspective of auditory perception mechanisms. For separating two arbitrary sound sources from monaural recordings, three different networks with different parameters are trained and achieve excellent performances. The networks’ output can obtain an average scale-invariant signal-to-distortion ratio improvement (SI-SDRi) higher than 10 dB, comparable with the human performance to separate natural sources. More importantly, the most intuitive principle—proximity—is explored through simultaneous and sequential organization experiments. Results show that regardless of network structures and parameters, the proximity principle is learned spontaneously by all networks. If components are proximate in frequency or time, they are not easily separated by networks. Moreover, the frequency resolution at low frequencies is better than at high frequencies. These behavior characteristics of all three networks are highly consistent with those of the human auditory system, which implies that the learned proximity principle is not accidental, but the optimal strategy selected by networks and humans when facing the same task. The emergence of the auditory-like separation mechanisms provides the possibility to develop a universal system that can be adapted to all sources and scenes.

Economic Assessment of South African Urban Green Spaces Using the Proximity Principle: Municipal Valuation vs. Market Value

<p>Urban green spaces (UGSs) deliver ecosystem services and potential economic benefits like increases in proximate residential property prices. The proximity principle (PP) premises that property prices increase as distance to UGS decreases. The PP has generally been confirmed by studies using municipal valuations and market values internationally. Conversely, South African studies have mostly employed municipal valuations and results have rejected the PP. There is an accepted interrelationship, but also often discrepancies, between municipal valuations and market values, presenting scope for this article to explore whether negative results are confirmed when market values replace municipal valuations in PP studies in the South African context. Accordingly, a statistical analysis of market values is completed in the Potchefstroom case study, where five test sites are replicated from studies that employed municipal valuations for longitudinal comparison. Results verify generally higher market values than municipal valuations and confirm the PP in two, but reject the PP in three, of five test sites. Previous studies employing municipal valuations in the case study confirmed the PP in one instance, thus presenting certain, but limited, inconsistencies between findings based on municipal valuation vs. market value. Results suggest that the market’s willingness to pay for UGS proximity is sensitive to the ecosystem services and disservices rendered by specific UGS, but not significantly more than reflected in municipal valuations. Overall, findings underscore the need to protect and curate features that encourage willingness to pay for UGS proximity to increase municipal valuations and property taxes to help finance urban greening.</p>

Better state pictures facilitating state machine characteristic conjecture

The mutual exclusion protocol invented by Mellor-Crummey and Scott (called MCS protocol) is used to exemplify that state picture designs based on which the state machine graphical animation (SMGA) tool produces graphical animations should be better visualized. Variants of MCS protocol have been used in Java virtual machines and therefore the 2006 Edsger W. Dijkstra Prize in Distributed Computing went to their paper on MCS protocol. The new state picture design of a state machine formalizing MCS protocol is assessed based on Gestalt principles, more specifically proximity principle and similarity principle. We report on a core part of a formal verification case study in which the new state picture design and the SMGA tool largely contributed to the successful completion of the formal proof that MCS protocol enjoys the mutual exclusion property. The lessons learned acquired through our experiments are summarized as two groups of tips. The first group is some new tips on how to make state picture designs. The second one is some tips on how to conjecture state machine characteristics by using the SMGA tool. We also report on one more case study in which the state picture design has been made for the mutual exclusion protocol invented by Anderson (called Anderson protocol) and some characteristics of the protocol have been discovered based on the tips.

Part 2 National and Regional Reports, Part 2.5 Latin America: Coordinated by Lauro Gama and José Antonio Moreno Rodríguez, 66 Venezuela: Venezuelan Perspectives on the Hague Principles

This chapter discusses Venezuelan perspectives on the Hague Principles. In Venezuela, the law applicable to international contracts depends on two main sources: the Inter-American Convention on Law Applicable to International Contracts, ratified in 1995 (the Mexico Convention), and the Venezuelan Act on Private International Law. Both the Mexico Convention and the Venezuelan Act adopt solutions that reflect the modern tendencies in the field of contract conflicts. Indeed, these instruments recognize the primacy of party autonomy and the subsidiary application of the proximity principle. The gaps of the Venezuelan Act could be filled, of course, with the solutions of the Mexico Convention. That would be welcome because no possible reform of the Act has been raised as yet. In fact, when scholars comment on the Venezuelan system of Private International Law, they analyse both the Mexico Convention and the Venezuelan Act, because, in short, the three rules of the Act reproduce, with some minor differences, the main rules of the Mexico Convention. Courts could even refer to the Hague Principles as persuasive authority.

Revisiting the Proximity Principle with Stakeholder Input: Investigating Property Values and Distance to Urban Green Space in Potchefstroom

Nature is essential to urban quality of life, yet green spaces are under pressure. In an attempt to strengthen the case for urban greening and to reclaim nature into cities, this research considered green spaces from an economic spatial perspective. The proximity principle, as part of hedonic price analysis, is employed to determine the impact of green spaces on property value in specifically selected residential areas within Potchefstroom, South Africa. Our statistical analysis indicated a rejection of the proximity principle in some areas, contradicting internationally accepted theory. To investigate local trends and possible reasons for the rejection, supporting quantitative data was gathered through structured questionnaires disseminated to local residents of Potchefstroom and Professional Planners in South Africa. Challenges pertaining to the planning of green spaces were emphasised, despite residents’ willingness to pay more for such green spaces in close proximity to residential areas, according to the cross-tabulations conducted. The research results contributed to the discourse on the economic benefits of green spaces and presented the trends of such benefits within the local context of Potchefstroom. The results emphasised the need to rethink the planning of green spaces within the local context, and provided recommendations on how to reclaim nature into cities from a spatial planning perspective.

The ability of oriental magpies (Pica serica) to solve baited multiple-string problems

Background Baited multiple-string problems are commonly used in avian laboratory studies to evaluate complex cognition. Several bird species possess the ability to use a string pull for obtaining food. Methods We initially tested and trained 11 magpies to determine whether the oriental magpie (Pica sericia) possesses the ability to solve baited multiple-string problems. Eight of the birds obtained the bait by pulling, and were selected for formal multiple-string tasks in the second stage. Second stage tests were divided into seven tasks based on string configurations. Results Only two magpies were able to solve two tasks: one solved the task of parallel strings, and the other solved the task of slanted strings with the bait farther from the middle point between the two strings and selected the short string in the task of long-short strings. When faced with more difficult tasks (i.e., the task of slanted strings with the bait closer to the middle point between the two strings, the task with two crossing strings, and the task of continuity and discontinuity), the birds initially observed the tasks and chose instead to adopt simpler strategies based on the proximity principle, side bias strategies and trial-and-error learning. Our results indicate that the oriental magpie had a partial understanding of the principle of multiple-string problems but adopted simpler strategies.

Perceiving actions before they happen: Psychological dimensions scaffold neural action prediction

The social world buzzes with action. People constantly walk, talk, eat, work, play, snooze, and so on. To interact with others successfully, we need to both recognize their current actions and predict their future actions. Here we used open fMRI data to test the hypothesis that people do both at the same time: when the brain perceives an action, it simultaneously encodes likely future actions. In the scanner, participants watched a naturalistic video. We automatically annotated the actions in that video using a deep learning algorithm, and then constructed a model which could decode participants’ action representations from multivoxel neural activity. Action representations here are defined as locations within a 6-dimensional action space identified by previous work. We hypothesized that within this space, actions are located close to other actions that they are likely to precede or follow. Using this proximity principle, we tested whether a participant’s representation of current actions predicted which actions actually occurred later in the video. Results indicated that neural representations correctly presaged actions later in the video, as-yet unseen by the participant. This finding suggests that the way the brain represents the others’ current behavior gives people an automatic glimpse into others’ future behavior.

Pedagogical grammar as the frameworkof tefl research.part 7. native language syntax acquisition: conceptual complexity

The paper presents the results of the experimental data analysis in the said sphere. Basing upon a high degree of the data similarity, the author makes a conclusion that the native language (English) syntax acquisition is a rule-governed process with its specific stages. Taking into account the said data, the author also offers some assumptions concerning the content of stages in the process of tag questions, negation, passive constructions and relative clauses acquisition in English as a native language, as well as regarding the conditions of fully inverted tag questions emergence in the child’s speech and the potential catalysts of this process. The paper analyses the influence of the minimal proximity principle on the correct interpretation of relative clauses and those containing passive constructions by the young children of different age groups. The author offers an assumption concerning the gradual character of the acquisition of conceptually complicated syntactic categories, in particular, about the dissimilar transfer speed of the passive construction use rule to the verbs indicating actions, on the one hand, and states – on the other. The paper lists the factors affecting the sequence and speed of the syntactic means acquisition, which include the semantic and grammatical complexity, frequency of use and perceptual salience. The author generalizes the stages of syntax acquisition, which include the acquirement of the sentence structure elements linear sequence (where the notions of ‘precedence’ and ‘succession’ are acquired); the acquisition of the rules, which do not take into account the sentence structure; the primary consolidation of sentence elements in terms of their surface features; the identification of sentence components on the basis of the minimum proximity principle; the formulation of the rule, which takes into account the sentence structure and its expansion to a small class of words; the gradual expansion of the latter rule to the entire class of words. The paper outlines the prospects of further research concerning the development of pedagogical grammar, taking into account the abovementioned conclusions.