1,721,023 research outputs found
SEPARATION POWER OF EQUIVARIANT NEURAL NETWORKS
No full textThe separation power of a machine learning model refers to its ability to distinguish between different inputs and is often used as a proxy for its expressivity. Indeed, knowing the separation power of a family of models is a necessary condition to obtain fine-grained universality results. In this paper, we analyze the separation power of equivariant neural networks, such as convolutional and permutation-invariant networks. We first present a complete characterization of inputs indistinguishable by models derived by a given architecture. From this results, we derive how separability is influenced by hyperparameters and architectural choices-such as activation functions, depth, hidden layer width, and representation types. Notably, all non-polynomial activations, including ReLU and sigmoid, are equivalent in expressivity and reach maximum separation power. Depth improves separation power up to a threshold, after which further increases have no effect. Adding invariant features to hidden representations does not impact separation power. Finally, block decomposition of hidden representations affects separability, with minimal components forming a hierarchy in separation power that provides a straightforward method for comparing the separation power of models
Genetic factors and early parent-infant interactions shape Instagram users' social behaviour
No full textThe interaction between parental care and oxytocin receptor gene affect social behaviours in adulthood. Serotonin transporter gene polymorphisms and early parent-infant interaction influence adults’ relationship with intimate partners. However, human conspecifics socially act towards the real and virtual world in displaying multiple and complex social interactions. Hence, the mechanisms underlying the virtual social relationship in humans remain largely unknown. This research investigates how the interaction between genetic factors (three polymorphisms from the oxytocin receptor, OXTR, and the serotonin transporter, 5-HTTLPR), parental bonding patterns (PBI) and adults’ interactional patterns (ASQ, ECR-R) shape Instagram virtual social behaviour. We hypothesized that adult Instagram users with a genetic risk factor and exposed to a negative social environment will show an enhanced virtual social activity compared to less vulnerable genetic carriers. To this purpose, self-reported questionnaires were used to evaluate participants’ (n = 57; 17 males) parental bonding patterns (Parental Bonding Instrument, formed by four subscales Maternal Care-Paternal Care-Maternal Overprotection-Paternal Overprotection), interactional patterns towards peers (Attachment Style Questionnaire formed by five subscales Confidence-Discomfort in Intimacy-Relationship as Secondary-Need for Approval-Worry about Relationship) and the quality of bond with intimate partners (Experience in Close Relationship Scale formed by two subscales Anxiety-Avoidance). Instagram activities were collected applying an automatic Python algorithm through which four indexes were extracted: number of followings, followers, published posts and tags from other users’ posts. To assess participants’ genetic predispositions, buccal mucosa cell samples were collected, and regions OXTR/rs2254298 (G/G homozygotes vs. A carriers), OXTR/rs53576 (A/A homozygotes vs. G carriers) and 5-HTTLPR/rs255331 (G/G homozygotes vs. A carriers) were genotyped. We expect that the main potential effects on the Instagram measured variables and on a “Social Recognition Index” (Followers/Following ratio) could represent an indirect pathway through which genes and parental behaviour interact to shape social media interactions
A Framework for Verifiable and Auditable Collaborative Anomaly Detection
Full text linkCollaborative and Federated Leaning are emerging approaches to manage cooperation between a group of agents for the solution of Machine Learning tasks, with the goal of improving each agent's performance without disclosing any data. In this paper we present a novel algorithmic architecture that tackle this problem in the particular case of Anomaly Detection (or classification of rare events), a setting where typical applications often comprise data with sensible information, but where the scarcity of anomalous examples encourages collaboration. We show how Random Forests can be used as a tool for the development of accurate classifiers with an effective insight-sharing mechanism that does not break the data integrity. Moreover, we explain how the new architecture can be readily integrated in a blockchain infrastructure to ensure the verifiable and auditable execution of the algorithm. Furthermore, we discuss how this work may set the basis for a more general approach for the design of collaborative ensemble-learning methods beyond the specific task and architecture discussed in this paper
Genetic predispositions and early parent-child relationship are related to Instagram behaviour
No full textpredispositions can influence in the same way online as well as offline virtual social abilities in adulthood. To better understand the interplay that influences social behaviour, this research explores the interaction between the oxytocin receptor (OXTr) gene and early parent-child relationship in shaping Instagram users' social attitudes.
Self-report questionnaire Parental Bonding Instrument was used to evaluate participants' (N = 57; 41 females) early parental care. Individuals' social parameters (number of followings, followers, posts) and indexes (the Social Desirability Index or SDI estimated by the ratio between followers' and followings' number) were collected using a Python tool and standardized by z-scores. Buccal mucosa cell samples were collected to assess participants' genetic predispositions. Regions OXTr/rs2254298 (G/G homozygotes vs A-carriers), OXTr/rs53576 (A/A homozygotes vs G-carriers) and 5-HTTLPR/rs255331 (T/T homozygotes vs C-carriers) were analyzed.
We hypothesized a cross interaction between genetic groups and parental bonding scores on the number of followings and posts. We identified a gene-environment interaction for OXTr/rs2254298 on posts' number. Moreover, two interaction effects were detected on the SDI between maternal overprotection and OXTr/rs2254298 and between maternal care and 5-HTTLPR/rs255331
Glass ingots, raw glass chunks, glass wastes and vessels from fifth century AD Palatine Hill (Rome, Italy)
No full textThe research focused on a collection of 20 glass fragments, including raw glass chunks and ingots, wastes and vessels found at the Palatine Hill in Rome and dated to the first half of the fifth century AD. The analyses performed by scanning electron microscopy (SEM)-energy-dispersive spectrometry (EDS), electron microprobe (EMPA), inductively coupled plasma mass spectrometry (ICP-MS), laser ablation-ICP-MS (LA-ICP-MS) and X-ray diffraction (XRD) showed that all samples are constituted by natron-based soda–lime–silica glass. De/colouring agents were lead stannates (yellow brownish), copper and lead antimonates (green), different Fe2+/Fe3+ ratios (green, light green and yellow-green), Fe2+ (prevailing over Fe3+; aqua blue), cobalt (blue), metallic copper (reddish) and manganese (colourless). As for provenance, two samples were of Levantine provenance, nine samples were likely of Egyptian origin (HIMT glass) and, similarly, seven samples (or maybe nine, adding ingots nos. 1–2) were likely of North African provenance (HIMT/RNCBGY 1 glass)
Valutazione dello sforzo in Quarter Horse impegnati in diverse discipline western: risultati preliminari
No full textImplicit associations among genetics, early care experiences, and adult relationships to social media behaviour.
No full textIntroduction
Humans are social beings whose experiences are ruled by complex interactions between genetic and environmental factors [1]. In-person social interactions represent the first means of sharing beliefs and making comparisons with others. Caregivers embody the first actors in social education. The attachment with one’s parents can be stable across development and can affect adult relationships [2]. Nowadays, social interaction has rapidly expanded through virtual environments like social networking sites (SNS). However, the quality and the frequency of online and in-person interactions may differ across individuals. We explored these variations in online interactions from the perspective of behavioural genetics, investigating whether the quality of early care experiences and of one’s adult relationships could affect the frequency in the online sociability of users who are genetically sensitive to experiences.
Method
The measures on the quality of in-person relationships we focused on are the parent-child attachment and the adult attachment, as recalled by Singaporean participants in the Parental Bonding Instrument (maternal care, maternal overprotection, paternal care, paternal overprotection; N = 57; 41 females) [3] and the Attachment Style Questionnaire (confidence, discomfort with closeness, relationships as secondary, need for approval, preoccupation with relationships; N = 58; 42 females) [4].
The genetic factors we examined are the polymorphisms within the regions rs2254298 (G/G homozygotes, A-carriers), rs53576 (A/A homozygotes, G-carriers) of the oxytocin receptor gene and the region rs25531 (T/T homozygotes, C-carriers) of the serotonin transporter gene. According to the sensitivity hypothesis [5], a genetic component can be either a risk or a protective factor to social distress in relationships.
From Instagram, we extracted three variables as a proxy for the frequency of online interactions: the number of a) published posts, b) people that the user follows ("followings") and c) followers. The Social Desirability Index (SDI) was estimated as the ratio of followers to followings.
Results
We hypothesized multiple interaction effects between genetic groups and attachment scores on Instagram parameters, independent of the gender. We found a gene*environment interaction for rs2254298 on the number of Instagram posts [6]. In line with our expectations, participants with a genetic risk factor (A-carriers) and a past of poor paternal care produced less Instagram posts than those without this risk factor (G/G genotype). A significant interaction between maternal overprotection and rs2254298 also emerged for Instagram SDI. Interestingly, the same pattern emerged between maternal care and rs25531 on the SDI [7]. Specifically, users genotypically more sensitive to environmental influences (A-carriers; T/T genotype) exhibited an increasing trend in the Instagram SDI when they experienced positive maternal caregiving. In contrast, a decreasing trend of the same index was observed for those who recalled a negative relationship with their mother. T/T homozygotes also showed a greater number of Instagram followings than C-carriers when they reported a high level of confidence towards people in adulthood [8].
Conclusion
A positive relationship with parents, as well as a high level of trust towards peers, affects the way genetically vulnerable Instagram users post, search for other users and increase their social likability
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
- …
