Gender | EPFL Graph Search

Gender includes the social, psychological, cultural and behavioral aspects of being a man, woman, or other gender identity. Depending on the context, this may include sex-based social structures (i.e. gender roles) and gender expression. Most cultures use a gender binary, in which gender is divided into two categories, and people are considered part of one or the other (boys/men and girls/women); those who are outside these groups may fall under the umbrella term non-binary. Some societies have specific genders besides "man" and "woman", such as the hijras of South Asia; these are often referred to as third genders (and fourth genders, etc.). Most scholars agree that gender is a central characteristic for social organization. In the mid-20th century, a terminological distinction in modern English (known as the sex and gender distinction) between biological sex and gender began to develop in the academic areas of psychology, sexology, and feminism. Before the mid-20th century, it was

Care. Emerging Natures

Pieter Versteegh

The work addresses research by design as a subject and a method. It explores theorizations, experimentations, and verifications, all of them engaged in writing, drawing and building. It is discovery-led, and starts from a blurry "state of the arts". It oscillates between architectural processes, products and performance without considering any of them isolated as research. Its relevance emerges. It is the design process that constructs, delimits the research corpus and produces architectural insight and knowledge. The first part addresses architectural meaning. It engages interdisciplinary connections: structuralism in linguistics and philosophy, color theory, and a displacement into deconstruction. It concludes that, although an architectural design process can deliberately try to produce and uphold determinate architectural meaning, such (objective) meaning is immediately deviated, notably by the appearance of architectural usages, scale, and possibilities of appropriation. The second part investigates deviation as a substantial part of architectural creation. If noise is irreducible, it must also structure architectural creation. It connects to domains resisting to reductionist thought: complexity sciences, literary critics, archaeological processes, and theorizations dealing with the secondary such as psychoanalysis and gender studies. The work is marked by a dilemma related to empowerment: the originality and authenticity of design and its delivery at a certain moment always seem to prohibit inhabitation, appropriation. The secondary, subject to dominant, single authorship, becomes primary. Architecture seems resistant to modernity in the Kantian sense. The work re-centers on psychiatry, gender studies and (health) care: the understanding of natures rather than appearance, structure or essence. In the third part, it moves from the production of difference to inscriptions of memories and future possibilities; from design as making to design as the initiation and care of a shared transformation process, accompaniment of differance. The understanding of architectural reality shifts from the finality of object to a process beyond control, co-construction between different authors/users occurring in minor events, in a very local and specific context. The work proposes a design process driven by care, encounter of authors and users in quality, accompanying different and specific emerging natures of (territorial) identity.

EPFL2012

On Learning to Identify Genders from Raw Speech Signal Using CNNs

Selen Hande Kabil, Hannah Muckenhirn

Automatic Gender Recognition (AGR) is the task of identifying the gender of a speaker given a speech signal. Standard approaches extract features like fundamental frequency and cepstral features from the speech signal and train a binary classifier. Inspired from recent works in the area of automatic speech recognition (ASR), speaker recognition and presentation attack detection, we present a novel approach where relevant features and classifier are jointly learned from the raw speech signal in end-to-end manner. We propose a convolutional neural networks (CNN) based gender classifier that consists of: (1) convolution layers, which can be interpreted as a feature learning stage and (2) a multilayer perceptron (MLP), which can be interpreted as a classification stage. The system takes raw speech signal as input, and outputs gender posterior probabilities. Experimental studies conducted on two datasets, namely AVspoof and ASVspoof 2015, with different architectures show that with simple architectures the proposed approach yields better system than standard acoustic features based approach. Further analysis of the CNNs show that the CNNs learn formant and fundamental frequency information for gender identification.

ISCA-INT SPEECH COMMUNICATION ASSOC2018

Doppelganger-based training: Imitating our virtual self to accelerate interpersonal skills learning

Marion Sandy Curdy, João Pedro de Matos Rodrigues, Maria del Carmen Sandi Perez

Interpersonal skills require mastering a wide range of competencies such as communication and adaptation to different situations. Effective training includes the use of videos in which role models perform the desired behaviours such that trainees can learn through behavioural mimicry. However, new technologies allow new ways of designing training. In the present study, given that virtual reality is emerging as a valuable training setting, we compare two different demonstration conditions within virtual reality by investigating the extent to which the use of doppelgangers as role models can boost trainees' interpersonal skills development as compared to a role model that does not resemble the trainees. We also assess trainees' level of self-efficacy and gender as potential moderators in this relationship. Participants delivered a speech in front of a virtual audience twice. Before delivering their second speech, they watched a role model giving a speech in front of the same audience. The role model was either their doppelganger or an avatar of the same gender depending on the condition they were randomly assigned to. Results showed that the doppelganger-based training was the most beneficial for male trainees low in self-efficacy. These findings have important implications for training design, suggesting that doppelganger-based training might be effective only for a specific subset of trainees.

2021

Care. Emerging Natures

Pieter Versteegh

EPFL2012