Republican Barbara Jordan as keynote speaker with Bella Abzug in the background

General note: Title and date provided by Bettye Lane.

Demonstration at NYU Law School against guest speaker Chief Justice Burger

General note: Title and date provided by Bettye Lane.

Gays demonstrate at NYU Law School against guest speaker Chief Justice Burger

General note: Title and date provided by Bettye Lane.

Nuclear disarmament demonstration Lesbian and Gay press conference with David Rothenberg on left (dem. Speaker), Betty Powell, and Charlotte Bunch (speaker)

General note: Title and date provided by Bettye Lane.

Young women listening to speaker at the Woman's Salon

Inscriptions: Verso: [stamped] Credit must be given to Freda Leinwand from Monkmeyer Press Photo Service.

From Aari to Zulu: Massively Multilingual Creation of Language Tools using Interlinear Glossed Text

Thesis (Ph.D.)--University of Washington, 2016-08

Functions of code-switching in multilingual student community

Dissertação de Mestrado, Ciências da Linguagem, Faculdade de Ciências Humanas e Sociais, Universidade do Algarve, 2013

Voice conversion with pre-trained representations for audio anonymization

The recording and processing of voice data raises increasing privacy concerns for users and service providers. One way to address these issues is to move processing on the edge device closer to the recording so that potentially identifiable information is not transmitted over the internet. However, this is often not possible due to hardware limitations. An interesting alternative is the development of voice anonymization techniques that remove individual speakers characteristics while preserving linguistic and acoustic information in the data. In this work, a state-of-the-art approach to sequence-to-sequence speech conversion, ini- tially based on x-vectors and bottleneck features for automatic speech recognition, is explored to disentangle the two acoustic information using different pre-trained speech and speakers representation. Furthermore, different strategies for selecting target speech representations are analyzed. Results on public datasets in terms of equal error rate and word error rate show that good privacy is achieved with limited impact on converted speech quality relative to the original method.

A novel polar-based human face recognition computational model

Motivated by a recently proposed biologically inspired face recognition approach, we investigated the relation between human behavior and a computational model based on Fourier-Bessel (FB) spatial patterns. We measured human recognition performance of FB filtered face images using an 8-alternative forced-choice method. Test stimuli were generated by converting the images from the spatial to the FB domain, filtering the resulting coefficients with a band-pass filter, and finally taking the inverse FB transformation of the filtered coefficients. The performance of the computational models was tested using a simulation of the psychophysical experiment. In the FB model, face images were first filtered by simulated V1- type neurons and later analyzed globally for their content of FB components. In general, there was a higher human contrast sensitivity to radially than to angularly filtered images, but both functions peaked at the 11.3-16 frequency interval. The FB-based model presented similar behavior with regard to peak position and relative sensitivity, but had a wider frequency band width and a narrower response range. The response pattern of two alternative models, based on local FB analysis and on raw luminance, strongly diverged from the human behavior patterns. These results suggest that human performance can be constrained by the type of information conveyed by polar patterns, and consequently that humans might use FB-like spatial patterns in face processing.

A Functional Verification Methodology Based on Parameter Domains for Efficient Input Stimuli Generation and Coverage Modeling

Modern Integrated Circuit (IC) design is characterized by a strong trend of Intellectual Property (IP) core integration into complex system-on-chip (SOC) architectures. These cores require thorough verification of their functionality to avoid erroneous behavior in the final device. Formal verification methods are capable of detecting any design bug. However, due to state explosion, their use remains limited to small circuits. Alternatively, simulation-based verification can explore hardware descriptions of any size, although the corresponding stimulus generation, as well as functional coverage definition, must be carefully planned to guarantee its efficacy. In general, static input space optimization methodologies have shown better efficiency and results than, for instance, Coverage Directed Verification (CDV) techniques, although they act on different facets of the monitored system and are not exclusive. This work presents a constrained-random simulation-based functional verification methodology where, on the basis of the Parameter Domains (PD) formalism, irrelevant and invalid test case scenarios are removed from the input space. To this purpose, a tool to automatically generate PD-based stimuli sources was developed. Additionally, we have developed a second tool to generate functional coverage models that fit exactly to the PD-based input space. Both the input stimuli and coverage model enhancements, resulted in a notable testbench efficiency increase, if compared to testbenches with traditional stimulation and coverage scenarios: 22% simulation time reduction when generating stimuli with our PD-based stimuli sources (still with a conventional coverage model), and 56% simulation time reduction when combining our stimuli sources with their corresponding, automatically generated, coverage models.

Conformationally constrained macrocycles that mimic tripeptide beta-strands in water and aprotic solvents

The beta-strand conformation is unknown for short peptides in aqueous solution, yet it is a fundamental building block in proteins and the crucial recognition motif for proteolytic enzymes that enable formation and turnover of all proteins. To create a generalized scaffold as a peptidomimetic that is preorganized in a beta-strand, we individually synthesized a series of 15-22-membered macrocyclic analogues of tripeptides and analyzed their structures. Each cycle is highly constrained by two trans amide bonds and a planar aromatic ring with a short nonpeptidic linker between them. A measure of this ring strain is the restricted rotation of the component tyrosinyl aromatic ring (DeltaG(rot) 76.7 kJ mol(-1) (16-membered ring), 46.1 kJ mol(-1) (17-membered ring)) evidenced by variable temperature proton NMR spectra (DMF-d(7), 200-400 K). Unusually large amide coupling constants ((3)J(NH-CHalpha) 9-10 Hz) corresponding to large dihedral angles were detected in both protic and aprotic solvents for these macrocycles, consistent with a high degree of structure in solution. The temperature dependence of all amide NH chemical shifts (Deltadelta/T7-12 ppb/deg) precluded the presence of transannular hydrogen bonds that define alternative turn structures. Whereas similar sized conventional cyclic peptides usually exist in solution as an equilibrium mixture of multiple conformers, these macrocycles adopt a well-defined beta-strand structure even in water as revealed by 2-D NMR spectral data and by a structure calculation for the smallest (15-membered) and most constrained macrocycle. Macrocycles that are sufficiently constrained to exclusively adopt a beta-strand-mimicking structure in water may be useful pre-organized and generic templates for the design of compounds that interfere with beta-strand recognition in biology.

Experiências formativas no início da docência mediadas pelo PIBID educação física da Universidade Estadual de Feira de Santana, Bahia

O presente trabalho objetiva compreender como o Coordenador, os Supervisores e os Bolsistas do PIBID Educação Física da Universidade Estadual de Feira de Santana (UEFS) no município de Feira de Santana – Bahia percebem os processos de iniciação à docência na formação inicial de professores de Educação Física no PIBID Educação Física – UEFS. Por meio da Pesquisa Narrativa, numa perspectiva de pesquisa e de formação, oportunizamos aos sujeitos da pesquisa entrevistas narrativas compostas pela reflexividade crítica necessária na constituição da docência. Analisamos a produção acadêmica sobre a temática da iniciação à docência no tempo da formação inicial nos bancos de dados da Associação Nacional de Pós-Graduação e Pesquisa em Educação (ANPED) e do Congresso Brasileiro de Ciências do Esporte (CONBRACE) e discutimos com Antonio Nóvoa, Carlos Marcelo e Molina Neto sobre as concepções de formação e as experiências vividas no início da docência e buscamos em Walter Benjamin e Jorge Larrosa ancoragens para o entendimento do conceito de experiência. Os resultados apontam para o reconhecimento do PIBID como um programa que aproxima do campo de atuação interferindo na escolha pela docência. Sinalizam que o programa possibilita a articulação entre a universidade e a escola fortalecendo a perspectiva de formação que acontece coletivamente. Confirmam que o PIBID permite a constatação dos desafios presentes na escola e da necessidade de transformação desta realidade com a articulação dos saberes acadêmicos e os escolares. Neste sentido, constatamos a afirmação da entrada na carreira, como uma experiência formativa que intensifica o desenvolvimento profissional docente e, portanto, deve ser concebida como política nacional institucionalizada de iniciação à docência.

A negação no português falado em Vitória/ES

Todas as línguas possuem algum recurso para expressar a negação verbal, porém cada uma apresenta estratégias próprias para sua realização. No português brasileiro (PB), há três estratégias de negação: 1) pré-verbal (Não+SV); 2) dupla negação (Não+SV+Não) e 3) pós-verbal (SV+Não). À luz da Sociolinguística Variacionista e com base na amostra PortVix (Português Falado na Cidade de Vitória), que tem por parâmetros sociais o gênero/sexo do falante, sua faixa etária e seu nível de escolaridade, o presente trabalho analisa a variação no uso das estruturas de negação no português falado na cidade de Vitória/ES, a fim de situar, a partir desse fenômeno, a variedade capixaba no cenário do PB. Também toma por base a proposta de Schwenter (2005) de que as três variantes se alternam apenas quando o conteúdo negado é ativado no discurso. Sendo assim, se a proposição negada apresentar um estatuto de uma informação nova, apenas a negação pré-verbal pode ser empregada. Desse modo, em nossa pesquisa, buscamos entender quais fatores influenciam a alternância das formas de negação e verificar os contextos linguístico-discursivos que comportam essa variação. Ao confrontarmos nossos resultados com os de outras pesquisas, observamos que a dupla negação é bastante produtiva na fala capixaba, representando 21,1% de um total de 2263 dados. Ao realizarmos rodadas em que foram amalgamadas duas variantes e contrapostas a uma outra, foram selecionados pelo programa Goldvarb X (SANKOFF; TAGLIAMONTE; SMITH, 2005) e, portanto, considerados estatisticamente relevantes para a dupla negação, os seguintes fatores: as sequências dialogais, a ausência de reforço negativo, a ausência de marcadores conversacionais e as orações absolutas. Para a negação pós-verbal, foram selecionadas as seguintes variáveis: as proposições negadas diretamente ativadas e as sequências dialogais. Para a negação pré-verbal, os fatores estatisticamente relevantes foram: as sequências narrativas e as argumentativas, a presença de reforço negativo, a presença de marcadores conversacionais, as orações principais e o gênero masculino. Os resultados revelaram que a variação no uso das estruturas negativas é um fenômeno marcadamente discursivo, mas também com atuação de alguns fatores sintáticos.

Derivative-free optimization and filter methods to solve nonlinear constrained problems

In real optimization problems, usually the analytical expression of the objective function is not known, nor its derivatives, or they are complex. In these cases it becomes essential to use optimization methods where the calculation of the derivatives, or the verification of their existence, is not necessary: the Direct Search Methods or Derivative-free Methods are one solution. When the problem has constraints, penalty functions are often used. Unfortunately the choice of the penalty parameters is, frequently, very difficult, because most strategies for choosing it are heuristics strategies. As an alternative to penalty function appeared the filter methods. A filter algorithm introduces a function that aggregates the constrained violations and constructs a biobjective problem. In this problem the step is accepted if it either reduces the objective function or the constrained violation. This implies that the filter methods are less parameter dependent than a penalty function. In this work, we present a new direct search method, based on simplex methods, for general constrained optimization that combines the features of the simplex method and filter methods. This method does not compute or approximate any derivatives, penalty constants or Lagrange multipliers. The basic idea of simplex filter algorithm is to construct an initial simplex and use the simplex to drive the search. We illustrate the behavior of our algorithm through some examples. The proposed methods were implemented in Java.