Machine learning systems are becoming increasingly ubiquitous. These systems’s adoption has been expanding, accelerating the shift towards a more algorithmic society, meaning that algorithmically informed decisions have greater potential for significant social impact. However, most of these accurate decision support systems remain complex black boxes, meaning their internal logic and inner workings are hidden to the user and even experts cannot fully understand the rationale behind their predictions. Moreover, new regulations and highly regulated domains have made the audit and verifiability of decisions mandatory, increasing the demand for the ability to question, understand, and trust machine learning systems, for which interpretability is indispensable. The research community has recognized this interpretability problem and focused on developing both interpretable models and explanation methods over the past few years. However, the emergence of these methods shows there is no consensus on how to assess the explanation quality. Which are the most suitable metrics to assess the quality of an explanation? The aim of this article is to provide a review of the current state of the research field on machine learning interpretability while focusing on the societal impact and on the developed methods and metrics. Furthermore, a complete literature review is presented in order to identify future directions of work on this field.

No abstract

No abstract

Water‐soluble polymers containing carboxylic acid, phosphonic acid, and sulfonic acid groups were investigated as polychelatogens under different experimental conditions. The maximum retention capacity (MRC) of these polymers was determined at different pH and polymer–metal ion ratios. The metal ions investigated were: Ag(I), Cu(II), Co(II), Ni(II), and Cr(III). The MRC values of the divalent metal ions were found to be very similar and higher than those of the trivalent cations, and there was an important effect of the pH on the MRC. This indicates that the polymer–metal ion interactions were dominated by electrostatic forces. Copyright © 2003 John Wiley & Sons, Ltd.

Iron toxicity is the most important stressor of rice in many lowland environments worldwide. Rice cultivars differ widely in their ability to tolerate excess iron. A physiological evaluation of iron toxicity in rice was performed using non-invasive photosynthesis, photorespiration and chlorophyll a fluorescence imaging measurements and chlorophyll content determination by SPAD. Four rice cultivars (BR IRGA 409; BR IRGA 412; BRA 041171 and BRA 041152) from the Brazilian breeding programs were used. Fe(2+) was supplied in the nutrient solution as Fe-EDTA (0.019, 4, 7 and 9 mM). Increases in shoot iron content due to Fe(2+) treatments led to changes in most of the non-invasive physiological variables assessed. The reduction in rice photosynthesis can be attributed to stomatal limitations at moderate Fe(2+) doses (4mM) and both stomatal and non-stomatal limitations at higher doses. Photorespiration was an important sink for electrons in rice cultivars under iron excess. A decreased chlorophyll content and limited photochemical ability to cope with light excess were characteristic of the more sensitive and iron accumulator cultivars (BRA 041171 and BRA 041152). Chlorophyll fluorescence imaging revealed a spatial heterogeneity of photosynthesis under excessive iron concentrations. The results showed the usefulness of non-invasive physiological measurements to assess differences among cultivars. The contributions toward understanding the rice photosynthetic response to toxic levels of iron in the nutrient solution are also discussed.

Photon trajectories in incoherent radiation trapping for Doppler, Lorentz and Voigt line shapes under complete frequency redistribution are shown to be Lévy flights. The jump length (r) distributions display characteristic long tails. For the Lorentz line shape, the asymptotic form is a strict power-law r −3/2 while for Doppler the asymptotic is r −2 (ln r) −1/2 . For the Voigt profile, the asymptotic form has always a Lorentz character, but the trajectory is a self-affine fractal with two characteristic Hausdorff scaling exponents.

No abstract

No presente artigo, é descrito um método de concepção para betão auto-compactável reforçado com fibras de aço (BACRFA) de custo competitivo, a ser usado na indústria de pré-fabricação. Uma das questões mais importantes na indústria da pré-fabricação é a da descofragem dos elementos, que deve ser feita com a maior brevidade possível. Deste modo, foi levado a cabo um programa experimental para estimar a influência da idade na resistência e na ductilidade do BACRFA desenvolvido. A relação tensão-abertura de fenda foi determinada com base na relação força-flecha obtida em ensaios realizados segundo as recomendações da RILEM (TC 162-TDF). Foi analisada a influência da idade do BACRFA nos parâmetros de fractura deste material. Foi analisada a influência da idade do BACRFA nos parâmetros de fractura deste material.