Learn decision trees with deep visual primitives

In this paper, we strive to propose a self-interpretable framework, termed PrimitiveTree, that incorporates deep visual primitives condensed from deep features with a conventional decision tree, bridging the gap between deep features extracted from deep neural networks (DNNs) and trees’ transparent decision-making processes. Specifically, we utilize a codebook, which embeds the continuous deep features into a finite discrete space (deep visual primitives) to distill the most common semantic information. The decision tree adopts the spatial location information and the mapped primitives to present the decision-making process of the deep features in a tree hierarchy. Moreover, the trained interpretable PrimitiveTree can inversely explain the constituents of the deep features, highlighting the most critical and semantic-rich image patches attributing to the final predictions of the given DNN. Extensive experiments and visualization results validate the effectiveness and interpretability of our method.     

Determinants of the intention to use non-medical insertable digital devices: The case of Chile and Colombia

In the past, thinking of carrying electronic devices inside our bodies was only posed by non-real scenarios. The emergence of insertable devices has changed this. Since this technology is still in its initial development stages, few studies have investigated factors that influence its acceptance. This paper analyzes the predictors of the intention to use non-medical insertable devices in two Latin American contexts. We used partial least squares structural equation modeling to examine whether six constructs predicted intention to use insertable devices. A questionnaire was administered to undergraduate students located in Colombia and Chile (n = 672). We also examined whether these predictors influenced intention differently for both of them. Four common constructs significantly and positively influenced both Chilean and Colombian respondents to use insertable devices (hedonic motivation, habit, performance expectancy, and social influence). Also, the habit has a complementary mediating effect on the relationship between social influence and behavioral intention. By contrast, effort expectations were a positive and significant predictor, but only among Chilean respondents. Findings suggest that when technologies are emerging, well-known predictors of intention (e.g., performance and effort expectations) are less influential than predictors related to self-efficacy (e.g., habit and hedonic motivation). The use of insertable devices has a significant impact on society. Thus, a better understanding of what motivates their use has implications for both academia and industry.     

Subset representations and eigenvalues of the universal intertwining matrix

We solve a combinatorial question concerning eigenvalues of the universal intertwining endomorphism of a subset representation.     

Off‐grid DoA estimation method of wideband signals

In this paper, an off‐grid direction of arrival (DoA) estimation method is proposed for wideband signals. This method is based on the sparse representation (SR) of the array covariance matrix. Similar to the time domain DoA estimation methods, the correlation function of the sources was assumed to be the same and known. A new measurement vector is obtained using the lower‐left triangular elements of the covariance matrix. The DoAs are estimated by quantizing the entire range of continuous angle space into discrete grid points. However, the exact DoAs may be located between two grid points; therefore, this estimation has errors. The accuracy of DoA estimation is improved by the minimization of the difference between the new measurement vector and its estimated values. Simulation results revealed that the proposed method can enhance the DoA estimation accuracy of wideband signals.     

Springer basic sets and modular Springer correspondence for classical types

We define the notion of basic set data for finite groups (building on the notion of basic set, but including an order on the irreducible characters as part of the structure), and we prove that the Springer correspondence provides basic set data for Weyl groups. Then we use this to determine explicitly the modular Springer correspondence for classical types (defined over a base field of odd characteristic $p$, and with coefficients in a field of odd characteristic $?\ne p$): the modular case is obtained as a restriction of the ordinary case to a basic set. In order to do so, we compare the order on bipartitions introduced by Dipper and James with the order induced by the Springer correspondence. We provide a quick proof, by sorting characters according to the dimension of the corresponding Springer fibre, an invariant which is directly computable from symbols.     

Design of broadband achromatic metasurface device based on phase-change material Ge2Sb2Te5

Based on the phase-change material Ge₂Sb₂Te₅ (GST), achromatic metasurface optical device in the longer-infrared wavelength is designed. With the combination of the linear phase gradient GST nanopillar and the adjustment of the crystalline fraction m value of GST, the polarization insensitive achromic metalenses and beam deflector metasurface within the longer-infrared wavelength 9.5 μm to 13 μm are realized. The design results show that the achromatic metalenses can be focused on the same focal plane within the working waveband. The simulation calculation results show that the full-width at half-maximum (FWHM) of the focusing spot reaches the diffraction limit at each wavelength. In addition, the same method is also used to design a broadband achromatic beam deflector metasurface with the same deflection angle of 19°. The method proposed in this article not only provides new ideas for the design of achromatic metasurfaces, but also provides new possibilities for the integration of optical imaging, optical coding and other related optical systems.     

CN decoration of dibenzofuran modified biphenyl for high triplet energy host for blue phosphorescent organic light-emitting diodes

A strongly electron deficient and high triplet energy host for blue emitters was developed by decorating a dibenzofuran modified biphenyl backbone structure with multiple CN units. Two hosts, 6,6′-bis(6-cyanodibenzo[b,d]furan-4-yl)-[1,1′-biphenyl]-3,3′-dicarbonitrile(CNDBF1) and 2,2′-bis(6-cyanodibenzo[b,d]furan-4-yl)-[1,1′-biphenyl]-4,4′-dicarbonitrile(CNDBF2), were derived from the CN decoration strategy for application in blue organic light-emitting diodes requiring high triplet energy host. They showed high triplet energy above 2.79 eV and acted as the electron transport type host based on the strong electron deficiency. The mixture of the CNDBF1 and CNDBF2 hosts with a hole transport type 3,3′-di(9H-carbazol-9-yl)-1,1′-biphenyl host performed as the exciplex host of a blue phosphor and accomplished high external quantum efficiency of 22.7% in the blue phosphorescent organic light-emitting diodes.