Design of Hierarchical Classifier to Improve Speech Emotion Recognition

Automatic Speech Emotion Recognition(SER)is used to recognize emotion from speech automatically.Speech Emotion recognition is working well in a laboratory environment but real-time emotion recognition has been influenced by the variations in gender,age,the cultural and acoustical background of the speaker.The acoustical resemblance between emotional expressions further increases the complexity of recognition.Many recent research works are concentrated to address these effects individually.Instead of addressing every influencing attribute individually,we would like to design a system,which reduces the effect that arises on any factor.We propose a two-level Hierarchical classifier named Interpreter of responses(IR).Thefirst level of IR has been realized using Support Vector Machine(SVM)and Gaussian Mixer Model(GMM)classifiers.In the second level of IR,a discriminative SVM classifier has been trained and tested with meta information offirst-level classifiers along with the input acoustical feature vector which is used in primary classifiers.To train the system with a corpus of versatile nature,an integrated emotion corpus has been composed using emotion samples of 5 speech corpora,namely;EMO-DB,IITKGP-SESC,SAVEE Corpus,Spanish emotion corpus,CMU's Woogle corpus.The hierarchical classifier has been trained and tested using MFCC and Low-Level Descriptors(LLD).The empirical analysis shows that the proposed classifier outperforms the traditional classifiers.The proposed ensemble design is very generic and can be adapted even when the number and nature of features change.Thefirst-level classifiers GMM or SVM may be replaced with any other learning algorithm.

Self-lubricating epoxy-based composite abradable seal coating eliminating adhesive transfer via hierarchical design

Aluminum-based composite abradable seal coatings are pivotal to improving the efficiency of aero engines or gas turbines.However,the adhesive transfer frequently occurs between metallic blade tips and aluminum-based composite coatings,resulting in engine vibration and even jam.Many past studies had tried to solve this problem by reducing coating hardness,improving lubrication,or strengthening blade tips,but all had failed.In this paper,we proposed a novel epoxy-based composite abradable seal coating,eliminating adhesive transfer by changing metal-to-metal scraping pair to metal-to-polymer scraping pair.The coating was developed via a hierarchical structure design.Large spherical pores were uniformly distributed in the continuous epoxy matrix with fine graphite dispersion.By adding 20 vol.%graphite and 50 vol.%hollow microspheres,a self-lubricating epoxy-based coating of 0.26 friction coefficient with thermal conductivity of 0.28 W/(m·K),coating HR15 Y hardness at 54.8,and bonding strength at 18.7 MPa can be reached.When the metallic blades scrape the epoxy-based composite coating,no adhesive transfer occurs.Besides,a smooth scraped surface is formed by pseudoplastic deformation.This epoxy-based composite abradable seal coating opens a new way to improve the efficiency and reliable operations of air engine compressors.

Interval Mapping of Quantitative Trait Loci Underlying Triploid Endosperm Traits Using F_3 Seeds

A statistical method for mapping quantitative trait loci （QTLs） underlying endosperm traits is proposed. The method is based on a genetic model containing both the direct effects and maternal effects of an endosperm QTL and on an experimental design termed two-stage hierarchical design, in which the trait information is obtained from F3 endosperms and molecular marker information is obtained from F2 plants and F3 embryos （plants）. Results of computer simulations indicate that the method can efficiently map endosperm QTLs and precisely estimate both the direct and maternal effects of endosperm QTLs when the sample size is sufficiently large.

Screening the Optimal Patterned Surfaces Consisting of Cell Morphology Mimicking Micro-pillars and Nanotube Arrays for the Design of Titanium Implants#br#

Micron/nano scale topographic modification has been a significant focus of interest in current titanium(Ti)surface design.However,the influence of micron/nano structured surface on cell or bacterium behavior on the Ti implant has rarely been systematically evaluated.Moreover,except for popular microgrooves,little work has been carried out on the reaction of cells to the bionic structure.In this study,several micro-pillars mimicking cell morphology were prepared on Ti surfaces by lithography and contact printing(ICP)method,and they were further decorated with nanotube arrays by anodization technology.These surface modifications remarkablly increased the surface roughness of pristine Ti surface from 91.17 nm±5.57 nm to be more than 1000 nm,and reduced their water contact angles from 68.3°±0.7°to be 16.9°±2.4°.Then,the effects of these hierarchical micron/nano scale patterns on the behaviors of MG63 osteoblasts,L929 fibroblasts,SCC epithelial cells and P.gingivalis were studied,aiming to evaluate their performance in osseointegration,gingival epithelial sealing and antibacterial ability.Through an innovative scoring strategy,our findings showed that square micro-pillars with 6μm width and 2μm height combined with 85 nm diameter nanotubes was suitable for implant neck design,while square micro-pillars with 3μm width and 3.6μm height combined with 55 nm diameter nanotubes was the best for implant body design.Our study reveals the synergistic effect of the hierarchical micron/nano scale patterns on MG63 osteoblasts,L929 fibroblasts,SCC epithelial cells and P.gingivalis functions.It provides insight into the design of biomedical implant surfaces.

Physical Implementation of the Eight-Core Godson-3B Microprocessor

The Godson-3B processor is a powerful processor designed for high performance servers including Dawning Servers. It offers significantly improved performance over previous Godson-3 series CPUs by incorporating eight CPU cores and vector computing units. It contains 582.6 M transistors within 300 mm2 area in 65 nm technology and is implemented in parallel with full hierarchical design flows. In Godson-3B, advanced clock distribution mechanisms including GALS （Globally Asynchronous Locally Synchronous） and clock mesh are adopted to obtain an OCV tolerable clock network. Custom-designed de-skew modules are also implemented to afford further latency balance after fabrication. The power reduction of Godson- 3B is maintained by MLMM （Multi Level Multi Mode） clock gating and multi-threshold-voltage cells substitution schemes. The highest frequency of Godson-3B is 1.05 GHz and the peak performance is 128 GFlops （double-precision） or 256 GFlops （single-precision） with 40 W power consumption.

A Robust and Power-Effcient SoC Implementation in 65nm

Godson2H is a complex SoC （System-on-Chip） of Godson series, which is a 117mm2, 152 million transistors chip fabricated in 65 nm CMOS LP/GP process technology. It integrates a 1 GHz processor core and abundant high or low speed peripheral IO interfaces. To overcome on-chip-variation problems in deep submicron designs, many methods are adopted in clock tree, and PVT detectors are integrated for debug. To meet the low power constraints in different applications, most of state-of-the-art low power methods are used carefully, such as dynamic voltage and frequency scaling, power gating and aggressive multi-voltage design.