In this study, we present extensive attention-based networks with data augmentation methods to participate in the IN-TERSPEECH 2019 ComPareE Challenge, specifically the three Sub-challenges: Styrian Dialect Recognition, Continuous Sleepiness Regression, and Baby Sound Classification. For Styrian Dialect Sub-challenge, these dialects are classified into Northern Styrian (NorthernS), Urban Sytrian (UrbanS), and Eastern Styrian (EasternS). Our proposed model achieves an UAR 49.5% on the test set, which is 2.5% higher than the baseline. For Continuous Sleepiness Sub-challenge, it is defined as a regression task with score range from 1 (extremely alert) to 9 (very sleepy). In this work, our proposed architecture achieves a Spearman correlation 0.369 on the test set, which surpasses the baseline model by 0.026. For Baby Sound Sub-challenge, the infant sounds are classified into canonical babbling, noncanonical babbling, crying, laughing and junk/other, and our proposed augmentation framework achieves an UAR of 62.39% on the test set, which outperforms the baseline by about 3.7%. Overall, our analyses demonstrate that by fusing attention network models with conventional support vector machine benefits the test set robustness, and the recognition rates of these paralinguistic attributes generally improve when performing data augmentation.
We read with interest Mitsiades et al's recent paper 1 purporting to define the intracellular factors regulating tumor necrosis factorrelated apoptosis-inducing ligand (TRAIL) activity in myeloma cells. They demonstrated quite clearly the crucial role for procaspase-8 activation in initiating TRAIL-induced apoptosis and the clear correlation between the efficiency of procaspase-8 activation and the degree of TRAIL-induced apoptosis. Furthermore, they were able to demonstrate that maneuvers capable of inhibiting antiapoptotic proteins or artificially elevating the levels of intracellular procaspase-8 enhanced the apoptosis-inducing capability of TRAIL. This preliminary data demonstrating the "sensitization" of previously TRAIL-resistant myeloma cells with novel agents are very interesting and, if it can be confirmed that these strategies do not also sensitize nonmalignant cells, may well provide a rationale for early-phase clinical trials.Of concern, however, is the claim that the degree of TRAIL resistance of the cell lines studied was associated with a low procaspase-8/cFLIP (FLICE inhibitory protein) ratio. No data are actually presented to support this assertion. Scrutiny of the data that are presented shows that there is no obvious correlation between the physiologic levels of cFLIP and/or procaspase-8 and the degree of TRAIL-induced apoptosis for the cell lines studied. This is clearly exemplified by the "sensitive" cell lines MM1S and RPMI showing quite different immunoblot findings, with high procaspase-8/low cFLIP ratios and low procaspase-8/high cFLIP ratios, respectively. Our own data from 5 authentic myeloma cell lines confirm the crucial relationship between the efficiency of procaspase-8 activation and TRAIL-induced apoptosis (Figure 1) and also show the lack of correlation between the physiologic procaspase-8/cFLIP levels and TRAIL-induced apoptosis (Figure 2).There are 4 known surface receptors for TRAIL, 2 of which (TRAIL-R1 and TRAIL-R2) appear to be capable of inducing apoptosis upon binding of TRAIL. 2 Our understanding of TRAIL-TRAIL receptor interactions beyond this is limited. But the earlier belief that the 2 "decoy" TRAIL receptors (TRAIL-R3 and TRAIL-R4) lacking intracellular death domains somehow protect cells from TRAIL has been shown by our group and others to be incorrect. [3][4][5] For TRAIL to be exploited maximally as an antitumor agent, how it works must be more thoroughly understood, particularly the exact roles of the known TRAIL receptors. Acceptance of the perhaps premature assertion that myeloma-cell TRAIL sensitivity is regulated by physiologic levels of procaspase-8 and/or cFLIP will do little to improve this lack of understanding.
In this study, we present a computational framework to participate in the Self-Assessed Affect Sub-Challenge in the INTER-SPEECH 2018 Computation Paralinguistics Challenge. The goal of this sub-challenge is to classify the valence scores given by the speaker themselves into three different levels, i.e., low, medium, and high. We explore fusion of Bi-directional LSTM with baseline SVM models to improve the recognition accuracy. In specifics, we extract frame-level acoustic LLDs as input to the BLSTM with a modified attention mechanism, and separate SVMs are trained using the standard ComParE 16 baseline feature sets with minority class upsampling. These diverse prediction results are then further fused using a decision-level score fusion scheme to integrate all of the developed models. Our proposed approach achieves a 62.94% and 67.04% unweighted average recall (UAR), which is an 6.24% and 1.04% absolute improvement over the best baseline provided by the challenge organizer. We further provide a detailed comparison analysis between different models.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.