End-to-end models have achieved impressive results on the task of automatic speech recognition (ASR). For low-resource ASR tasks, however, labeled data can hardly satisfy the demand of end-to-end ...models. Self-supervised acoustic pre-training has already shown its impressive ASR performance, while the transcription is still inadequate for language modeling in end-to-end models. In this work, we fuse a pre-trained acoustic encoder (wav2vec2.0) and a pre-trained linguistic encoder (BERT) into an end-to-end ASR model. The fused model only needs to learn the transfer from speech to language during fine-tuning on limited labeled data. The length of the two modalities is matched by a monotonic attention mechanism without additional parameters. Besides, a fully connected layer is introduced for the hidden mapping between modalities. We further propose a scheduled fine-tuning strategy to preserve and utilize the text context modeling ability of the pre-trained linguistic encoder. Experiments show our effective utilizing of pre-trained modules. Our model achieves better recognition performance on CALLHOME corpus (15 hours) than other end-to-end models.
The innate hypoxic microenvironment of most solid tumors has a major influence on tumor growth, invasiveness, and distant metastasis. Here, a hypoxia‐activated self‐immolative prodrug of paclitaxel ...(PTX2‐Azo) was synthesized and encapsulated by a peptide copolymer decorated with the photosensitizer chlorin e6 (Ce6) to prepare light‐boosted PTX nanoparticle (Ce6/PTX2‐Azo NP). In this nanoparticle, PTX2‐Azo prevents premature drug leakage and realizes specific release in hypoxic tumor microenvironment and the photosensitizer Ce6 not only efficiently generates singlet oxygen under light irradiation but also acts as a positive amplifier to promote the release of PTX. The combination of photodynamic therapy (PDT) and chemotherapy results in excellent antitumor efficacy, demonstrating the great potential for synergistic cancer therapy.
A light‐boosted hypoxia‐activated self‐immolative paclitaxel prodrug nanosystem was designed for synergistic photodynamic therapy and chemotherapy. Upon irradiation, severe hypoxia occurred and amplified the specific release of paclitaxel from prodrug bridged with azobenzene. The nanoparticle showed superior antitumor efficacy with little toxicity to other organs.
In this study, a high‐entropy perovskite oxide Sr(Zr0.2Sn0.2Hf0.2Ti0.2Nb0.2)O3 (SZSHTN) was first introduced to Na0.5Bi0.5TiO3 (NBT) lead‐free ferroelectric ceramics to boost both the ...high‐temperature dielectric stability and energy storage performance. Excellent comprehensive performance was simultaneously obtained in the 0.8NBT–0.2SZSHTN ceramic with high ε′ value (> 2000), wide ε′‐temperature stable range (TCC < 5%, 52.4–362°C), low tanδ value in a wide range (<0.01, 90–341°C) and high energy storage performance (Wrec = 3.52 J/cm3, Wrec and η varies ±6.08% and ±7.4% from 20 to 150°C), which endows it the promising potential to be used in high‐temperature environments.
Copper has been used as an antimicrobial agent long time ago. Nowadays, copper-containing nanoparticles (NPs) with antimicrobial properties have been widely used in all aspects of our daily life. ...Copper-containing NPs may also be incorporated or coated on the surface of dental materials to inhibit oral pathogenic microorganisms. This review aims to detail copper-containing NPs’ antimicrobial mechanism, cytotoxic effect and their application in dentistry.
Plasma wakefield acceleration in the blowout regime is particularly promising for high-energy acceleration of electron beams because of its potential to simultaneously provide large acceleration ...gradients and high energy transfer efficiency while maintaining excellent beam quality. However, no equivalent regime for positron acceleration in plasma wakes has been discovered to date. We show that after a short propagation distance, an asymmetric electron beam drives a stable wakefield in a hollow plasma channel that can be both accelerating and focusing for a positron beam. A high charge positron bunch placed at a suitable distance behind the drive bunch can beam-load or flatten the longitudinal wakefield and enhance the transverse focusing force, leading to high efficiency and narrow energy spread acceleration of the positrons. Three-dimensional quasistatic particle-in-cell simulations show that an over 30% energy extraction efficiency from the wake to the positrons and a 1% level energy spread can be simultaneously obtained. Further optimization is feasible.
•Ground temperature prediction models basing on ANN and LS-SVM are established, respectively.•A new method that correlating fuzzy theory with LS-SVM is proposed.•Fuzzy LS-SVM model is validated to be ...superior to the other two models in both validation accuracy and calculation speed.
Ground source heat pump (GSHP) system has received more and more attentions for its energy-conserving and environmental-friendly properties. Acquisition of the undisturbed ground temperature is the prerequisite for designing of GSHP system. Measurement by burying temperature sensors underground is the conventional means for obtaining the ground temperature data. However, this way is usually time consuming and high investment, and also easily encounter with certain technical difficulties. The rapid development of intelligent computation algorithm provides solutions for many realistic difficult problems. Basing on a great number of the measured data of the ground temperature from two boreholes with 100m depth located in Chongqing, ground temperature prediction models basing on artificial neural network (ANN) and support vector machine based on least square (LS-SVM) are established, respectively. And then, two kinds of validation works, i.e., holdout validation and k-fold validation are conducted toward the two models, respectively. Furthermore, a new method that correlating fuzzy theory with LS-SVM is proposed to solve the big computation burden problem encountered by LS-SVM model. By comparing with the above two models, it is concluded that the newly proposed model can not only improve the calculation speed obviously but also be able to promote the prediction accuracy, especially superior to the single LS-SVM model.
Previous research has shown that consummatory ERP components are sensitive to contextual valence. The present study investigated the contextual valence effect across anticipatory and consummatory ...phases by requiring participants to play a simple gambling task during a gain context and a loss context. During the anticipatory phase, the cue‐P3 was more positive in the gain context compared to the loss context, whereas the stimulus‐preceding negativity (SPN) was comparable across the two contexts. With respect to the consummatory phase, the feedback‐related negativity (FRN) in response to the zero‐value outcome was more negative in the gain versus loss context, whereas the feedback P3 (fb‐P3) in response to the zero‐value outcome was insensitive to contextual valence. These findings suggest that contextual valence effect occurs at a relative early stage of both the reward anticipation and consumption. Moreover, across the gain and loss contexts, the SPN was selectively correlated with the FRN, whereas the cue‐P3 was selectively associated with the fb‐P3, pointing to a close association between the anticipatory and consummatory phases in reward dynamics.
This paper concerns the problem of formation‐containment control for general‐linear multi‐agent systems (MASs) with both communication delays and switching interaction topologies. On the one hand, ...the leaders can communicate with each other to form the desired formation and on the other, the followers need to enter the convex envelope spanned by the multiple leaders. Firstly, by using the neighbouring relative information, formation‐containment protocols are designed for each leader and follower, where an edge‐based state observer is incorporated into the formation‐containment controller to evaluate the whole leaders' state. Secondly, according to the linear matrix inequality technology, an algorithm is given to determine the unknown feedback matrixes in the protocol. Then, based on Lyapunov theory, the formation‐containment error is proved to be convergent and formation feasibility conditions are also presented for the MASs to achieve formation‐containment. Finally, a simulation on several MASs is provided to demonstrate the theoretical results.
•Combining MAE with UADLLME and HPLC for simultaneous determination of pyrethroids residues in Litchi fruit.•Extraction conditions of MAE and UADLLME were optimized by single-factor experiments and ...RSM.•Extraction yields and enrichment factors of pyrethroids were improved by MAE-UADLLME.•The distribution of pyrethroids residues in the pericarp and the pulp was investigated.
A novel method for simultaneous determination of pyrethroids residues in Litchi fruit has been developed by HPLC-UV detection using microwave-assisted extraction (MAE) coupled with ultrasonic-assisted dispersive liquid-liquid microextraction (UADLLME). Extraction conditions of MAE and UADLLME were respectively investigated by single-factor experiments and response surface methodology. Optimized experimental conditions included 310μL of chlorobenzene as extraction solvent, 1.3mL of ethanol as dispersive solvent and 3min of extraction time for UADLLME. In the case of MAE, extraction temperature of 70°C, extraction time of 4min and solvent-to-materials ratio of 40:1 were adopted. Results demonstrated that the proposed method had good performance with linearity of 0.0050–4.98mg/L, recovery of 83.3–91.5%, RSDs below 5.6% and detection limit (LOD) of 1.15–2.46μg/L for six pyrethroids, offering higher extraction efficiency and larger enrichment factor. MAE-UADLLME provided a sensitive and efficient alternative to determination of trace amounts of pesticides residues in food samples.
Abstract
Background
As a hot method in machine learning field, the forests approach is an attractive alternative approach to Cox model. Random survival forests (RSF) methodology is the most popular ...survival forests method, whereas its drawbacks exist such as a selection bias towards covariates with many possible split points. Conditional inference forests (CIF) methodology is known to reduce the selection bias via a two-step split procedure implementing hypothesis tests as it separates the variable selection and splitting, but its computation costs too much time. Random forests with maximally selected rank statistics (MSR-RF) methodology proposed recently seems to be a great improvement on RSF and CIF.
Methods
In this paper we used simulation study and real data application to compare prediction performances and variable selection performances among three survival forests methods, including RSF, CIF and MSR-RF. To evaluate the performance of variable selection, we combined all simulations to calculate the frequency of ranking top of the variable importance measures of the correct variables, where higher frequency means better selection ability. We used Integrated Brier Score (
IBS
) and c-index to measure the prediction accuracy of all three methods. The smaller
IBS
value, the greater the prediction.
Results
Simulations show that three forests methods differ slightly in prediction performance. MSR-RF and RSF might perform better than CIF when there are only continuous or binary variables in the datasets.
For variable selection performance,
When there are multiple categorical variables in the datasets, the selection frequency of RSF seems to be lowest in most cases. MSR-RF and CIF have higher selection rates, and CIF perform well especially with the interaction term.
The fact that correlation degree of the variables has little effect on the selection frequency indicates that three forest methods can handle data with correlation.
When there are only continuous variables in the datasets, MSR-RF perform better. When there are only binary variables in the datasets, RSF and MSR-RF have more advantages than CIF.
When the variable dimension increases, MSR-RF and RSF seem to be more robustthan CIF
Conclusions
All three methods show advantages in prediction performances and variable selection performances under different situations. The recent proposed methodology MSR-RF possess practical value and is well worth popularizing. It is important to identify the appropriate method in real use according to the research aim and the nature of covariates.