Search

Abstract Handling emotions in human‐computer dialogues has emerged as a challenging task which requires artificial intelligence systems to generate emotional responses by jointly perceiving the emotion involved in the input posts and incorporating it into the generation of semantically coherent and emotionally reasonable responses. However, most previous works generate emotional responses solely from input posts, which do not take full advantage of the training corpus and suffer from generating generic responses. In this study, we introduce a hierarchical semantic‐emotional memory module for emotional conversation generation (called HSEMEC), which can learn abstract semantic conversation patterns and emotional information from the large training corpus. The learnt semantic and emotional knowledge helps to enrich the post representation and assist the emotional conversation generation. Comprehensive experiments on a large real‐world conversation corpus show that HSEMEC can outperform the strong baselines on both automatic and manual evaluation. For reproducibility, we release the code and data publicly at: https://github.com/siat‐nlp/HSEMEC‐code‐data.

Abstract Personalized voice triggering is a key technology in voice assistants and serves as the first step for users to activate the voice assistant. Personalized voice triggering involves keyword spotting (KWS) and speaker verification (SV). Conventional approaches to this task include developing KWS and SV systems separately. This paper proposes a single system called the multi-task deep cross-attention network (MTCANet) that simultaneously performs KWS and SV, while effectively utilizing information relevant to both tasks. The proposed framework integrates a KWS sub-network and an SV sub-network to enhance performance in challenging conditions such as noisy environments, short-duration speech, and model generalization. At the core of MTCANet are three modules: a novel deep cross-attention (DCA) module to integrate KWS and SV tasks, a multi-layer stacked shared encoder (SE) to reduce the impact of noise on the recognition rate, and soft attention (SA) modules to allow the model to focus on pertinent information in the middle layer while preventing gradient vanishing. Our proposed model demonstrates outstanding performance in the well-off test set, improving by 0.2%, 0.023, and 2.28% over the well-known SV model emphasized channel attention, propagation, and aggregation in time delay neural network (ECAPA-TDNN) and the advanced KWS model Convmixer in terms of equal error rate (EER), minimum detection cost function (minDCF), and accuracy (Acc), respectively.

IntroductionAlzheimer’s disease (AD) is the most prevalent neurodegenerative disease characterized by extracellular senile plaques including amyloid-β peptides and intracellular neurofibrillary tangles consisting of abnormal Tau. Depression is one of the most common neuropsychiatric symptoms in AD, and clinical evidence demonstrates that depressive symptoms accelerate the cognitive deficit of AD patients. However, the underlying molecular mechanisms of depressive symptoms present in the process of AD remain unclear.MethodsDepressive-like behaviors and cognitive decline in hTau mice were induced by chronic restraint stress (CRS). Computational prediction and molecular experiments supported that an asparagine endopeptidase (AEP)-derived Tau fragment, Tau N368 interacts with peroxisome proliferator-activated receptor delta (PPAR-δ). Further behavioral studies investigated the role of Tau N368-PPAR-δ interaction in depressive-like behaviors and cognitive declines of AD models exposed to CRS.ResultsWe found that mitochondrial dysfunction was positively associated with depressive-like behaviors and cognitive deficits in hTau mice. Chronic stress increased Tau N368 and promoted the interaction of Tau N368 with PPAR-δ, repressing PPAR-δ–mediated transactivation in the hippocampus of mice. Then we predicted and identified the binding sites of PPAR-δ. Finally, inhibition of AEP, clearance of Tau N368 and pharmacological activation of PPAR-δ effectively alleviated CRS-induced depressive-like behaviors and cognitive decline in mice.ConclusionThese results demonstrate that Tau N368 in the hippocampus impairs mitochondrial function by suppressing PPAR-δ, facilitating the occurrence of depressive-like behaviors and cognitive decline. Therefore, our findings may provide new mechanistic insight in the pathophysiology of depression-like phenotype in mouse models of Alzheimer’s disease.

The Voronoi diagram is a fundamental data structure for spatial neighborhood relations used in various fields related to geometric information. Taking four asphalt mixtures as the research object, this study established the skeleton contact Voronoi diagram model to characterize the contact meso-structure between coarse aggregates, and extracted the quantitative information of skeleton contact space. The results show that the skeleton contact Voronoi diagram model can represent important spatial information of asphalt mixtures from the meso-structure including the homogeneity of asphalt mixtures, the extrusion relationship between coarse and fine aggregates, and the skeleton contact distribution characteristics, etc. The shape of the connected tree reflects the coarse aggregate contact distribution and load transfer trend, and the order of the connected tree reflects the contact connectivity strength of the skeleton. We used Voronoi diagram area Vs, connected tree area TVs and the total length of contact points L to quantitatively evaluate the skeleton contact of asphalt mixtures. In the design of asphalt mixtures with dense skeletons, the evaluation criteria of skeleton contact characteristics include Vs (30–180 mm2) proportion ≥75 %, Tr-1 and Tr-2 proportion ≤75 % and L ≥70. The results provide new ideas to promote the quality of asphalt mixtures and pavement construction.

The current understanding of the influence of smoke suppressants on the smoke inhibition effect and working performance of different types of asphalt is limited. In this study, portable gas detector (PGD) tests were used to investigate the influence of the smoke suppressant on different types of asphalt and the smoke purification effect. Subsequently, gas chromatography-mass spectrometry (GC-MS) was used to analyze the influence of the smoke suppressant on the volatile organic compounds (VOCs) in different types of asphalt fume, so as to compare and verify the PGD tests results. Finally, scanning electron microscopy and energy dispersive spectrometry were used to observe and analyze the influence of the smoke suppressant on the microstructure and element composition of different types of asphalt. The results show that the smoke suppressant can purify various harmful gases volatilized from different types of asphalt, but the effect and the optimal dosage of smoke suppressant are slightly different. The purification rate of the smoke suppressant for SO2 or H2S in the fumes of various asphalts can reach >70% and the NO and VOC purification rates can reach >40%. The influence of the smoke suppressant on the VOC purification rate of different asphalts detected through the GC-MS and PGD tests is similar. The smoke suppressant has little effect on the asphalt surface microstructure and on the changes in the element content in asphalt, indicating that it has no effect on the main properties of the asphalt itself. This study is helpful to further understand the relationship between smoke suppressants and the smoke inhibition effect and the working performance of different types of asphalt. Additionally, it provides basis and guidance for the development of low-smoke asphalt with a better emission reduction effect, which has a certain guiding and application significance.

Insulin-resistance (IR) is one of the most important precursors of type 2 diabetes (T2D). Recent evidence suggests an association of depression with the onset of T2D. Accumulating evidence shows that depression and T2D share common biological origins, and DNA methylation examination might reveal the link between lifestyle, disease risk, and potential therapeutic targets for T2D. Here we hypothesize that integrative mining of IR and depression cohort data will facilitate predictive biomarkers identification for T2D. We utilized a newly proposed method to extract gene-level information from probe level data on genome-wide DNA methylation array. We identified a set of genes associated with IR and depression in clinical cohorts. By overlapping the IR-related nutraceutical-gene network with depression networks, we identified a common subnetwork centered with Vitamin D Receptor (VDR) gene. Preliminary clinical validation of gene methylation set in a small cohort of T2D patients and controls was established using the Sequenome matrix-assisted laser desorption ionization-time flight mass spectrometry. A set of sites in the promoter regions of VDR showed a significant difference between T2D patients and controls. Using a logistic regression model, the optimal prediction performance of these sites was AUC = 0.902，and an odds ratio = 19.76. Thus, monitoring the methylation status of specific VDR promoter region might help stratify the high-risk individuals who could potentially benefit from vitamin D dietary supplementation. Our results highlight the link between IR and depression, and the DNA methylation analysis might facilitate the search for their shared mechanisms in the etiology of T2D.

Oxide dispersion strengthened (ODS) superalloys are widely used in aerospace and nuclear reactors due to their exceptional creep strength and fatigue crack resistance at elevated temperature. In this work, an ODS NiCrFeY alloy for high temperature applications was fabricated by laser powder bed fusion (L-PBF). The microstructure and mechanical properties at room and elevated temperature of the L-PBF processed ODS NiCrFeY alloy were investigated. It was demonstrated that equiaxed grains with large fraction of low angle grain boundaries and cellular structures with high density of dislocations were obtained in the as-built alloy. Dispersed Y2O3 nanoparticles were in-situ formed in the ODS NiCrFeY alloy. Most of the Y2O3 nanoparticles were located at the cellular walls and pinned with a large number of dislocations. Due to the strengthening effects by dispersed Y2O3 nanoparticles and fine cellular structures, the ODS NiCrFeY alloy showed high ultimate tensile strength of 810, 560 and 250 ​MPa at room temperature, 600 ​°C and 800 ​°C, respectively.

In the Knowledge Grounded Dialogue (KGD) generation, the explicit modeling of instance-variety of knowledge specificity and its seamless fusion with the dialogue context remains challenging. This paper presents an innovative approach, the Knowledge Interpolated conditional Variational auto-encoder (KIV), to address these issues. In particular, KIV introduces a novel interpolation mechanism to fuse two latent variables: independently encoding dialogue context and grounded knowledge. This distinct fusion of context and knowledge in the semantic space enables the interpolated latent variable to guide the decoder toward generating more contextually rich and engaging responses. We further explore deterministic and probabilistic methodologies to ascertain the interpolation weight, capturing the level of knowledge specificity. Comprehensive empirical analysis conducted on the Wizard-of-Wikipedia and Holl-E datasets verifies that the responses generated by our model performs better than strong baselines, with notable performance improvements observed in both automatic metrics and manual evaluation.

A current popular feature extraction method of classifying cognitive states and task engagements from electroencephalographs (EEG) is common spatial patterns (CSP). However, the classical CSP only focuses on the correlation between the signals and ignores all characteristics of the signals in the time domain and the frequency domain. In this paper we propose Common cross-spectral patterns (CCSP) a novel EEG feature extraction method for combining spectral and spatial patterns based on cross-spectral density (CSD) to overcome the disadvantages of classical CSP. In CCSP method. the cross-power matrices (CPMs) are extracted to measure the spatial correlation of each task in the band of interest. Then, an orthogonal linear transformation is constructed by simultaneously diagonalizing the CPMs of two tasks. Finally, each band's logarithmic power of the transformed signals is extracted for the support vector machine (SVM) classifier. The experiment results on multiple datasets showed that the CCSP algorithm is fully applicable to multi-channel EEG for reliable multi-cognitive-task identification.

Thiamine metabolism dysfunction syndrome 2 (THMD2) is a rare metabolic disorder caused by SLC19A3 mutations, inherited in autosomal recessive pattern. As a treatable disease, early diagnosis and therapy with vitamin supplementation is important to improve the prognosis. So far, the reported cases were mainly from Saudi Arab regions, and presented with relatively simple clinical course because of the hot spot mutation (T422A). Rare Chinese cases were described until now. In this study, we investigated 18 Chinese THMD2 patients with variable phenotypes, and identified 23 novel SLC19A3 mutations, which expanded the genetic and clinical spectrum of the disorder. Meanwhile, we reviewed all 146 reported patients from different countries. Approximately 2/3 of patients presented with classical BTBGD, while 1/3 of patients manifested as much earlier onset and poor prognosis, including infantile Leigh-like syndrome, infantile spasms, neonatal lactic acidosis and infantile BTBGD. Literature review showed that elevated lactate in blood and CSF, as well as abnormal OXPHOS activities of muscle or skin usually correlated with infantile phenotypes, which indicated poor outcome. Brainstem involvement on MRI was more common in deceased cases. Thiamine supplementation is indispensable in the treatment of THMD2, whereas combination of biotin and thiamine is not superior to thiamine alone. But biotin supplementation does work in some patients. Genotypic-phenotypic correlation remains unclear which needs further investigation, and biallelic truncated mutations usually led to more severe phenotype.

Neural attention mechanism has achieved many successes in various tasks in natural language processing. However, existing neural attention models based on a densely connected network are loosely related to the attention mechanism found in psychology and neuroscience. Motivated by the finding in neuroscience that human possesses the template-searching attention mechanism, we propose to use convolution operation to simulate attentions and give a mathematical explanation of our neural attention model. We then introduce a new network architecture, which combines a recurrent neural network with our convolution-based attention model and further stacks an attention-based neural model to build a hierarchical sentiment classification model. The experimental results show that our proposed models can capture salient parts of the text to improve the performance of sentiment classification at both the sentence level and the document level.

There is a body of evidence that the aging immune system is linked to cancer. In this study, with aging- and immune-related DNA methylation data, we investigated the DNA methylation regulation changes in promoters with other regions of genes during aging and their association with the immune-cell proportion in the circulating whole blood of individuals. The analyses for aging- and CD4+ T cell proportion-derived differential genes showed that ubiquitination plays an important role in the aging immune system and tumorigenesis. Therefore, starting from a set of pre-annotated ubiquitination genes, we found that among the differentially ubiquitinated genes, DZIP3, an E3 ubiquitin ligase with no reports on its function in immune cells and tumorigenesis, was significantly associated with both aging (P-value = 3.86e-06) and CD4+ T cell proportion (P-value = 1.97e-05) in circulating blood. By collecting a cohort of 100 colon cancer patients and 50 healthy individuals, we validated that the 1st exon DNA methylation of DZIP3 could predict the onset of early stage (AUC = 0.833, OR = 8.82) and all pTNM stages of colorectal cancer (AUC = 0.782, OR = 5.70). Thus, the epigenetically regulated ubiquitination machine plays an important role in immune aging and tumorigenesis.

Aiming at classifying the polarities over aspects, aspect-based sentiment analysis (ABSA) is a fine-grained task of sentiment analysis. The vector representations of current models are generally constrained to real values. Based on mathematical formulations of quantum theory, quantum language models have drawn increasing attention. Words in such models can be projected as physical particles in quantum systems, and naturally represented by representation-rich complex-valued vectors in a Hilbert Space, rather than real-valued ones. In this paper, the Hilbert Space representation for ABSA models is investigated and the complexification of three strong real-valued baselines are constructed. Experimental results demonstrate the effectiveness of complexification and the outperformance of our complex-valued models, illustrating that the complex-valued embedding can carry additional information beyond the real embedding. Especially, a complex-valued RoBERTa model outperforms or approaches the previous state-of-the-art on three standard benchmarking datasets.

Abstract Background Protein secondary structure is the three dimensional form of local segments of proteins and its prediction is an important problem in protein tertiary structure prediction. Developing computational approaches for protein secondary structure prediction is becoming increasingly urgent. Results We present a novel deep learning based model, referred to as CNNH_PSS, by using multi-scale CNN with highway. In CNNH_PSS, any two neighbor convolutional layers have a highway to deliver information from current layer to the output of the next one to keep local contexts. As lower layers extract local context while higher layers extract long-range interdependencies, the highways between neighbor layers allow CNNH_PSS to have ability to extract both local contexts and long-range interdependencies. We evaluate CNNH_PSS on two commonly used datasets: CB6133 and CB513. CNNH_PSS outperforms the multi-scale CNN without highway by at least 0.010 Q8 accuracy and also performs better than CNF, DeepCNF and SSpro8, which cannot extract long-range interdependencies, by at least 0.020 Q8 accuracy, demonstrating that both local contexts and long-range interdependencies are indeed useful for prediction. Furthermore, CNNH_PSS also performs better than GSM and DCRNN which need extra complex model to extract long-range interdependencies. It demonstrates that CNNH_PSS not only cost less computer resource, but also achieves better predicting performance. Conclusion CNNH_PSS have ability to extracts both local contexts and long-range interdependencies by combing multi-scale CNN and highway network. The evaluations on common datasets and comparisons with state-of-the-art methods indicate that CNNH_PSS is an useful and efficient tool for protein secondary structure prediction.

Background/Objectives: Incidence of herpes zoster (HZ) increases with age. The vaccine to prevent HZ is available for those who are 50 and older. This study aims to investigate the cost-effectiveness of HZ vaccine on population older than 50 years old from a societal perspective in Singapore. Methods: The established decision analytic (Markov) model was utilized to simulate the results of vaccination for adults who were older than 50 years of age, compared to those without vaccination. The total number of HZ cases, postherpetic neuralgia (PHN) cases, total direct cost of treatment and indirect cost due to productivity loss were estimated based on the characteristics of population, and treatment costs were collected from a local institute in Singapore. The public health consensus for the ICER benchmark in Singapore was S$65,000. Results: One million individuals were simulated in this health economics analysis. Vaccination would reduce 48% of HZ cases and 33% of PHN cases over their life time. When a vaccination strategy for adults who are 50 years old and older was compared to a strategy with no vaccination, incremental cost-effectiveness ratio (ICER) was S$49,023 per quality adjusted life years (QALY) gained from the societal perspective. Conclusion: The herpes zoster vaccination is likely to be cost-effective in societal perspective for adult over 50 years of age in Singapore.

Abstract Background Prediction of DNA-binding residue is important for understanding the protein-DNA recognition mechanism. Many computational methods have been proposed for the prediction, but most of them do not consider the relationships of evolutionary information between residues. Results In this paper, we first propose a novel residue encoding method, referred to as the Position Specific Score Matrix (PSSM) Relation Transformation (PSSM-RT), to encode residues by utilizing the relationships of evolutionary information between residues. PDNA-62 and PDNA-224 are used to evaluate PSSM-RT and two existing PSSM encoding methods by five-fold cross-validation. Performance evaluations indicate that PSSM-RT is more effective than previous methods. This validates the point that the relationship of evolutionary information between residues is indeed useful in DNA-binding residue prediction. An ensemble learning classifier (EL_PSSM-RT) is also proposed by combining ensemble learning model and PSSM-RT to better handle the imbalance between binding and non-binding residues in datasets. EL_PSSM-RT is evaluated by five-fold cross-validation using PDNA-62 and PDNA-224 as well as two independent datasets TS-72 and TS-61. Performance comparisons with existing predictors on the four datasets demonstrate that EL_PSSM-RT is the best-performing method among all the predicting methods with improvement between 0.02–0.07 for MCC, 4.18–21.47% for ST and 0.013–0.131 for AUC. Furthermore, we analyze the importance of the pair-relationships extracted by PSSM-RT and the results validates the usefulness of PSSM-RT for encoding DNA-binding residues. Conclusions We propose a novel prediction method for the prediction of DNA-binding residue with the inclusion of relationship of evolutionary information and ensemble learning. Performance evaluation shows that the relationship of evolutionary information between residues is indeed useful in DNA-binding residue prediction and ensemble learning can be used to address the data imbalance issue between binding and non-binding residues. A web service of EL_PSSM-RT ( http://hlt.hitsz.edu.cn:8080/PSSM-RT_SVM/ ) is provided for free access to the biological research community.

Numerous studies have shown that changes in the epigenome are an important cause of human biochemical or metabolic parameter changes. Biochemical/metabolic parameter disorders of the human body are usually closely related to the occurrence of disease. Therefore, constructing credible DNA methylation site-biochemical/metabolic parameter associations are key in interpreting the pathogenesis of diseases. However, there is a lack of research on systematic integration analysis of DNA methylation with biochemical/metabolic parameter and diseases. In this study, we attempted to use the four-people, multiple time point detected data from the long-term isolation experiment to conduct a correlation analysis. We used the biclustering algorithm FABIA to cluster the DNA methylation site-parameter correlation matrixes into 28 biclusters. The results of the biological function analysis for these biclusters were consistent with the biochemical/metabolic parameter change characteristics of the human body during long-term isolation, demonstrating the reliability of the biclusters identified by our method. In addition, from these biclusters, we obtained highly credible biochemical/metabolic parameter-disease associations, which is supported by several studies. Our results indicate that there is an overlap of biochemical/metabolic parameter-disease associations derived from a small sample, multiple time point data in healthy populations and the associations obtained from a large sample data in patients during disease development. These findings provide insights into understanding the role of the epigenome in biochemical/metabolic parameter change and disease development and has potential applications in biology and medicine research.

Abstract Background Studies investigating the relationship between maternal tea drinking and risk of preterm birth have reached inconsistent results. Methods The present study analyzed data from a birth cohort study including 10,179 women who delivered a singleton live birth were conducted in Lanzhou, China between 2010 and 2012. Results Drinking tea (OR = 1.36, 95 % CI: 1.09–1.69), and specifically green (OR = 1.42, 95 % CI: 1.08–1.85) or scented tea (OR = 1.61, 95 % CI: 1.04–2.50), was associated with an increased risk of preterm birth. Drinking tea was associated with both moderate preterm (OR = 1.41, 95 % CI: 1.12–1.79) and spontaneous preterm birth (OR = 1.41, 95 % CI: 1.09–1.83). Risk of preterm birth increased with decreasing age of starting tea drinking (

Quantum-inspired language models have been introduced to Information Retrieval due to their transparency and interpretability. While exciting progresses have been made, current studies mainly investigate the relationship between density matrices of difference sentence subspaces of a semantic Hilbert space. The Hilbert space as a whole which has a unique density matrix is lack of exploration. In this paper, we propose a novel Quantum Expectation Value based Language Model (QEV-LM). A unique shared density matrix is constructed for the Semantic Hilbert Space. Words and sentences are viewed as different observables in this quantum model. Under this background, a matching score describing the similarity between a question-answer pair is naturally explained as the quantum expectation value of a joint question-answer observable. In addition to the theoretical soundness, experiment results on the TREC-QA and WIKIQA datasets demonstrate the computational efficiency of our proposed model with excellent performance and low time consumption.

In recent years, there has been an increasing interest in learning a distributed representation of word sense. Traditional context clustering based models usually require careful tuning of model parameters, and typically perform worse on infrequent word senses. This paper presents a novel approach which addresses these limitations by first initializing the word sense embeddings through learning sentence-level embeddings from WordNet glosses using a convolutional neural networks. The initialized word sense embeddings are used by a context clustering based model to generate the distributed representations of word senses. Our learned representations outperform the publicly available embeddings on half of the metrics in the word similarity task, 6 out of 13 sub tasks in the analogical reasoning task, and gives the best overall accuracy in the word sense effect classification task, which shows the effectiveness of our proposed distributed distribution learning model.

Transcription factor binding sites (TFBSs) play an important role in gene expression regulation. Many computational methods for TFBS prediction need sufficient labeled data. However, many transcription factors (TFs) lack labeled data in cell types. We propose a novel method, referred to as DANN_TF, for TFBS prediction. DANN_TF consists of a feature extractor, a label predictor, and a domain classifier. The feature extractor and the domain classifier constitute an Adversarial Network, which ensures that learned features are common features across different cell types. DANN_TF is evaluated on five TFs in five cell types with a total of 25 cell-type TF pairs and compared to a baseline method which does not use Adversarial Network. For both data augmentation and cross-cell-type prediction, DANN_TF performs better than the baseline method on most cell-type TF pairs. DANN_TF is further evaluated by an additional 13 TFs in the five cell types with a total of 65 cell-type TF pairs. Results show that DANN_TF achieves significantly higher AUC than the baseline method on 96.9% pairs of the 65 cell-type TF pairs. This is a strong indication that DANN_TF can indeed learn common features for cross-cell-type TFBS prediction.