RESUMO
The comprehension of spoken language is a crucial aspect of dialogue systems, encompassing two fundamental tasks: intent classification and slot filling. Currently, the joint modeling approach for these two tasks has emerged as the dominant method in spoken language understanding modeling. However, the existing joint models have limitations in terms of their relevancy and utilization of contextual semantic features between the multiple tasks. To address these limitations, a joint model based on BERT and semantic fusion (JMBSF) is proposed. The model employs pre-trained BERT to extract semantic features and utilizes semantic fusion to associate and integrate this information. The results of experiments on two benchmark datasets, ATIS and Snips, in spoken language comprehension demonstrate that the proposed JMBSF model attains 98.80% and 99.71% intent classification accuracy, 98.25% and 97.24% slot-filling F1-score, and 93.40% and 93.57% sentence accuracy, respectively. These results reveal a significant improvement compared to other joint models. Furthermore, comprehensive ablation studies affirm the effectiveness of each component in the design of JMBSF.
Assuntos
Idioma , Semântica , Processamento de Linguagem Natural , Intenção , Estimulação AcústicaRESUMO
Robustness against background noise and reverberation is essential for many real-world speech-based applications. One way to achieve this robustness is to employ a speech enhancement front-end that, independently of the back-end, removes the environmental perturbations from the target speech signal. However, although the enhancement front-end typically increases the speech quality from an intelligibility perspective, it tends to introduce distortions which deteriorate the performance of subsequent processing modules. In this paper, we investigate strategies for jointly training neural models for both speech enhancement and the back-end, which optimize a combined loss function. In this way, the enhancement front-end is guided by the back-end to provide more effective enhancement. Differently from typical state-of-the-art approaches employing on spectral features or neural embeddings, we operate in the time domain, processing raw waveforms in both components. As application scenario we consider intent classification in noisy environments. In particular, the front-end speech enhancement module is based on Wave-U-Net while the intent classifier is implemented as a temporal convolutional network. Exhaustive experiments are reported on versions of the Fluent Speech Commands corpus contaminated with noises from the Microsoft Scalable Noisy Speech Dataset, shedding light and providing insight about the most promising training approaches.
Assuntos
Percepção da Fala , Fala , RuídoRESUMO
BACKGROUND: To provide satisfying answers, medical QA system has to understand the intentions of the users' questions precisely. For medical intent classification, it requires high-quality datasets to train a deep-learning approach in a supervised way. Currently, there is no public dataset for Chinese medical intent classification, and the datasets of other fields are not applicable to the medical QA system. To solve this problem, we construct a Chinese medical intent dataset (CMID) using the questions from medical QA websites. On this basis, we compare four intent classification models on CMID using a case study. METHODS: The questions in CMID are obtained from several medical QA websites. The intent annotation standard is developed by the medical experts, which includes four types and 36 subtypes of users' intents. Besides the intent label, CMID also provides two types of additional information, including word segmentation and named entity. We use the crowdsourcing way to annotate the intent information for each Chinese medical question. Word segmentation and named entities are obtained using the Jieba and a well-trained Lattice-LSTM model. We loaded a Chinese medical dictionary consisting of 530,000 for word segmentation to obtain a more accurate result. We also select four popular deep learning-based models and compare their performances of intent classification on CMID. RESULTS: The final CMID contains 12,000 Chinese medical questions and is organized in JSON format. Each question is labeled the intention, word segmentation, and named entity information. The information about question length, number of entities, and are also detailed analyzed. Among Fast Text, TextCNN, TextRNN, and TextGCN, Fast Text and TextCNN models have achieved the best results in four types and 36 subtypes intent classification, respectively. CONCLUSIONS: In this work, we provide a dataset for Chinese medical intent classification, which can be used in medical QA and related fields. We performed an intent classification task on the CMID. In addition, we also did some analysis on the content of the dataset.
Assuntos
Benchmarking , Intenção , China , HumanosRESUMO
"Treatise on Febrile Diseases" is an important classic book in the academic history of Chinese material medica. Based on the knowledge map of traditional Chinese medicine established by the study of "Treatise on Febrile Diseases", a question-answering system of traditional Chinese medicine was established to help people better understand and use traditional Chinese medicine. Intention classification is the basis of the question-answering system of traditional Chinese medicine, but as far as we know, there is no research on question intention classification based on "Treatise on Febrile Diseases". In this paper, the intent classification research is carried out based on the Chinese material medica-related content materials in "Treatise on Febrile Diseases" as data. Most of the existing models perform well on long text classification tasks, with high costs and a lot of memory requirements. However, the intent classification data of this paper has the characteristics of short text, a small amount of data, and unbalanced categories. In response to these problems, this paper proposes a knowledge distillation-based bidirectional Transformer encoder combined with a convolutional neural network model (TinyBERT-CNN), which is used for the task of question intent classification in "Treatise on Febrile Diseases". The model used TinyBERT as an embedding and encoding layer to obtain the global vector information of the text and then completed the intent classification by feeding the encoded feature information into the CNN. The experimental results indicated that the model outperformed other models in terms of accuracy, recall, and F1 values of 96.4%, 95.9%, and 96.2%, respectively. The experimental results prove that the model proposed in this paper can effectively classify the intent of the question sentences in "Treatise on Febrile Diseases", and provide technical support for the question-answering system of "Treatise on Febrile Diseases" later.
Assuntos
Intenção , Redes Neurais de Computação , Humanos , IdiomaRESUMO
Collaborative state recognition is a critical issue for physical human-robot collaboration (PHRC). This paper proposes a contact dynamics-based state recognition method to identify the human-robot collaborative grinding state. The main idea of the proposed approach is to distinguish between the human-robot contact and the robot-environment contact. To achieve this, dynamic models of both these contacts are first established to identify the difference in dynamics between the human-robot contact and the robot-environment contact. Considering the reaction speed required for human-robot collaborative state recognition, feature selections based on Spearman's correlation and random forest recursive feature elimination are conducted to reduce data redundancy and computational burden. Long short-term memory (LSTM) is then used to construct a collaborative state classifier. Experimental results illustrate that the proposed method can achieve a recognition accuracy of 97% in a period of 5 ms and 99% in a period of 40 ms.
RESUMO
BACKGROUND AND OBJECTIVE: This study describes the integration of a spoken dialogue system and nursing records on an Android smartphone application intending to help nurses reduce documentation time and improve the overall experience of a healthcare setting. The application also incorporates with collecting personal sensor data and activity labels for activity recognition. METHODS: We developed a joint model based on a bidirectional long-short term memory and conditional random fields (Bi-LSTM-CRF) to identify user intention and extract record details from user utterances. Then, we transformed unstructured data into record inputs on the smartphone application. RESULTS: The joint model achieved the highest F1-score at 96.79%. Moreover, we conducted an experiment to demonstrate the proposed model's capability and feasibility in recording in realistic settings. Our preliminary evaluation results indicate that when using the dialogue-based, we could increase the percentage of documentation speed to 58.13% compared to the traditional keyboard-based. CONCLUSIONS: Based on our findings, we highlight critical and promising future research directions regarding the design of the efficient spoken dialogue system and nursing records.
Assuntos
Registros de Enfermagem , Smartphone , Coleta de Dados , Registros Eletrônicos de Saúde , HumanosRESUMO
Upper limb and hand functionality is critical to many activities of daily living and the amputation of one can lead to significant functionality loss for individuals. From this perspective, advanced prosthetic hands of the future are anticipated to benefit from improved shared control between a robotic hand and its human user, but more importantly from the improved capability to infer human intent from multimodal sensor data to provide the robotic hand perception abilities regarding the operational context. Such multimodal sensor data may include various environment sensors including vision, as well as human physiology and behavior sensors including electromyography and inertial measurement units. A fusion methodology for environmental state and human intent estimation can combine these sources of evidence in order to help prosthetic hand motion planning and control. In this paper, we present a dataset of this type that was gathered with the anticipation of cameras being built into prosthetic hands, and computer vision methods will need to assess this hand-view visual evidence in order to estimate human intent. Specifically, paired images from human eye-view and hand-view of various objects placed at different orientations have been captured at the initial state of grasping trials, followed by paired video, EMG and IMU from the arm of the human during a grasp, lift, put-down, and retract style trial structure. For each trial, based on eye-view images of the scene showing the hand and object on a table, multiple humans were asked to sort in decreasing order of preference, five grasp types appropriate for the object in its given configuration relative to the hand. The potential utility of paired eye-view and hand-view images was illustrated by training a convolutional neural network to process hand-view images in order to predict eye-view labels assigned by humans.