In this paper, a novel method is proposed to extract the yarn positional information for the structure analysis of a textile fabric from its three-dimensional (3D) image obtained from its X-ray computed tomography (CT) images. In this paper, the sequence of the points on the yarn centerline is defined as the yarn positional information. The sequence is extracted by tracing the yarn. The yarn is traced using the yarn direction obtained by estimating the directions of its filaments and averaging the estimated filament directions. The filament direction is estimated by correlating the filament part in the 3D CT image with a 3D filament model. The trajectory of the yarn tracing corresponds to the yarn positional information. The validity of the proposed method is discussed by experimentally applying the proposed method to a 3D CT image of a plain knitted fabric. Furthermore, the usefulness is discussed through an experiment, in which the proposed method is applied to a 3D CT image of a plain woven fabric, which is made of a two-folded yarn.

Sequence-to-sequence models are a common approach to develop a chatbot. They can train a conversational model in an end-to-end manner. One significant drawback of such a neural network based approach is that the response generation process is a black-box, and how a specific response is generated is unclear. To tackle this problem, an interpretable response generation mechanism is desired. As a step toward this direction, we focus on dialogue-acts (DAs) that may provide insight to understand the response generation process. In particular, we propose a method to predict a DA of the next response based on the history of previous utterances and their DAs. Experiments using a Switch Board Dialogue Act corpus show that compared to the baseline considering only a single utterance, our model achieves 10.8% higher F1-score and 3.0% higher accuracy on DA prediction.

Advanced pre-trained models for text representation have achieved state-of-the-art performance on various text classification tasks. However, the discrepancy between the semantic similarity of texts and labelling standards affects classifiers, i.e. leading to lower performance in cases where classifiers should assign different labels to semantically similar texts. To address this problem, we propose a simple multitask learning model that uses negative supervision. Specifically, our model encourages texts with different labels to have distinct representations. Comprehensive experiments show that our model outperforms the stateof-the-art pre-trained model on both singleand multi-label classifications, sentence and document classifications, and classifications in three different languages.

We propose a novel automatic weave diagram construction method from yarn positional data of woven fabric. In this work, a set of yarn positional data implies a sequence of center points of a yarn obtained using the yarn tracing method proposed in our previous study. For constructing the weave diagram, the sets of yarn positional data are first divided into warp and weft based on the gradient of a line, which approximates the yarn positional data. The intersections between warp and weft are then calculated using the approximated lines. Then the weave diagram is constructed by comparing the warp positions with the weft positions at all the intersections. The effectiveness of the automatic weave diagram construction method is confirmed by experimentally applying this method to the yarn positional data of a double-layered woven fabric.

A sequence-to-sequence model tends to generate generic responses with little information for input utterances. To solve this problem, we propose a neural model that generates relevant and informative responses. Our model has simple architecture to enable easy application to existing neural dialogue models. Specifically, using positive pointwise mutual information, it first identifies keywords that frequently co-occur in responses given an utterance. Then, the model encourages the decoder to use the keywords for response generation. Experiment results demonstrate that our model successfully diversifies responses relative to previous models.