In multi-modal dialogue systems, it is important to allow the use of images as part of a multi-turn conversation. Training such dialogue systems generally requires a large-scale dataset consisting of multi-turn dialogues that involve images, but such datasets rarely exist. In response, this paper proposes a 45k multimodal dialogue dataset created with minimal human intervention. Our method to create such a dataset consists of (1) preparing and pre-processing text dialogue datasets, (2) creating image-mixed dialogues by using a textto-image replacement technique, and (3) employing a contextual-similarity-based filtering step to ensure the contextual coherence of the dataset. To evaluate the validity of our dataset, we devise a simple retrieval model for dialogue sentence prediction tasks. Automatic metrics and human evaluation results on such tasks show that our dataset can be effectively used as training data for multi-modal dialogue systems which require an understanding of images and text in a context-aware manner. Our dataset and generation code is available at https://github.com/shh1574/ multi-modal-dialogue-dataset.

Tag recommendation relies on either a ranking function for top-k tags or an autoregressive generation method. However, the previous methods neglect one of two seemingly conflicting yet desirable characteristics of a tag set: orderlessness and inter-dependency. While the ranking approach fails to address the interdependency among tags when they are ranked, the autoregressive approach fails to take orderlessness into account because it is designed to utilize sequential relations among tokens. We propose a sequence-oblivious generation method for tag recommendation, in which the next tag to be generated is independent of the order of the generated tags and the order of the ground truth tags occurring in training data. Empirical results on two different domains, Instagram and Stack Overflow, show that our method is significantly superior to the previous approaches.

Current study attempted to investigate the evidence utility of psychological autopsy based on precedents accumulated over the past 10 years. Precedents analysis was conducted by applying topic modeling, and the main topics of psychological autopsy were confirmed. For the study, the case law was selected based on the information of the Korean court and the National Legal Information Center. If the lower court and the appeal were the same case, the final 46 precedents were integrated into one document and used as analysis data, and the above data were classified by time and case. Reviewing the original and topic modeling results showed that ‘Interpersonal factors', including all the personal environment of office worker suicide, stressful work environment and socio-economic factors. ‘Inner suicide risk factor', which urged the suicide of the deceased, which is important to suicide prediction and can affect suicide behavior. ‘Purpose of psychological autopsy’ which can be obtained through psychological autopsy. ‘Organizational-related factors' including suicide-related factors in the military, indicating that the psychological state before suicide is often caused by job stress. ‘Source of information on psychological autopsy' which can obtain the information necessary to conduct psychological autopsy is an issue of psychological autopsy precedents found in precedents. These results can be used as evidence to determine policy development and precedent trends related to the evidence utility of psychological autopsy in the future.

No abstract

In multi-modal dialogue systems, it is important to allow the use of images as part of a multi-turn conversation. Training such dialogue systems generally requires a large-scale dataset consisting of multi-turn dialogues that involve images, but such datasets rarely exist. In response, this paper proposes a 45k multimodal dialogue dataset created with minimal human intervention. Our method to create such a dataset consists of (1) preparing and pre-processing text dialogue datasets, (2) creating image-mixed dialogues by using a textto-image replacement technique, and (3) employing a contextual-similarity-based filtering step to ensure the contextual coherence of the dataset. To evaluate the validity of our dataset, we devise a simple retrieval model for dialogue sentence prediction tasks. Automatic metrics and human evaluation results on such tasks show that our dataset can be effectively used as training data for multi-modal dialogue systems which require an understanding of images and text in a context-aware manner. Our dataset and generation code is available at https://github.com/shh1574/ multi-modal-dialogue-dataset.