With technological advancements in transmitting and storing large video files, more and more organizations are incorporating asynchronous video interviews as part of their personnel selection process. Automatic evaluation of these videos is a challenging machine learning setting because the samples are composed of time series input data but only one overall label is available. It is unclear which segments of the time series input (i.e., videos) are the most important ones for prediction. Not all nonverbal features, spoken words, and utterances contribute equally to the prediction; some segments of the videos might even introduce noise to the model. Processing all multimodal information is therefore inefficient. To address this challenge, we propose a framework to model the content of the answer via the full transcription and the speaking patterns of the interviewee via short clips. Our model learns to automatically select the most informative segment by previewing the acoustic modality using a technique called differentiable masking. The results show that our method outperforms existing approaches while being more efficient since only partial multimodal data are processed, and the interpretability of the model is enhanced.

Personality traits are known to have a high correlation with job performance. On the other hand, there is a strong relationship between language and personality. In this paper, we presented a neural network model for inferring personality and hirability. Our model was trained only from linguistic features but achieved good results by incorporating transfer learning and multi-task learning techniques. The model improved the F1 score 5.6

Generative language models such as ChatGPT are good at producing answers for questions whose answers are publicly available. For questions about private organization documents, such models cannot perform well due to the fact that they were not trained on these documents, hence the knowledge was not embedded in their parameters. On the other hand, traditional Question Answering (QA) systems with a Retriever and a Reader model are interpretable and can be trained quickly on private documents, but they require laborious annotation. The output of QA systems is a span of text, which is not friendly to the end users as well. In this paper, we proposed a framework for generating closed-domain QA data set in a semi-automatic manner, reducing human efforts. An organization-specific QA data set was created based on this framework. Additionally, we fine-tuned a traditional open-domain QA model on the newly created data set, and combine the output of this model with recently released language models to improve the naturalness of the response. Finally, we present our results and discuss the findings.

One of the challenges in constructing Knowledge Graphs from text is verifying the correctness of the produced results. Each language has its unique characteristics, so a Knowledge Graphs construction system may perform better on certain languages and worse on others. In order to detect the most suitable Knowledge Graph construction systems for Vietnamese, in this paper, we propose a method to classify triples extracted from such systems into two categories: Existent and Non-existent. Vietnamese is a low-resource language with limited natural language processing tools and datasets. By combining BERT with a self-constructed Vietnamese Knowledge Graph, we build a classification model to verify the existence of triples in paragraphs. Our results suggest that BERT can learn contextual relations between words from a large amount of text, even for a low-resource language like Vietnamese. BERT’s adaptive capability to detect meaningful triples is also shown and discussed. The outcome of this paper could potentially be used to build more sophisticated systems to solve Knowledge Graph construction and Triple Classification tasks in low resource languages.

Most data in the world is in form of text. Therefore, we can say text stores large amount of the knowledge of human beings. Extracting useful knowledge from text, however, is not a simple task. In this paper, we present a complete pipeline to extract knowledge from paragraph. This pipeline combines state-of-the-art systems in order to yield optimal results. There are some other Knowledge Graphs such as Google Knowledge Graph, YAGO, or DBpedia. Most of the data in these Knowledge Graphs is in English. On the other hand, the results from our system is used to build a new Knowledge Graph in Vietnamese of Vietnam Tourism. We use the rich resources language like English to process a low resources language like Vietnamese. We utilize the NLP tools of English such as Google translate, Stanford parser, Co-referencing, ClausIE, MinIE. We develop Google Search to find the text describing the entities in the Internet. This text is in Vietnamese. Then, we translate the Vietnamese text into English text and use English NLP tools to extract triples. Finally, we translate the triples back into Vietnamese and build the knowledge graph of Vietnam tourism. We conduct experiment and discover the advantages and disadvantages of our method.

The simplest and effective way to store human knowledge through centuries was using text. Along with the advancement of technology nowadays, the volume of text has grown to be larger and larger. To extract useful information from this amount of text becomes an exceptionally complex task. As an effort to solve that problem, in this paper, we present a pipeline to extract core knowledge from large quantity text using distributed computing. The components of our pipeline are systems that were known to yield good results. The outputs of our proposed system are stored in a knowledge graph. A knowledge graph is a graph for storing knowledge in the form of triples (head, relation, tail). Some of the existing knowledge graphs in the world are Google knowledge graph, YAGO, DBLP, or DBpedia. These knowledge graphs have one thing in common-they are in English. The English language is studied by many researchers in the world and it had become a rich-resource language (with many natural language processing tools and data set). Vietnamese, on the other hand, is a low-resource language. Therefore, we use cross-lingual transfer method to build a Vietnamese knowledge graph. Firstly, we collect data in form of text about Vietnam tourism, which was written mostly in Vietnamese, using Google search and Wikipedia. In the next step, we translate them into English with Google Translate and use English Natural Language Processing tools like Stanford Parser, Co-referencing, ClausIE, MinIE to extract useful triples from this text. Lastly, the triples are translated back to Vietnamese to build a Vietnam tourism knowledge graph. Since we are working with massive text, we develop a distributed algorithm to extract triples from sentences of massive text. This is a distributed version of MinIE, which was originally developed for a single machine model. In Apache Spark framework, we divide massive text into many smaller parts and move them to the worker nodes with distributed MinIE function. Spark distributed MinIE will extract the triples of sentences in the local text of this worker node in parallel. Finally, the result of worker nodes will be sent back to the master node for building the knowledge graph. We conduct experiments with the distributed MinIE on spark cluster to prove the outperformance of our proposed algorithm.