Text API service is based on Korean pre-trained language model. The Korean pre-trained language model studied written Korean texts from news articles and Korean Wikipedia to serve various business use cases. In addition, it uses a patented model compression architecture, differentiating itself from other Korean language models in the market. Therefore, it can be used for various business situations such as VOC systems and chatbots.
Text API uses the abstractive summarization approach for text summaries. While the extractive summarization approach summarizes key sentences, abstractive summarization uses Natural Language Generation (NLG) technique to capture the entire context and create a new sentence. With text API, analysts generating business insights through texts can benefit greatly from time and cost savings.
Semantic Textual Similarity(STS) helps evaluate two texts in terms of meaning or context. For this purpose, Siamese Neural Network(SNN) and Convolutional Neural Network(CNN) are used, reducing processing time. Capturing the contextual similarity can later be utilized for search features.
- Summary API, Semantic Textual Similarity API
- Text API functional/performance testing through demo
- Sentence summary : Summarize sentences (200 to 1,000 characters)
- Semantic textual similarity : Compare sentences (20 to 1,000 characters)
- View call status by API and by day/week/month
- View the number of calls by API (All/success/error)
- View response time for successful API