H. Lee

Junwoo Park, Youngwoo Cho, Haneol Lee et al.

Question Answering (QA) is a widely-used framework for developing and evaluating an intelligent machine. In this light, QA on Electronic Health Records (EHR), namely EHR QA, can work as a crucial milestone towards developing an intelligent agent in healthcare. EHR data are typically stored in a relational database, which can also be converted to a directed acyclic graph, allowing two approaches for EHR QA: Table-based QA and Knowledge Graph-based QA. We hypothesize that the graph-based approach is more suitable for EHR QA as graphs can represent relations between entities and values more naturally compared to tables, which essentially require JOIN operations. In this paper, we propose a graph-based EHR QA where natural language queries are converted to SPARQL instead of SQL. To validate our hypothesis, we create four EHR QA datasets (graph-based VS table-based, and simplified database schema VS original database schema), based on a table-based dataset MIMICSQL. We test both a simple Seq2Seq model and a state-of-the-art EHR QA model on all datasets where the graph-based datasets facilitated up to 34% higher accuracy than the table-based dataset without any modification to the model architectures. Finally, all datasets are open-sourced to encourage further EHR QA research in both directions.

Kihong Kim, Haneol Lee, Jihye Park et al.

Generating high-quality videos that synthesize desired realistic content is a challenging task due to their intricate high-dimensionality and complexity of videos. Several recent diffusion-based methods have shown comparable performance by compressing videos to a lower-dimensional latent space, using traditional video autoencoder architecture. However, such method that employ standard frame-wise 2D and 3D convolution fail to fully exploit the spatio-temporal nature of videos. To address this issue, we propose a novel hybrid video diffusion model, called HVDM, which can capture spatio-temporal dependencies more effectively. The HVDM is trained by a hybrid video autoencoder which extracts a disentangled representation of the video including: (i) a global context information captured by a 2D projected latent (ii) a local volume information captured by 3D convolutions with wavelet decomposition (iii) a frequency information for improving the video reconstruction. Based on this disentangled representation, our hybrid autoencoder provide a more comprehensive video latent enriching the generated videos with fine structures and details. Experiments on video generation benchamarks (UCF101, SkyTimelapse, and TaiChi) demonstrate that the proposed approach achieves state-of-the-art video generation quality, showing a wide range of video applications (e.g., long video generation, image-to-video, and video dynamics control).