医药知识图谱自动问答系统实现，包括构建知识图谱、基于知识图谱的流水线问答以及前端实现。实体识别（基于词典+BERT_CRF）、实体链接（Sentence-BERT做匹配）、意图识别（基于提问词+领域词词典）。

[NeurIPS 2022 Spotlight] RLIP: Relational Language-Image Pre-training and a series of other methods to solve HOI detection and Scene Graph Generation.

Implementation of BERT-Based Span Entity and Relation Prediction Models for Question Answering over Wikidata

Code for rendering images for NeurRIPS 2020 paper "Rel3D: A Minimally Contrastive Benchmark for Grounding Spatial Relations in 3D"

Master Thesis: Extraction of Gateways in Process Model Generation from Text

We present an open-source end-to-end framework for disease-specific knowledge discovery from raw text. This repository includes two annotated datasets focused on Rett syndrome and Alzheimer's disease, designed to detect semantic relations between biomedical entities. Extensive benchmarking explores various relation and entity representations.