With the current spike in the popularity of AI models, we are witnessing an era of redefining the way we seek information. At Erdyn, we advocate for more efficient knowledge consumption to foster innovation and continue imagining and building the world of tomorrow. To this end, Question Answering models can retrieve an answer to a question from a given text, but it is commonly known in the AI community that real-world implementations raise a challenging set of practical issues. In our quest for augmenting our consultants and assisting them in their investigations, we studied the hurdles standing in the way of the successful adoption of information-seeking QA in our environment, and published a paper tackling the portability of such systems on scientific papers to real-life application scenarios, where we provide a characterization of practical aspects and discuss the limitations of current state-of-the-art development in Natural Language Processing. You can check it out here: https://aclanthology.org/2022.wiesp-1.8/
The paper was presented at the First Workshop on Information Extraction from Scientific Publications, held at AACL-IJCNLP 2022.