A Dataset of Reader-Aware Multi-Document Summarization (RA-MDS): http://www1.se.cuhk.edu.hk/~textmine/dataset/ra-mds/
We are delighted to announce the release of a dataset for Reader-Aware Multi- Document Summarization. In the RA-MDS task, one should jointly consider news documents and news reader comments for generating the summaries. This dataset contains 45 topics from those 6 predefined categories. Each topic contains 10 news documents and 4 model summaries. Definitely, this dataset could also be used for conventional MDS. Moreover, the human-written summaries are abstractive, thus, it is also a good dataset for text generation. This is a *HIGH-QUALITY* dataset. We put significant effort in preparing the dataset in the past three years, and spent tens of thousands US dollars. During the preparation, we employed experts with particular background, such as Journalist and English Literature, to conduct the tasks of data collection, aspect annotation, and summary writing, as well as scrutinizing. Thus, we are pretty confident that this dataset will bring in interesting new things to the researchers in our community. Please enjoy the dataset, and forward us questions if any. Best, -- Name: 李丕绩, Piji Li Email: *p...@se.cuhk.edu.hk <p...@se.cuhk.edu.hk>* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Dept. of Systems Engineering & Engineering Management, Rm 711, William M. W. Mong Engineering Building, The Chinese University of Hong Kong Shatin, NT, Hong Kong
_______________________________________________ uai mailing list uai@ENGR.ORST.EDU https://secure.engr.oregonstate.edu/mailman/listinfo/uai
