Paper Title: "Generation of Highlights from Research Papers Using Pointer-Generator Networks and SciBERT Embeddings" authored by Tohida Rehman, Debarshi Kumar Sanyal, Samiran Chattopadhyay, Plaban Kumar Bhowmick, Partha Pratim Das. Author Written Research Highlights:

1. A new multi-disciplinary dataset called MixSub is also presented. It contains URIs of papers from different domains. Each paper contains research highlights in addition to the abstract. The performance of the proposed method on MixSub is also described.
2. Conducted crawling and publishing activities to compile the CSPubSum dataset, comprising 10,147 computer science publications sourced from ScienceDirect.
3. An approach for extracting research highlights from scientific papers is proposed.
4. The proposed approach combines SciBERT pre-trained word embeddings with a pointer-generator network that also uses a coverage mechanism.
5. The proposed approach outperforms existing methods on CSPubSum dataset in terms of ROUGH, METEOR , and BERTScore metrics.

MixSub Dataset:

1. We propose a new dataset called MixSub that contains research articles from multiple domains.
2. To prepare MixSub, we crawled the ScienceDirect website and curated articles published in various journals in year 2020.
3. We removed the articles that did not contain research highlights.
4. Finally, we got 19785 articles with author-written research highlights as shown in Table given below.
5. Each example in this dataset is organized as (abstract, author-written research highlights).
6. We have also segmented the dataset into training, validation and test subsets in the ratio of 80:10:10.

Subject-wise URL/data contents count in MixSub dataset

| ----------- | ---------------------------------------------- | ------ | ------ | ---- | ----- |

Domain Name	Subject Name	#Total	#Train	#Val	#Test
Biological	Agricultural and Biological Sciences	2156	1726	216	214
	Biochemistry, Genetics and Molecular Biology	976	806	71	99
	Immunology and Microbiology	233	195	24	14
	Neuroscience	962	771	96	95
-----------	----------------------------------------------	------	------	----	-----
Chemistry	Chemical Engineering	2140	1713	214	213
	Chemistry	2282	1919	240	123
	Materials Science	735	572	82	81
-----------	----------------------------------------------	------	------	----	-----
Energy	Energy	1313	1025	145	143
	Environmental Science	677	517	81	79
-----------	----------------------------------------------	------	------	----	-----
Management	Business, Management and Accounting	698	560	70	68
	Decision Sciences	947	759	95	93
	Economics, Econometrics and Finance	421	324	56	41
-----------	----------------------------------------------	------	------	----	-----
Nursing	Health Sciences	823	796	12	15
	Nursing and Health Professions	61	47	8	6
	Pharmacology, Toxicology, Pharmaceutical Science	1184	949	118	117
	Psychology	28	24	3	1
	Veterinary Science and Veterinary Medicine	186	156	19	11
-----------	----------------------------------------------	------	------	----	-----
Physics	Earth and Planetary Sciences	1354	1038	159	157
	Mathematics	288	232	29	27
	Physics and Astronomy	1469	1177	147	145
-----------	----------------------------------------------	------	------	----	-----
Social Science	Social Sciences	852	654	100	98
-----------	----------------------------------------------	------	------	----	-----

CSPubSum Dataset:

1. Conducted crawling and publishing activities to compile the CSPubSum dataset, comprising 10,147 computer science publications sourced from ScienceDirect.
2. The URLs for the dataset were extracted based on the work by Collins et al. \cite{collins2017supervised}.
3. Each document in the dataset is meticulously structured, featuring essential components such as abstracts and author-provided research highlights, ensuring comprehensive coverage and accessibility for researchers and enthusiasts.
4. We have also segmented the dataset into training, validation and test subsets in the ratio of 80:10:10.

Our paper is available on IEEE Xplore and arXiv.

If you find our dataset/method useful, please cite our paper:

1. IEEE Access: Rehman, T., Sanyal, D. K., Chattopadhyay, S., Bhowmick, P. K., & Das, P. P. (2023). "Generation of Highlights From Research Papers Using Pointer-Generator Networks and SciBERT Embeddings." IEEE Access, 11, 91358-91374.
2. arXiv: Rehman, T. et al. (2023). "Generation of Highlights from Research Papers Using Pointer-Generator Networks and SciBERT Embeddings." arXiv preprint arXiv:2302.07729."

All rights reserved to the author.