Title: Approach for Unsupervised Bug Report Summarization
Work done on this paper was it used four unsupervised techniques (Cen-troid, MMR, Divrank, Grasshopper) and compared their efficacy with supervised approach(BRC,EC,EMC). The efficacy of the unsupervised techniques is enhanced by noise removal technique and deducting the useless code. There were few significant limitation observed that were- Observed results are particular results cant be used for other setups and they need reconfiguration depending on subject to which noise reduction is applied . 
2.2 Authors: Rafael Lotufo, Zeeshan Malik, Krzysztof Czarnecki
Title: Modelling the hurried bug reports reading process to summarize bug reports 
This paper focused on unsupervised techniques for bug summarization and apply noise reduction, two of the unsupervised technique became scalable for larger size bug reports. Its algorithm resulted into following limitations- it dont provide an evaluation showing how user would choose two input parameter, which are summary percentage length and minimum relevance threshold.
2.3 Authors: Sarah Rastkar, Gail C Murphy
Title: Automatic Summarization of Bug Reports
They made a summarizer that produces summaries that are statistically better than produced by existing conversation- based generators. In this paper, they have researched the automatic generation of one sort of software skill, bug answers, to give designers the advantages others encounter every day in different areas. They found that current conversation based extractive summary generators can create summaries for bug reports that are better than any random classifier. They likewise discovered that an extractive summary generator prepared on bug reports delivers the best outcomes. It resulted into following limitations that is Na?ve/non-experts cant create summaries and more than one annotator is required.
2.4 Authors: Elder Cirilo, Fernando Mourao
Title: Bug Report Summarization: An Evaluation of Ranking Technique
They provided a solution using extractive summaries where summaries are based on comments instead of one based on isolated sentences. In this paper, they propose a novel methodology where summaries depend on remarks, rather than the ones dependent on confined sentences, as proposed by past works. Exact outcomes prove with our arguments that positioning the most relevant comments would enable developers to discover more appropriate information. They could observe that summaries generated by conventional ranking algorithms are precise concerning developers expected data, when contrasted with reference summaries made manually, offers applicable summaries in general. Conclusion lead to limitation like Size and amount of reports may be a threat to the conclusion of their study and few bug reports contain different types of structured information which cant be treated in the algorithm.