The algorithm that will change the ways of information consumption

Meet Uri Ernst, a PhD student in the Faculty of Engineering and the Department of Computer Science, who researches automatic summarization of texts

Machine learning - review of texts. Image: depositphotos.com
Machine learning - review of texts. Image: depositphotos.com

Meet Uri Ernst, a PhD student in the Department of Computer Science, researching automatic summarization of texts. Uri Ernst's research deals with summarizing a collection of documents on a certain topic, with the help of machine learning algorithms and artificial intelligence. In each such collection of documents, Ernst tries to simulate the process of human summarization, which naturally breaks down the problem into small problems such as - finding important pieces of information, examining repetitive pieces of information, and unifying the pieces of information into a readable and coherent exhaustive summary, up to 100 words.

As a doctoral student in the natural language processing laboratory, in the department ofComputer Science at Bar-Ilan University, under the guidance of Prof. Ido Dagan and Prof. Yaakov Goldberger from the Faculty of Engineering, Ernst and his colleagues build a computational model for each of the tasks. Combining all the models together, produces higher quality summaries than those obtained in previous models. The recently published study won an award Best paper runner up award.

Ernst says that the results of his research will make our lives much easier in the future. "The information that exists today on the net is growing at an enormous speed, and automatic summary methods will enter our lives in the coming years, and will change the way we consume information." For example, at the end of each work meeting, the computer will send all participants an automatic summary of the meeting by e-mail, including the distribution of tasks to continue. Locating information in the search engine will not only yield links where there is a high probability of an answer, but will also read the content of the sites and summarize an answer that includes information from several sources. The same is true for summarizing long legal documents, summarizing complex medical documents, summarizing intelligence transcripts and identifying the important intelligence information from huge databases.

Ernst is a graduate of the Faculty of Engineering in Bar-Ilan. In his doctoral thesis, he combines studies at the Faculty of Engineering and the Department of Computer Science. "Throughout the entire period, I received comprehensive support from both the academic and administrative staff. In academic research, I have had excellent supervisors over the years who not only promoted the research directly, but also spent a lot of time in personal pedagogy - taught and guided how research should be conducted, what mistakes to avoid and how to focus the subject of the study".

Alongside his research work, he works as a practitioner in courses at the Faculty of Engineering, supervises a number of master's degree students and advises on a number of projects. At the beginning of his doctorate - thanks to the President's scholarship he won, Ernest began volunteering to accompany families in the "Bells" organization - an organization that accompanies families in financial distress and educates them on proper financial management. "I have already accompanied several families who managed to improve their financial situation thanks to the hard work they did, and I hope I can continue this project even after the doctorate," he says proudly.