Automated Essay Scoring by Beata Beigman Klebanov and Nitin Madnani

Automated essay scoring is concerned with the development of language technologies that make it possible for an essay -or any piece of writing for that matter -to be evaluated or scored by a computer.These technologies find their utility primarily in the context of educational measurement, where they serve a dual purpose.On the one hand, they provide crucial support to educators and institutions, facilitating the assessment of students' writing skills and content knowledge.A good example is the TOEFL iBT ® , the Internet-based Test of English as a Foreign Language, administered by the Educational Testing Service and widely adopted by institutions worldwide.On the other hand, these technologies benefit writers themselves, including students, by offering a platform to assess and enhance their writing skills.One illustrative tool for this purpose is the Write & Improve software developed at the University of Cambridge.The field of automated essay scoring emerged in the pioneering era of artificial intelligence and computational linguistics, with the inception of Ellis Page's Project Essay Grade system at the University of Connecticut in 1964, and the subsequent publication of his seminal article The Imminence of... Grading Essays by Computer in 1966.Page's automated scoring system is seen by the authors of this book as one of the first concrete applications of natural language processing, after the Audrey system for speech recognition and the Georgetown-IBM demonstration of machine translation.But just like speech recognition and machine translation, the industrialization of automated essay scoring mostly gained momentum in the 1990s.This decade saw an increase in richly annotated language data and corpora, which enabled the use of statistical and supervised machine learning in developing essay-scoring systems.Pearson's Intelligent Essay Assessor ™ , released in 1998, is a prominent example of this evolution.A year later, in 1999, the e-rater ® scoring engine was launched by the Educational Testing Service, commonly abbreviated as ETS, a private organization overseeing several standardized tests and high-stakes examinations such as TOEFL ® and GRE ® .This book is written by Beata Beigman Klebanov and Nitin Madnani, two researchers at ETS with many years of research experience and numerous peer-reviewed publications in the field of automated essay scoring.Their book provides a concise yet indispensable introduction to the field.After this short introduction, the eager reader is invited to take a deep dive into the scientific literature, encompassing slightly over half of the book's content, approximately 115 pages.The literature review primarily caters to computational linguists and NLP practitioners, as it delves comprehensively into diverse machine learning models -ranging from linear regression to artificial neural 1