Search
605 resources
-
Sami Baral, Anthony F. Botelho, John A. ...|Dec 1st, 2021|conferencePaperSami Baral, Anthony F. Botelho, John A. ...Dec 1st, 2021
-
Randy Elliot Bennett, Mo Zhang, Sandip S...|Dec 1st, 2021|journalArticleRandy Elliot Bennett, Mo Zhang, Sandip S...Dec 1st, 2021
This study examined differences in the composition processes used by educationally at-risk males and females who wrote essays as part of a high-school equivalency examination. Over 30,000 individuals were assessed, each taking one of 12 forms of the examination’s language arts writing subtest in 23 US states. Writing processes were inferred using features extracted from keystroke logs and aggregated into seven composite indicators. Results showed that females earned higher essay and total...
-
Guembe Blessing, Ambrose Azeta, Sanjay M...|Dec 1st, 2021|bookSectionGuembe Blessing, Ambrose Azeta, Sanjay M...Dec 1st, 2021
-
Rishi Bommasani, Drew A. Hudson, Ehsan A...|Dec 1st, 2021|journalArticleRishi Bommasani, Drew A. Hudson, Ehsan A...Dec 1st, 2021
AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical...
-
Elizabeth Clark, Tal August, Sofia Serra...|Dec 1st, 2021|preprintElizabeth Clark, Tal August, Sofia Serra...Dec 1st, 2021
Human evaluations are typically considered the gold standard in natural language generation, but as models' fluency improves, how well can evaluators detect and judge machine-generated text? We run a study assessing non-experts' ability to distinguish between human- and machine-authored text (GPT2 and GPT3) in three domains (stories, news articles, and recipes). We find that, without training, evaluators distinguished between GPT3- and human-authored text at random chance level. We explore...
-
Elizabeth Clark, Tal August, Sofia Serra...|Dec 1st, 2021|preprintElizabeth Clark, Tal August, Sofia Serra...Dec 1st, 2021
Human evaluations are typically considered the gold standard in natural language generation, but as models' fluency improves, how well can evaluators detect and judge machine-generated text? We run a study assessing non-experts' ability to distinguish between human- and machine-authored text (GPT2 and GPT3) in three domains (stories, news articles, and recipes). We find that, without training, evaluators distinguished between GPT3- and human-authored text at random chance level. We explore...
-
Saad Khan, Jesse Hamer, Tiago Almeida|Dec 1st, 2021|conferencePaperSaad Khan, Jesse Hamer, Tiago AlmeidaDec 1st, 2021
We present Generate, a AI-human hybrid system to help education content creators interactively generate assessment content in an efficient and scalable manner. Our system integrates advanced natural language generation (NLG) approaches with subject matter expertise of assessment developers to efficiently generate a large number of highly customized and valid assessment items. We utilize the powerful Transformer architecture which is capable of leveraging substantive pretraining on several...
-
Vivekanandan S. Kumar, David Boulanger|Sep 1st, 2021|journalArticleVivekanandan S. Kumar, David BoulangerSep 1st, 2021
-
Michael McTear, Michael McTear|Dec 1st, 2021|bookSectionMichael McTear, Michael McTearDec 1st, 2021
-
Gabriel Oliveira dos Santos, Esther Luna...|Dec 1st, 2021|preprintGabriel Oliveira dos Santos, Esther Luna...Dec 1st, 2021
This paper shows that CIDEr-D, a traditional evaluation metric for image description, does not work properly on datasets where the number of words in the sentence is significantly greater than those in the MS COCO Captions dataset. We also show that CIDEr-D has performance hampered by the lack of multiple reference sentences and high variance of sentence length. To bypass this problem, we introduce CIDEr-R, which improves CIDEr-D, making it more flexible in dealing with datasets with high...
-
University of Wolverhampton, UK, Hadeel ...|Dec 1st, 2021|conferencePaperUniversity of Wolverhampton, UK, Hadeel ...Dec 1st, 2021
-
Masaki Uto|Jul 1st, 2021|journalArticleMasaki UtoJul 1st, 2021
Abstract Automated essay scoring (AES) is the task of automatically assigning scores to essays as an alternative to grading by humans. Although traditional AES models typically rely on manually designed features, deep neural network (DNN)-based AES models that obviate the need for feature engineering have recently attracted increased attention. Various DNN-AES models with different characteristics have been proposed over the past few years. To our knowledge, however, no study has...
-
Cong Wang, Xiufeng Liu, Lei Wang|Apr 1st, 2021|journalArticleCong Wang, Xiufeng Liu, Lei WangApr 1st, 2021
-
Tianqi Wang, Hiroaki Funayama, Hiroki Ou...|Dec 1st, 2021|journalArticleTianqi Wang, Hiroaki Funayama, Hiroki Ou...Dec 1st, 2021
-
Zichao Wang, Andrew Lan, Richard Baraniu...|Dec 1st, 2021|conferencePaperZichao Wang, Andrew Lan, Richard Baraniu...Dec 1st, 2021
-
Kim Christopher Williamson, René F. Kizi...|Dec 1st, 2021|conferencePaperKim Christopher Williamson, René F. Kizi...Dec 1st, 2021
-
Mike Wu, Noah Goodman, Chris Piech|Dec 1st, 2021|journalArticleMike Wu, Noah Goodman, Chris PiechDec 1st, 2021
High-quality computer science education is limited by the difficulty of providing instructor feedback to students at scale. While this feedback could in principle be automated, supervised approaches to predicting the correct feedback are bottlenecked by the intractability of annotating large quantities of student code. In this paper, we instead frame the problem of providing feedback as few-shot classification, where a meta-learner adapts to give feedback to student code on a new programming...
-
Khensani Xivuri, Hossana Twinomurinzi, D...|Dec 1st, 2021|bookSectionKhensani Xivuri, Hossana Twinomurinzi, D...Dec 1st, 2021
-
Jing Xu, Da Ju, Margaret Li|Dec 1st, 2021|conferencePaperJing Xu, Da Ju, Margaret LiDec 1st, 2021
-
Goh Ying Yingsoon, Niaz Chowdhury, Ganes...|Dec 1st, 2021|bookSectionGoh Ying Yingsoon, Niaz Chowdhury, Ganes...Dec 1st, 2021
The teaching of Chinese as a foreign language can be supported by using AI technology. Traditionally, the non-native learners can only interact with the instructors and depend on them solely for speaking practices. However, with the advancement of AI technology, the learners can use AI technology for interactive speaking skill development. In this study, the learners were instructed to download an application at https://m.wandoujia.com/apps/6790950. The process on the preparation of this AI...