Guest Column: The Psychotherapist with Fur and Four Legs!  |  Health Monitor: Dealing With Post Covid Syndrome  |  National Edu News: Secretary Higher Education urges students to emerge as job creators  |  National Edu News: PM addresses the 18th Convocation of Tezpur University, Assam  |  Leadership Instincts: Experts highlight the need for strengthening centre-state cooperation  |  Policy Indications: India’s global position rises both in innovations & publications  |  Education Information: Written Result of Indian Economic Service/Indian Statistical Service Examination  |  National Edu News: AstroSat’s Ultraviolet Imaging Telescope spots rare ultraviolet-bright stars  |  Parent Interventions: Randomized trials could help to return children safely to schools   |  Parent Interventions: How fellow students improve your own grades   |  Parent Interventions: School-made lunch 'better' for children  |  Teacher Insights: Second Anniversary of India Science, Nation’s OTT Channel  |  Leadership Instincts: Participation of MGIEP in the Implementation of NEP 2020  |  Teacher Insights: World of Puzzling Patterns  |  Education Information: HKUST Collaborates with Hang Lung to Foster Young Mathematics Talent  |  
July 27, 2019 Saturday 09:51:27 AM IST

Statistical Method Identified to Detect Fake or Computer-Generated Text

By Felix Litchenfeld

Researchers at Harvard John A Paulson School of Engineering and Applied Sciences (SEAS) and IBM have developed a statistical method to detect computer-generated or fake text from human generated text.
Researchers Sebastian Gehrmann and Hendrik Strobert(IBM) found that natural-language generators are trained on tens of millions of online texts and mimic human language by predicting the words that most often follow one another. For eg I followed by 'have' and 'am'. Using this idea they developed a method that identifies predictable text instead of flagging errors. Gehrmann and Strobelt’s method, known as GLTR, is based on a model trained on 45 million texts from websites — the public version of the OpenAI model, GPT-2. Because it uses GPT-2 to detect generated text, GLTR works best against GPT-2, but it does well against other models, too.


https://news.harvard.edu/gazette/story/2019/07/researchers-develop-a-method-to-identify-computer-generated-text/?utm_medium=Feed&utm_source=Syndication



Comments