Limits of Detecting Text Generated by Large-Scale Language Models
Authors:
- 2 February 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Abstract
Some consider large-scale language models that can generate long and coherent pieces of text as dangerous, since they may be used in misinformation campaigns. Here we formulate large-scale language model output detection as a hypothesis testing problem to classify text as genuine or generated. We show that error exponents for particular language models are bounded in terms of their perplexity, a standard measure of language generation performance. Under the assumption that human language is stationary and ergodic, the formulation is ex-tended from considering specific language models to considering maximum likelihood language models, among the class of k-order Markov approximations; error probabilities are characterized. Some discussion of incorporating semantic side information is also given.Keywords
All Related Versions
This publication has 18 references indexed in Scilit:
- Authentication theory and hypothesis testingIEEE Transactions on Information Theory, 2000
- Measure concentration for a class of random processesProbability Theory and Related Fields, 1998
- Bounding $\bar{d}$-distance by informational divergence: a method to prove measure concentrationThe Annals of Probability, 1996
- Adaptive tests for stochastic processes in the ergodic caseStochastic Processes and their Applications, 1993
- A convergent gambling estimate of the entropy of EnglishIEEE Transactions on Information Theory, 1978
- Codes based on inaccurate source probabilitiesIEEE Transactions on Information Theory, 1971
- On estimating the entropy of random fieldsInformation and Control, 1970
- A note on the entropy of partially-known languagesInformation and Control, 1968
- The reconstruction of abbreviated printed messages.Journal of Experimental Psychology, 1954
- Prediction and Entropy of Printed EnglishBell System Technical Journal, 1951