The End is Nigh: Generic Solving of Text-based CAPTCHAs
By Elie Bursztein, Jonathan Aigrain, A. Moscicki, John C. Mitchell. 2014
157 citations recorded by the source corpus.
Subfields: Natural Language Processing, Safety, Robustness & Interpretability
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Natural Language Processing
- José Hernandez-Orallo
- Atilim Gunes Baydin
- Dan Boneh
- Tuan Anh Le
- Frank D. Wood
- Florian Tramèr
- Brooks Paige
- Jan-Willem van de Meent
In Safety, Robustness & Interpretability
- Florian Tramèr
- Dan Boneh
- Alexander Lavin
- José Hernandez-Orallo
- Pascal Dupré
- Giancarlo Pellegrino
- Gili Rusak
- Nicholas Carlini
Citations in the corpus
- An Introduction to Probabilistic Programming (2018, background): “” A non-probabilistic programming approach would require gathering a very large number of Captchas, hand-labeling them all, then designing and training a neural network to regress from the image to a text string (Bursztein et al., 2014).”
- Inference Compilation and Universal Probabilistic Programming (2016, methodology): “With the resulting artifacts, running inference on a test Captcha takes 100 ms, whereas durations ranging from 500 ms (Starostenko et al., 2015) to 7.95 s (Bursztein et al., 2014) have been reported with segment-and-classify approaches.”
- Evaluation in artificial intelligence: from task-oriented to ability-oriented measurement (2017, background): “Logically, when character recognition systems and other techniques improve, current CAPTCHAs are broken (see, e.g., Bursztein et al, 2014), and CAPTCHAs need to be updated to more distorted words or to other tasks that are beyond AI technology.”
- A generative vision model that trains with high data efficiency and breaks text-based CAPTCHAs (2017, methodology): “A recent deep-learning approach for parsing one specific CAPTCHA style required millions of labeled examples from it (6), and earlier approaches mostly relied on handcrafted style-specific heuristics to segment out the character (3, 7), whereas humans can solve new styles without explicit training (”
- AdVersarial: Perceptual Ad Blocking meets Adversarial Machine Learning (2018, background): “As ML models can solve CAPTCHAs [15, 94], one may wonder why transcribing ad disclosures is harder.”