Tortured Phrases in Scientific Writing
"Tortured phrases" is a term coined by researchers to describe strange, incorrect, or unnatural substitutions of technical terms that often appear in low-quality, AI-generated or paper-mill–generated scientific texts.
Here are some examples:
Artificial Intelligence & Computing- Artificial Intelligence (AI) → counterfeit consciousness, fake intelligence
- Machine Learning → apparatus learning, mechanical acquiring
- Deep Learning → profound learning, deep tutoring
- Neural Network → neuron net, brain organization
- Convolutional Neural Network (CNN) → folded brain arrangement, convoluted neuron grid
- Random Forest → arbitrary woodland, stochastic trees
- Support Vector Machine (SVM) → prop-up vector contraption, help outline motor
- Gradient Descent → incline reduction, slope fall
Mathematics & Statistics- Mean Square Error (MSE) → average squared fault, mean square slip-up
- Principal Component Analysis (PCA) → main constituent examination, fundamental section review
- Logarithm → log book, proportionate mathematics
- Eigenvalue/Eigenvector → characteristic worth, own vector
- Regression Analysis → backslide examination
Engineering & Physics- Signal-to-noise ratio (SNR) → flag-to-commotion proportion, signal vs. turbulence
- Finite Element Method (FEM) → restricted component strategy
- Quantum Entanglement → quantum perplexity, quantum complexity
- Thermodynamics → warmth elements, heat force study
Biology & Medicine- Gene Expression → quality articulation
- Polymerase Chain Reaction (PCR) → polymerase loop response
- Molecular Docking → atomic harborage, molecule mooring
- Placebo Effect → fake treatment impact
- Antimicrobial Resistance (AMR) → microbe opposition
Why These Are Red Flags- They signal paper mills or paraphrased plagiarism.
- They reduce scientific clarity and searchability.
- No genuine researcher would use these phrasings in proper context.
|
