Pronounce epoch, gradient, stochastic, perceptron, and hyperparameter with confidence — the core ML training vocabulary for engineers and researchers.
0 / 5 completed
1 / 5
How is 'epoch' pronounced in a machine learning context?
Epoch is pronounced /ˈiːpɒk/ — 'EE-pok'. The word comes from the Greek ἐποχή (epochē), meaning a fixed point in time. 'E' = /iː/ (long 'ee'). 'poch' = /pɒk/ (short /ɒ/ as in 'lock', 'rock'). Two syllables: EE-pok, stress on the first. In machine learning, an epoch is one complete pass through the entire training dataset during model training. Training typically runs for multiple epochs; the number of epochs is a hyperparameter that affects model convergence and overfitting.
2 / 5
How is 'gradient' pronounced?
Gradient is pronounced /ˈɡreɪdiənt/ — 'GRAY-dee-unt'. The word comes from Latin 'gradiens' (stepping, walking). 'Gra' = /ɡreɪ/ (diphthong /eɪ/ as in 'grey'). '-di' = /diː/ (long 'ee'). '-ent' = /ənt/ (schwa). Three syllables: GRAY-dee-unt, stress on the first. In ML, the gradient is the vector of partial derivatives of a loss function with respect to model parameters, indicating the direction and rate of steepest increase. Gradient descent minimises the loss by moving in the opposite direction.
3 / 5
How is 'stochastic' pronounced?
Stochastic is pronounced /stəˈkæstɪk/ — 'stuh-KAS-tik'. The word comes from the Greek στοχαστικός (stochastikos), meaning 'skilful in aiming' or 'of guessing'. 'Sto' = /stə/ (schwa, unstressed). 'chas' = /kæs/ (short /æ/). '-tic' = /tɪk/. Three syllables: stuh-KAS-tik, stress on the second. In ML, Stochastic Gradient Descent (SGD) uses a random subset (mini-batch) of training data to compute the gradient at each step, making training faster and introducing beneficial noise that can help escape local minima.
4 / 5
How is 'perceptron' pronounced?
Perceptron is pronounced /pərˈsɛptrɒn/ — 'pur-SEP-tron'. The name was coined by Frank Rosenblatt in 1957, blending 'perception' + '-tron' (a suffix from Greek for instrument). 'Per' = /pər/ (schwa). 'cep' = /sɛp/ (stressed, short /ɛ/). '-tron' = /trɒn/ (short /ɒ/). Three syllables: pur-SEP-tron, stress on the second. A perceptron is the simplest type of artificial neuron, computing a weighted sum of inputs and applying an activation function — the building block of neural networks and the precursor to modern deep learning.
5 / 5
How is 'hyperparameter' pronounced?
Hyperparameter is pronounced /ˌhaɪpərpəˈræmɪtər/ — 'HY-pur-puh-RAM-ih-tur'. The word combines 'hyper-' (beyond, above) + 'parameter'. 'Hy' = /haɪ/ (diphthong). 'per' = /pər/ (schwa). 'pa' = /pə/ (schwa). 'ram' = /ræm/ (short /æ/, the stressed syllable). '-eter' = /ɪtər/ (short /ɪ/ + schwa). Five syllables: HY-pur-puh-RAM-ih-tur, primary stress on the first syllable. A hyperparameter is a parameter set before training begins — such as learning rate, batch size, and number of layers — as opposed to model parameters learned during training.