Machine learning develops algorithms that use data or experience to improve performance on a specified task. The task, the experience available to the algorithm and the performance criterion must all be stated. A statistical classification system might learn from labelled training data; an unsupervised learning method instead looks for structure without those labels.^[ref]

For example, a linear regression model can learn coefficients predicting a numerical outcome from measured inputs. Minimizing its error on the training data determines a fitted model, but does not establish how well it predicts new observations. A separate test set estimates performance on data withheld from fitting. Overfitting occurs when fitting peculiarities of the training data impairs this generalization; data leakage can make the evaluation misleading.^[ref]

Statistical learning supplies tools for reasoning about these prediction and uncertainty questions. A useful learned model is therefore assessed through an explicit task and evaluation, alongside the running time and other resources needed to train and use it.

An antibiotic that interferes with bacterial protein synthesis. Its resistance markers have been used in molecular genetics, including the 1973 plasmid construction experiment, linking an experimentally selectable growth trait with carriage of a genetic construct.

An aminoglycoside antibiotic that interferes with bacterial protein synthesis. Resistance can provide a selectable marker in molecular genetics, as in the 1973 plasmid construction experiment; survival under selection identifies a phenotype rather than every detail of the inserted DNA.

The ability of a microorganism to grow under antibiotic exposure that inhibits a susceptible comparison, often reflected in an increased minimum inhibitory concentration. Survival alone does not establish resistance: antibiotic tolerance and bacterial persistence concern killing and survival without necessarily changing that growth threshold.

A dipole antenna uses two conducting arms carrying an oscillating current, commonly fed across a central gap. Its changing charge and current distributions produce electromagnetic radiation. A short straight dipole has far-zone power proportional to ${\sin }^2\theta$ relative to its axis: it radiates most strongly sideways and has axial nulls. A practical half-wave dipole has a nonuniform current distribution and a different exact pattern and input impedance. Geometric wire length, electrical length and the surrounding dielectric or conducting structures all affect resonance and matching.

An antenna beam pattern gives an antenna's directional response as a function of observation angle. A normalized power pattern divides radiation intensity by its peak; a field-amplitude pattern is its square root under a consistent impedance convention. The beam solid angle integrates the normalized power pattern over direction, so ideal directivity is $4\pi /{\Omega }_A$. A short dipole antenna has a power pattern proportional to ${\sin }^2\theta$, zero along its axis and maximum broadside. A narrow main lobe can coexist with sidelobes, so beamwidth alone does not specify the complete response.

The antenna temperature expresses the noise power delivered by a receiving antenna as an equivalent input noise temperature. It depends on the sky and surroundings weighted by the antenna beam pattern, as well as losses in the antenna; it need not be the physical temperature of either the antenna or a single astronomical source.

Antenna gain compares the far-field energy flux in a specified direction with the flux that an isotropic radiator would produce using the same input power. Thus $S_r=G{P}_{\mathrm{in}}/(4\pi r^2)$. Gain includes radiation efficiency; it must be distinguished from a voltage or current amplifier gain and from directivity when losses matter.

An antenna aperture describes the collecting area associated with a receiving antenna. The antenna effective area need not equal its geometric projected area.

The Friis transmission equation describes a far-field free-space antenna link with aligned directions, matched polarizations and negligible additional loss. It follows by multiplying the transmitter's axial energy flux $P_t{G}_t/(4\pi d^2)$ by the receiver's antenna effective area ${\lambda }^2{G}_r/(4\pi )$. A receiving impedance mismatch multiplies the available received power by $1-|{\Gamma }_L{|}^2$. Wavelength scaling must state which gains, apertures and impedances are held fixed.

Radiation resistance represents emitted electromagnetic power as a real input resistance of an antenna. It does not describe conversion to heat inside the conductor. If peak current amplitude is used, $P_{\mathrm{rad}}=|I_0{|}^2{R}_{\mathrm{rad}}/2$. Real ohmic loss adds a separate resistance and lowers radiation efficiency. For an ideal short uniform-current element of length $d$, the Hertzian dipole gives $R_{\mathrm{rad}}=80{\pi }^2(d/\lambda {)}^2$ in ohms, using the usual vacuum impedance approximation $120\pi \Omega$.

A phased array combines radiation from spatially separated antennas with controlled relative phases. Changing those phases steers the direction of constructive interference without moving the elements. For an equally spaced linear array with spacing $d$, free-space wavenumber $k$ and phase increment $\psi$, a principal beam at angle ${\theta }_0$ from the array normal satisfies $\psi =kd\sin {\theta }_0$ when the time dependence is $e^{-i\omega t}$ and the element drives are $e^{ij\psi }$.

Histology studies the microscopic organization of tissues. Thin sections, imaging and histological staining reveal how cells and extracellular matrix form a working structure. A two-dimensional section samples a three-dimensional arrangement, so section orientation and preparation influence interpretation.

Single-sideband modulation transmits one of the two translated message bands generated by amplitude modulation. For a real signal, the transmitted negative-frequency band is the complex conjugate of the positive-frequency band. Removing the redundant band reduces the required signal bandwidth. The carrier wave may also be suppressed; its retention or removal must be specified separately.

An aerospace engineer applies aerospace engineering methods to flight vehicles or their components. The work can involve modeling aerodynamic loads, calculating trajectories, designing propulsion or evaluating structural and thermal response. A spacecraft instrument mount, for example, must retain alignment through launch vibration and temperature changes. The analysis links environmental assumptions to measurable component requirements and to how the entire vehicle operates.

An ordinary acoustic wave involving a pressure and density oscillation in local thermodynamic equilibrium. In a Fermi liquid, this hydrodynamic response requires collisions fast compared with the oscillation period; zero sound instead describes a collisionless collective mode.

Acoustic attenuation is a reduction of a propagating sound signal beyond that attributed to the specified geometric spreading. Absorption converts acoustic energy to other forms, while scattering redirects it away from the receiving path. The inferred attenuation therefore depends on what the detector collects and how the spreading reference is defined.

For a plane wave with pressure amplitude $p(x)=p(0)e^{-\alpha x}$, the corresponding sound intensity decays as $I(x)=I(0)e^{-2\alpha x}$ in a fixed medium. An attenuation coefficient defined from intensity has twice the numerical value of this amplitude coefficient. Measuring the signal at several separations can estimate the slope, provided the source response, reflections and geometric corrections are controlled.

Reverberation is the persistence of acoustic waves after a source stops, produced by repeated reflections in an enclosure. A diffuse-room approximation gives an approximately exponential decay of mean acoustic energy; an individual normal mode may instead be described as a damped resonator.

The conventional reverberation time $T_{60}$ is the time for the level of decaying acoustic energy to decrease by $60$ decibels, a factor $10^6$ in energy or $10^3$ in pressure wave amplitude. For exponential energy decay $E(t)=E(0)e^{-t/{\tau }_E}$, $T_{60}=6\ln (10){\tau }_E$. It is not the exponential relaxation time itself and cannot be inferred exactly from an unspecified audibility threshold.

A particle is an entity described as a localized constituent in an appropriate physical model. In classical mechanics, a point particle has a position and trajectory while its internal size is neglected. This approximation can describe an extended object when the omitted structure does not affect the question being studied.

In quantum mechanics, a quantum particle has a state described by a wave function or a more general state description, with probabilistic measurement outcomes. Quantum field theory describes particle states as excitations of fields. A proton can be treated as a particle in one experiment while its quark and gluon structure matters in another. The particle description therefore specifies what is being resolved, rather than asserting that every particle is a structureless classical point.