# Algebraic geometry Algebraic geometry is the branch of [[Mathematics]] that studies varieties — the solution sets of systems of polynomial equations — by running a two-way dictionary between [[Geometry]] and the rings, ideals, and [[Field_(mathematics)|fields]] of [[Abstract_algebra]]. A circle is the variety x² + y² = 1; the reachable poses of a robot arm form a variety cut out by kinematic constraints. Hilbert's Nullstellensatz (1893) guarantees that over an algebraically closed field the dictionary loses nothing, so geometric questions become algebra that a machine can push through [[Computer_algebra]]. The dividends radiate outward: the [[Genus_(mathematics)|genus]] that classifies curves, the Zariski [[Topology|topology]] that makes "generic" an exact word, Gröbner-basis [[Algorithm|algorithms]] that solve polynomial systems in [[Robotics]] and [[Computer_vision]], elliptic curves that secure modern [[Cryptography]], and codes built from curves over finite fields that push [[Error_detection_and_correction]] past classical bounds. ## The ideal–variety dictionary and Hilbert's two pillars An affine variety V ⊆ kⁿ is the common zero locus of polynomials f₁, …, f_m over a [[Field_(mathematics)|field]] k; conversely, any point set has an ideal I(V) of polynomials vanishing on it. Hilbert supplied both load-bearing theorems. The basis theorem (1890) says every ideal of k[x₁, …, x_n] is finitely generated, so every variety needs only finitely many equations — the finiteness that [[Computer_algebra]] exploits. The Nullstellensatz (1893) says that over an algebraically closed field, I(V(J)) = √J, an exact match between radical ideals and varieties. [[Geometry]] thereby becomes commutative [[Abstract_algebra|algebra]]: points are maximal ideals, irreducible components are minimal primes, dimension is the length of a chain of primes, and tangent spaces are kernels of Jacobians, plain [[Linear_algebra]]. Even the [[Graph_of_a_function|graph of a polynomial map]] is itself a variety. The Zariski [[Topology|topology]], whose closed sets are exactly the varieties, is deliberately coarse: on an irreducible variety every nonempty open set is dense, which is why "generically true" is a theorem-grade notion — a discipline that [[Dynamical_system|dynamical systems]] and [[Statistics]] borrow whenever they argue about almost-all parameter values. ## Projective completion and Bézout's exact count Intersection theory only becomes exact after two repairs: add points at infinity, and count multiplicity. In the [[Projective_geometry|projective]] plane over the complex numbers, Bézout's theorem (1779) states that curves of degrees m and n with no common component meet in exactly m×n points. Over ℝ the count can fail — a line can miss a conic entirely — so the passage to [[Complex_analysis|complex]] coordinates is doing real work. Classical geometers mined this exactness for integer invariants: every smooth cubic surface contains exactly 27 lines (Cayley and Salmon, 1849), and a smooth plane curve of degree d has genus g = (d−1)(d−2)/2, tying the [[Genus_(mathematics)|genus]] to raw degree data. Such numbers behave like the [[Combinatorics|combinatorial]] invariants of [[Surface_(topology)|surfaces]]: integers, stable under deformation, computable by bookkeeping — the enumerative habit that Schubert systematized in 1879 and that [[Mathematical_physics]] later re-derived from string-theoretic mirror symmetry. ## Curves as Riemann surfaces: the genus dividend Over ℂ, a smooth projective curve is a compact Riemann surface — the bridge Riemann built in 1857 between algebra and [[Complex_analysis]] — and its [[Genus_(mathematics)|genus]], the handle count of the underlying [[Surface_(topology)|surface]], governs everything. Genus 0 curves admit rational parametrizations. Genus 1 curves are elliptic curves carrying a [[Group_theory|group]] law on their points; that structure over finite [[Field_(mathematics)|fields]] is the engine of elliptic-curve [[Cryptography]] proposed independently by Koblitz and Miller in 1985. Genus ≥ 2 curves have only finitely many rational points (Faltings, 1983). Riemann–Roch (1865) counts functions with prescribed poles, and evaluating such functions at rational points of curves over finite fields yields Goppa's algebraic-geometry codes: in 1982 Tsfasman, Vlăduţ, and Zink showed they beat the Gilbert–Varshamov bound, a landmark for [[Error_detection_and_correction]] and [[Information_theory]]. ## Schemes: geometry rebuilt from rings alone Grothendieck's refoundation (1960s) assigns a geometric object, the spectrum, to *every* commutative ring — the move that makes [[Abstract_algebra|arithmetic]] and geometry one subject. Nilpotent elements track infinitesimal thickenings, doing algebraically what derivatives do in [[Calculus]]; schemes glue from local pieces the way a [[Manifold]] glues from charts; and the functor-of-points view runs on [[Category_theory]], with foundational care (Grothendieck universes) bordering [[Set_theory]]. The payoff was the Weil conjectures: counting solutions over finite [[Field_(mathematics)|fields]] is governed by the [[Topology|topological]] invariants of the complex solution set, with point counts packaged into zeta functions like a [[Combinatorics|generating function]]. Deligne completed the proof in 1974, confirming that a variety's discrete, arithmetic shadow and its continuous shape carry the same information. ## Gröbner bases turn the dictionary into an algorithm Buchberger's Gröbner bases (1965) made the ideal–variety dictionary executable: fix a monomial order, complete your generators, and ideal membership, projection (elimination of variables), and solving all reduce to a generalized long division. The worst case is brutal — ideal membership is EXPSPACE-hard (Mayr–Meyer, 1982), with doubly exponential degree growth — a caution from the [[Theory_of_computation]] that practical [[Algorithmic_efficiency]] here is input-dependent. Yet tuned implementations inside [[Computer_algebra]] systems and dedicated [[Mathematical_software]] routinely dispatch real instances: inverse kinematics in [[Robotics]] (the general Stewart platform has 40 assembly configurations), minimal problems in [[Computer_vision]] such as five-point relative pose, steady states of mass-action [[Ordinary_differential_equation|ODE]] models in [[Systems_biology]], and sums-of-squares relaxations in polynomial [[Mathematical_optimization]] (Lasserre, 2001). ## Where varieties meet the systems sciences Algebraic statistics treats conditional-independence models as varieties in the space of [[Probability_distribution|distributions]], giving [[Statistics]] exact tests of model geometry; phylogenetic invariants do the same for tree models in [[Mathematical_and_theoretical_biology]]. Identifiability of low-rank [[Tensor]] decompositions — workhorses of modern [[Machine_learning]] — is a question about secant varieties. Structural identifiability of [[Ordinary_differential_equation|ODE]] models, central to [[System_identification]] and [[Control_theory]], is decided by differential-algebraic elimination. Classically integrable [[Dynamical_system|dynamical systems]], like the Kovalevskaya top (1889), linearize on abelian varieties. And mirror symmetry (1991) predicted curve counts on the quintic threefold — beginning with its 2,875 lines — that enumerative geometers later verified, closing a loop between [[Mathematical_physics]] and the oldest counting problems in the subject. **On the spine:** [[Abstract_algebra]] · [[Projective_geometry]] · [[Genus_(mathematics)]] · [[Computer_algebra]] · [[Field_(mathematics)]]. ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Algebraic_geometry) : [Wikitube](https://en.wikitube.io/wiki/Algebraic_geometry) ## Previous hub tags Hubs: `Systems`. Portals: [[PORTAL_Systems]], [[PORTAL_Graph_theory]], [[PORTAL_Dynamical_system]], [[PORTAL_Information_theory]], [[PORTAL_Control_theory]]. --- *Repopulated 2026-08-12 · redlink fill · 0 deletions.*