Historical task source. The Eve task board holds current task state.
Phase 109: Kalika Frontier Methods — Tropical Geometry, Positive Geometry, Resurgence, Bootstrap, Applied Category Theory, Differentiable Physics & Quantum Simulation#
Implements the cutting-edge computational methods that are known primarily in specialized research communities. These are the tools that most software platforms don't even attempt — representing genuine competitive differentiation. Informed by deep research into the latest work of Arkani-Hamed (positive geometry), Francis Brown (motivic amplitudes), Simmons-Duffin (conformal bootstrap), Borinsky (tropical methods), and others at the absolute frontier of computational physics and mathematics.
109.1 Tropical Geometry (@kalika/tropical)#
- 109.1.1.1 Implement tropical semiring: min-plus algebra (a ⊕ b = min(a,b), a ⊗ b = a + b) and max-plus algebra, tropical polynomial evaluation, tropical root finding (corner loci of piecewise-linear functions), and tropical convexity — the algebraic foundation for all tropical computation
- 109.1.1.2 Implement tropical variety computation: given a polynomial ideal, compute its tropicalization (the set of points where the tropical polynomial achieves its minimum at least twice) using Gfan-style algorithms — matching polymake's tropical geometry capabilities
- 109.1.1.3 Implement tropical Grassmannian computation: Trop(Gr(2,n)) (space of phylogenetic trees / metric trees) and Trop(Gr(k,n)) for higher k — relevant for tree-level string amplitudes and cluster algebras for N=4 SYM amplitudes (Drummond-Foster-Gürdogan-Kalousios)
- 109.1.1.4 Implement tropical Monte Carlo integration for Feynman integrals (Borinsky's method): use the tropical limit of Symanzik polynomials to identify dominant integration regions, construct importance sampling distributions from the tropical fan, and evaluate multi-loop Feynman integrals with dramatically reduced variance — this is the cutting-edge technique for numerical amplitude computation
- 109.1.1.5 Implement Newton polytope computation: for Symanzik polynomials of Feynman graphs, compute Newton polytopes and their fine triangulations — the combinatorial data encoding the singularity structure of amplitudes
109.2 Motivic Amplitudes & Periods (@kalika/periods)#
- 109.2.1.1 Implement multiple zeta value (MZV) algebra: MZV computation to arbitrary precision, stuffle and shuffle product relations, double shuffle relations, Hoffman basis, MZV datamine database integration (Blumlein-Broadhurst-Vermaseren), motivic MZV arithmetic (coaction, period polynomial) — the algebra of periods that Feynman integrals evaluate to
- 109.2.1.2 Implement multiple polylogarithm computation: Li*{s1,...,sk}(z1,...,zk) in all standard representations (iterated integral, series, shuffle regularized), Chen iterated integrals on M*{0,n}, symbol map (projecting to tensor algebra), and coproduct computation — matching GiNaC/PolyLogTools capabilities
- 109.2.1.3 Implement HyperInt-style Feynman integral evaluation: evaluate Feynman parameter integrals by iterating one-dimensional integrals using hyperlogarithm methods (Panzer's algorithm), automatically determining the function class of the result (MZVs, alternating sums, multiple polylogarithms) — targeting massless propagator-type integrals to high loop order
- 109.2.1.4 Implement motivic coaction computation: given a Feynman integral period, compute its motivic coaction (Brown-Dupont), decomposing the result into simpler periods — providing structural constraints on amplitudes beyond what numerical evaluation reveals
- 109.2.1.5 Implement single-valued projection: the map sv: MZVs → single-valued MZVs connecting open string amplitudes to closed string amplitudes — relevant for string theory amplitude computation
- 109.2.1.6 Implement elliptic multiple polylogarithm evaluation: for Feynman integrals that go beyond genus zero (e.g., sunrise/banana diagrams), evaluate in terms of iterated integrals on elliptic curves — matching Broedel-Duhr-Dulat-Tancredi methods
109.3 Positive Geometry (@kalika/positive-geometry)#
- 109.3.1.1 Implement positive geometry framework: abstract positive geometry (interior, boundary, canonical form), canonical form computation via triangulation (decompose positive geometry into simplices, sum simplex forms), and Jeffrey-Kirwan residue computation — the core infrastructure for amplituhedron and cosmological polytope calculations
- 109.3.1.2 Implement amplituhedron computation: amplituhedron
boundaries and triangulations for tree-level and one-loop planar N=4 SYM,
canonical form extraction (= scattering amplitude), BCFW triangulation —
matching Lukowski-Parisi-Williams'
amplituhedronBoundariespackage - 109.3.1.3 Implement associahedron and permutohedron: realization of associahedra as convex polytopes, facet structure, canonical form = biadjoint scalar tree amplitude (Arkani-Hamed-Bai-He-Yan), generalized associahedra from cluster algebras
- 109.3.1.4 Implement cosmological polytopes: polytopes encoding wavefunction coefficients of the universe in FRW spacetimes, cosmological optical theorem, singularity structure of cosmological correlators from facet geometry — implementing Arkani-Hamed-Benincasa-Postnikov's cosmological polytope program
- 109.3.1.5 Implement positroid stratification: positroid varieties in the Grassmannian Gr(k,n), plabic graphs (bicolored planar graphs), perfect orientations, boundary measurement map — the combinatorial infrastructure underlying the amplituhedron
109.4 Resurgence & Trans-Series#
- 109.4.1.1 Implement Borel summation: given a divergent asymptotic series (coefficients growing as n!), compute the Borel transform (term-by-term division by n!), analytically continue the Borel transform (Padé approximant or conformal mapping), and compute the Laplace integral — extracting finite, well-defined values from divergent perturbation series
- 109.4.1.2 Implement Borel-Padé analysis: diagonal/near-diagonal Padé approximants of the Borel transform, identification of Borel plane singularities (locations = non-perturbative actions, residues = Stokes constants), and Stokes automorphism computation — matching the techniques used by Aniceto-Schiappa-Vonk
- 109.4.1.3 Implement trans-series computation: formal expansion in powers of coupling g AND in exp(-A/g) with possible log(g) factors, trans-series parameter (Stokes multiplier) determination from Borel plane analysis, lateral Borel resummation (above/below Stokes lines), median resummation — the full resurgent analysis pipeline
- 109.4.1.4 Implement alien calculus (Écalle's formalism): alien derivative Δ_ω mapping between trans-series sectors, bridge equations connecting perturbative and non-perturbative data, resurgence relations (perturbative coefficients encode non-perturbative information at all orders) — for extracting instanton/renormalon physics from perturbation theory
- 109.4.1.5 Implement high-order perturbative coefficient computation: using Kalika's CAS engine, compute perturbative coefficients to hundreds of terms for standard QFT and QM models (anharmonic oscillator, Mathieu equation, double-well, phi^4 theory) — the raw data needed for resurgent analysis, requiring arbitrary-precision arithmetic
109.5 Conformal & S-Matrix Bootstrap#
- 109.5.1.1 Implement conformal block computation: scalar conformal blocks in d=2,3,4 dimensions (Zamolodchikov recursion for d=2, Dolan-Osborn for d=4, recursion relations for general d), spinning conformal blocks (seed blocks + differential operators), and conformal Casimir differential equations — the building blocks of the bootstrap
- 109.5.1.2 Implement crossing symmetry equations: from a four-point correlator <O_1 O_2 O_3 O_4>, derive the crossing equations (s-channel expansion = t-channel expansion) as a linear functional equation constraining OPE coefficients and operator dimensions
- 109.5.1.3 Implement semidefinite programming (SDP) for bootstrap: formulate crossing symmetry + unitarity as a semidefinite feasibility/optimization problem, implement or interface with SDPB (Simmons-Duffin's arbitrary-precision SDP solver) — the core numerical engine of the conformal bootstrap that produced world-record 3D Ising critical exponents
- 109.5.1.4 Implement navigator function method (Reehorst-Rychkov): replace the binary feasibility search with smooth optimization over CFT data space, accelerating the bootstrap by orders of magnitude for multi-correlator systems — the 2024-era improvement to the bootstrap
- 109.5.1.5 Implement S-matrix bootstrap: positivity constraints on scattering amplitudes (forward-limit dispersion relations + crossing symmetry), EFThedron bounds on Wilson coefficients of effective field theories — the amplitude-space analogue of the conformal bootstrap
- 109.5.1.6 Implement analytic bootstrap: Lorentzian inversion formula (Caron-Huot) for extracting OPE data from the double-discontinuity of correlators, large-spin expansion, and light-cone bootstrap — complementing the numerical bootstrap with analytical control
109.6 Applied Category Theory in Physics#
- 109.6.1.1 Implement ZX-calculus engine: ZX-diagrams (green/red spiders with phases), rewrite rules (spider fusion, color change, Hadamard, bialgebra), circuit-to-ZX and ZX-to-circuit conversion, T-count optimization (reducing the number of expensive T gates in quantum circuits) — matching PyZX capabilities but integrated into Kalika's quantum computing library
- 109.6.1.2 Implement string diagram engine: monoidal category string diagrams as a computational data structure, composition (vertical stacking), tensor product (horizontal juxtaposition), braiding, cups/caps (for rigid categories), and diagram rewriting — matching DisCoPy but in TypeScript with Kalika's renderer for visualization
- 109.6.1.3 Implement TQFT computation: Reshetikhin-Turaev invariants from modular tensor categories, Turaev-Viro state sum models, knot/link invariants from ribbon categories, and Dijkgraaf-Witten theory for finite gauge groups — connecting category theory to physically meaningful topological invariants
- 109.6.1.4 Implement compositional system modeling (matching Catlab.jl): represent physical systems as morphisms in monoidal categories, compose systems via categorical operations, and automatically derive system dynamics from compositional structure — for modeling coupled physical systems (circuits, mechanical systems, thermodynamic processes) with guaranteed compositionality
109.7 Differentiable Physics#
- 109.7.1.1 Implement differentiable CAS: make Kalika's symbolic computation pipeline end-to-end differentiable — given a symbolic expression with parameters, compute gradients of the evaluated result with respect to parameters via automatic differentiation through the symbolic computation graph (not numerical finite differences) — enabling gradient-based optimization of symbolic models
- 109.7.1.2 Implement FiniteFlow-style dataflow evaluation: define symbolic computation graphs (IBP reduction, tensor contraction, polynomial algebra) as dataflow programs, evaluate numerically over finite fields (mod large primes), reconstruct rational results via Chinese Remainder Theorem + Newton interpolation — the breakthrough technique enabling previously impossible multi-loop QFT calculations (matching Peraro's FiniteFlow)
- 109.7.1.3 Implement differentiable physics simulation: end-to-end differentiable N-body simulator, differentiable fluid dynamics (lattice Boltzmann), and differentiable quantum mechanics (variational quantum eigensolver with gradient-based optimization) — using Kalika's numerical engine with automatic differentiation, enabling gradient-based discovery of physical laws from simulation data
- 109.7.1.4 Implement neural operator training pipeline: infrastructure for training FNO (Fourier Neural Operator) and DeepONet models on PDE solution data generated by Kalika's spectral/FEM solvers — creating learned surrogate models that are 1000x faster than direct numerical solution for parametric PDE families
109.8 Quantum Simulation on Real Hardware#
- 109.8.1.1 Implement quantum circuit generation for physics simulation: given a Hamiltonian (from @kalika/quantum-mechanics or @kalika/lattice), generate quantum circuits implementing time evolution via Trotter-Suzuki decomposition, variational ansätze (UCCSD, ADAPT-VQE), and qubitization — targeting NISQ and early fault-tolerant hardware
- 109.8.1.2 Implement quantum hardware interface: submit quantum circuits to real quantum computers via Qiskit (IBM), Cirq (Google), or TKET (Quantinuum) APIs, with error mitigation (zero-noise extrapolation, probabilistic error cancellation, symmetry verification) — enabling Kalika users to run physics simulations on actual quantum processors
- 109.8.1.3 Implement hybrid quantum-classical optimization: variational algorithms (VQE, QAOA) with classical optimizer (L-BFGS, COBYLA, Adam) driving the quantum circuit parameter updates, with Kalika's classical engine handling the optimization loop and the quantum hardware handling the expectation value estimation
109.9 Machine Learning for Algebraic Geometry#
- 109.9.1.1 Implement ML-based Hodge number prediction: train neural networks on the Kreuzer-Skarke database (polytope weight matrix → Hodge numbers h^{1,1}, h^{2,1}) achieving >99% accuracy — matching He-Lukas work, providing fast approximate Hodge number computation for landscape scans without expensive algebraic computation
- 109.9.1.2 Implement neural Calabi-Yau metric approximation (matching cymetric): train neural networks to approximate Ricci-flat metrics on Calabi-Yau manifolds by minimizing the Monge-Ampère residual — solving the problem of numerically computing CY metrics where no analytic solution exists. Use Kalika's ML training infrastructure (Phase 108)
- 109.9.1.3 Implement RL-based string landscape navigation: reinforcement learning agent that explores the string landscape (flux vacua on Calabi-Yau compactifications), learning to navigate toward vacua with Standard Model-like properties (correct gauge group, three generations, small cosmological constant) — matching Ruehle's RL landscape exploration
109.10 Topological Data Analysis for Physics#
- 109.10.1.1 Implement persistent homology computation: Vietoris-Rips complex construction (from point cloud or distance matrix), boundary matrix reduction (standard algorithm), persistence diagram and barcode computation, Wasserstein/bottleneck distances — using Ripser-style algorithms for performance
- 109.10.1.2 Implement TDA for phase transitions: given lattice model configurations (from @kalika/lattice Monte Carlo), compute persistent homology of the configuration as a function of temperature/coupling, detect topological phase transitions (Betti number discontinuities, persistence diagram bifurcations) — matching Donato-Giri et al.'s approach
- 109.10.1.3 Implement TDA for cosmic large-scale structure: given a point cloud of galaxy positions (or N-body simulation output), compute persistent homology to identify filaments (H_1), voids (H_2), and their scale distribution — matching Xu-Cisewski-Kehe et al.'s approach for cosmological applications
109.11 Tests#
- 109.11.1.1 Write tropical geometry tests: verify tropical polynomial roots, tropical Grassmannian Trop(Gr(2,n)) structure for n=4,5,6,7, and tropical sampling convergence for known Feynman integrals (massless bubble, triangle) — minimum 30 test cases
- 109.11.1.2 Write MZV/polylogarithm tests: verify known MZV values (zeta(2)=pi^2/6, zeta(3)=Apery's constant, zeta(2,1)=zeta(3)), shuffle/stuffle relations, and Feynman integral evaluation against published results (Broadhurst's datasets) — minimum 50 test cases
- 109.11.1.3 Write conformal bootstrap tests: reproduce published OPE bounds for 3D Ising model (Delta_sigma, Delta_epsilon, central charge C_T) to published precision using SDPB — minimum 10 test cases against Simmons-Duffin et al.'s landmark results
- 109.11.1.4 Write resurgence tests: verify Borel summation of known series (Euler series, Airy function asymptotics, anharmonic oscillator), check Stokes constants against published values — minimum 20 test cases
- 109.11.1.5 Write differentiable physics tests: verify gradients of CAS computations against finite difference approximations, verify FiniteFlow reconstruction against exact symbolic computation for IBP test cases — minimum 30 test cases