Searcharxiv⌕ Search

arXiv subjects

Yaqian Tang

Publications and source records attributed to Yaqian Tang.

4 recordsLinked to original sources

Coordinate-Free Automatic Differentiation

Automatic differentiation (AD) is a cornerstone of modern machine learning. Since all AD frameworks implicitly fix a basis of Euclidean space to define gradients and Jacobians, they natively process mathematical objects that are arrays of a fixed set of coordinates. This makes it awkward to express objects such as rotations, volume forms, and curvatures that arise, for example, in applications of AI to science. Bespoke workarounds lack reusability. We show that (i) a single, basis-independent derivative operator can represent many geometric differential operations, and (ii) this operator can be expressed entirely in terms of existing AD primitives. This reusable transformation can be composed with JAX batching, compilation, and hardware acceleration. Geometry enters as algebraic data. We experimentally demonstrate its reuse across different manifold operators, its composition through a spacetime inverse-problem pipeline, and its use in optimization over different rotor subgroups.

cs.LG↗

Towards a Broad-Band Coherent Perfect Absorption in systems without Scale-Invariance

We experimentally and theoretically challenge the concept of coherent perfect absorption (CPA) as a narrow frequency resonant mechanism associated with scattering processes that respect scale-invariance. Using a microwave platform, consisting of a lossy nonlinear resonator coupled to two interrogating antennas, we show that a coherent incident excitation can trigger a self-induced perfect absorption once its intensity exceeds a critical value. Importantly, a (near) perfect absorption persists for a broad band frequency range around the nonlinear CPA condition. Its origin is traced to a quartic behavior that the absorbance spectrum acquires in the proximity of a CPA associated with a new kind of exceptional point degeneracy related to the zeros of the nonlinear scattering operator.

physics.app-ph↗

Tverberg's theorem, disks, and Hamiltonian cycles

For a finite set $S$ of points in the plane and a graph with vertices on $S$ consider the disks with diameters induced by the edges. We show that for any odd set $S$ there exists a Hamiltonian cycle for which these disks share a point, and for an even set $S$ there exists a Hamiltonian path with the same property. We discuss high-dimensional versions of these theorems and their relation to other results in discrete geometry.

math.CO↗