arXiv subjects

Yichao Tang

Publications and source records attributed to Yichao Tang.

At least 19 recordsLinked to original sources

arXiv2026

Bootstrapping Giant Graviton Correlators

We develop bootstrap methods for mixed heavy-light four-point correlators $\langle GGOO\rangle$ in $\mathcal N=4$ super-Yang--Mills theory at large $N$, where $O\equiv {\cal O}_2$ is the chiral primary operator in the stress-tensor multiplet and $G$ are (dual) giant graviton operators with dimension of order $N$, including the maximal determinant case. The loop integrand is expanded in a basis of labelled $f$-graphs -- necessarily including non-planar topologies due to the dimension-$N$ nature of the giant gravitons -- and the coefficients are fixed by various bootstrap conditions including double-triangle and triangle rules in the cusp and OPE limits, integrated correlators from supersymmetric localization, and a ten-dimensional hidden symmetry, the latter also allowing extension to correlators involving generic chiral primaries $\mathcal{O}_k$. Together, these inputs uniquely determine the correlator through three loops, passing further non-trivial consistency checks. For the maximal determinant operator, we reproduce the known results through two loops and obtain the full three-loop correction.

hep-th↗

arXiv2026

MATA-Former & SIICU: Semantic Aware Temporal Alignment for High-Fidelity ICU Risk Prediction

Forecasting evolving clinical risks relies on intrinsic pathological dependencies rather than mere chronological proximity, yet current methods struggle with coarse binary supervision and physical timestamps. To align predictive modeling with clinical logic, we propose the Medical-semantics Aware Time-ALiBi Transformer (MATA-Former), utilizing event semantics to dynamically parameterize attention weights to prioritize causal validity over time lags. Furthermore, we introduce Plateau-Gaussian Soft Labeling (PSL), reformulating binary classification into continuous multi-horizon regression for full-trajectory risk modeling. Evaluated on SIICU -- a newly constructed dataset featuring over 506k events with rigorous expert-verified, fine-grained annotations -- and the MIMIC-IV dataset, our framework demonstrates superior efficacy and robust generalization in capturing risks from text-intensive, irregular clinical time series.

cs.LG↗

arXiv2025

SimuFreeMark: A Noise-Simulation-Free Robust Watermarking Against Image Editing

The advancement of artificial intelligence generated content (AIGC) has created a pressing need for robust image watermarking that can withstand both conventional signal processing and novel semantic editing attacks. Current deep learning-based methods rely on training with hand-crafted noise simulation layers, which inherently limit their generalization to unforeseen distortions. In this work, we propose $\textbf{SimuFreeMark}$, a noise-$\underline{\text{simu}}$lation-$\underline{\text{free}}$ water$\underline{\text{mark}}$ing framework that circumvents this limitation by exploiting the inherent stability of image low-frequency components. We first systematically establish that low-frequency components exhibit significant robustness against a wide range of attacks. Building on this foundation, SimuFreeMark embeds watermarks directly into the deep feature space of the low-frequency components, leveraging a pre-trained variational autoencoder (VAE) to bind the watermark with structurally stable image representations. This design completely eliminates the need for noise simulation during training. Extensive experiments demonstrate that SimuFreeMark outperforms state-of-the-art methods across a wide range of conventional and semantic attacks, while maintaining superior visual quality.

cs.CV↗

arXiv2025

Spinning Mellin amplitudes

We propose a definition of Mellin amplitudes for conformal correlators involving arbitrary spinning operators in tensor representations of the Lorentz group. These representations cover all bosonic local operators. Our strategy is to perform discrete Mellin transforms on all scalar products involving polarization vectors, so that each polarization vector can be interpreted as the position of a fictitious scalar operator. We establish the general pole structures and factorization properties of these spinning Mellin amplitudes. We also provide a systematic algorithm to derive factorization formulas with arbitrary spinning exchanges, yielding new explicit results up to spin-4. To illustrate the practicality of our formalism, we bootstrap the 3- and 4-point current correlators in a 4d $\mathcal{N}=2$ superconformal field theory, which are dual to gluon scattering amplitudes in $\mathrm{AdS}_5 \times \mathrm{S}^3$. The results agree with the snowflake channel of 6- and 8-point scalar supergluon amplitudes in the literature.

hep-th↗

arXiv2025

A Hidden Permutation Symmetry of Squared Amplitudes in ABJM Theory

We define the square amplitudes in planar Aharony-Bergman-Jafferis-Maldacena theory (ABJM), analogous to that in $\mathcal{N}{=}4$ super-Yang-Mills theory (SYM). Surprisingly, the $n$-point $L$-loop integrands with fixed $N{:=}n{+}L$ are unified in a single generating function. Similar to the SYM four-point half-BPS correlator integrand, the generating function enjoys a hidden $S_N$ permutation symmetry in the dual space, allowing us to write it as a linear combination of weight-3 planar $f$-graphs. Remarkably, through Gram identities it can also be represented as a linear combination of bipartite $f$-graphs which manifest the important property that no odd-multiplicity amplitude exists in the theory. The generating function and these properties are explicitly checked against squared amplitudes for all $n$ with $N{=}4,6,8$. By drawing analogies with SYM, we conjecture some graphical rules the generating function satisfy, and exploit them to bootstrap a unique $N{=}10$ result, which provides new results for $n{=}10$ squared tree amplitudes, as well as integrands for $(n,L){=}(4,6),(6,4)$. Our results strongly suggest the existence of a "bipartite correlator" in ABJM theory that unifies all squared amplitudes and satisfies physical constraints underlying these graphical rules.

hep-th↗

arXiv2025

AdS$\times$S Mellin Bootstrap, Hidden 10d Symmetry and Five-point Kaluza-Klein Functions in $\mathcal{N}=4$ SYM

We propose an AdS$\times$S factorization formula at the level of the generating function for correlators with arbitrary Kaluza-Klein configurations, and implement it in the supergravity limit of $\mathcal{N}=4$ super Yang-Mills. By incorporating this mechanism into Mellin space bootstrap, together with an observed $Z_2$ symmetry under AdS$\ \leftrightarrow\ $S, we manage to simultaneously work out unified formulas both for all five-point half-BPS correlators and for all four-point correlators with one superdescendant. This AdS\times$S bootstrap method is directly applicable to generic multi-point computation at tree level.

hep-th↗

arXiv2025

Pose State Perception of Interventional Robot for Cardio-cerebrovascular Procedures

In response to the increasing demand for cardiocerebrovascular interventional surgeries, precise control of interventional robots has become increasingly important. Within these complex vascular scenarios, the accurate and reliable perception of the pose state for interventional robots is particularly crucial. This paper presents a novel vision-based approach without the need of additional sensors or markers. The core of this paper's method consists of a three-part framework: firstly, a dual-head multitask U-Net model for simultaneous vessel segment and interventional robot detection; secondly, an advanced algorithm for skeleton extraction and optimization; and finally, a comprehensive pose state perception system based on geometric features is implemented to accurately identify the robot's pose state and provide strategies for subsequent control. The experimental results demonstrate the proposed method's high reliability and accuracy in trajectory tracking and pose state perception.

cs.RO↗

arXiv2025

GenPTW: Latent Image Watermarking for Provenance Tracing and Tamper Localization

The proliferation of generative image models has revolutionized AIGC creation while amplifying concerns over content provenance and manipulation forensics. Existing methods are typically either unable to localize tampering or restricted to specific generative settings, limiting their practical utility. We propose \textbf{GenPTW}, a \textbf{Gen}eral watermarking framework that unifies \textbf{P}rovenance tracing and \textbf{T}amper localization in latent space. It supports both in-generation and post-generation embedding without altering the generative process, and is plug-and-play compatible with latent diffusion models (LDMs) and visual autoregressive (VAR) models. To achieve precise provenance tracing and tamper localization, we embed the watermark using two complementary mechanisms: cross-attention fusion aligned with latent semantics and spatial fusion providing explicit spatial guidance for edit sensitivity. A tamper-aware extractor jointly conducts provenance tracing and tamper localization by leveraging watermark features together with high-frequency features. Experiments show that GenPTW maintains high visual fidelity and strong robustness against diverse AIGC-editing.

cs.CR↗

arXiv2025

The Four-Point Correlator of Planar sYM at Twelve Loops

We determine the 4-point correlation function and amplitude in planar, maximally supersymmetric Yang-Mills theory to 12 loops. We find that the recently-introduced 'double-triangle' rule in fact implies the previously described square and pentagon rules; and when applied to 12 loops, it fully determines the 11-loop correlator and fixes all but 3 of the (22,024,902) 12-loop coefficients; these remaining coefficients can be subsequently fixed using the '(single-)triangle' rule. Not only do we confirm the Catalan conjecture for anti-prism graphs, but we discover evidence for a greatly generalized Catalan conjecture for the coefficients of all polygon-framed fishnet graphs. We provide all contributions through 12 loops as ancillary files to this work.

hep-th↗

arXiv2024

The Cusp Limit of Correlators and A New Graphical Bootstrap for Correlators/Amplitudes to Eleven Loops

We consider the universal behavior of half-BPS correlators in $\mathcal{N}=4$ super-Yang-Mills in the cusp limit where two consecutive separations $x_{12}^2,x_{23}^2$ become lightlike. Through the Lagrangian insertion procedure, the Sudakov double-logarithmic divergence of the $n$-point correlator is related to the $(n+1)$-point correlator where the inserted Lagrangian "pinches" to the soft-collinear region of the cusp. We formulate this constraint as a new graphical rulefor the $f$-graphs of the four-point correlator, which turns out to be the most constraining rule known so far. By exploiting this single graphical rule, we bootstrap the planar integrand of the four-point correlator up to ten loops ($n=14$) and fix all 22024902 but one coefficient at eleven loops ($n=15$); the remaining coefficient is then fixed using the triangle rule. We verify the "Catalan conjecture" for the coefficients of the family of $f$-graphs known as "anti-prisms" where the coefficient of the twelve-loop ($n=16$) anti-prism is found to be $-42$ by a local analysis of the bootstrap equations. We also comment on the implication of our graphical rule for the non-planar contributions.

hep-th↗

arXiv2024

Supergluon scattering in AdS: constructibility, spinning amplitudes, and new structures

We elaborate on a new recursive method proposed in arXiv:2312.15484 for computing tree-level $n$-point supergluon amplitudes as well as those with one gluon, i.e., spinning amplitudes, in ${\rm AdS}_5 \times S^3$. We present an improved proof for the so-called "constructibility" of supergluon and spinning amplitudes based on their factorizations and flat-space limit, which allows us to determine these amplitudes in Mellin space to all $n$. We present explicit and remarkably simple expressions for up to $n=7$ supergluon amplitudes and $n=6$ spinning amplitudes, which can be viewed as AdS generalizations of the scalar-scaffolded gluon amplitudes proposed recently. We then reveal a series of hidden structures of these AdS amplitudes including (1) an understanding of general pole structures especially the precise truncation on descendent poles (2) a derivation of simple "Feynman rules" for the all-$n$ amplitudes with the simplest R-symmetry structures, and (3) certain universal behavior analogous to the soft/collinear limit of flat-space amplitudes.

hep-th↗

arXiv2023

Constructibility of AdS supergluon amplitudes

We prove that all tree-level $n$-point supergluon (scalar) amplitudes in AdS$_5$ can be recursively constructed, using factorization and flat-space limit. Our method is greatly facilitated by a natural R-symmetry basis for planar color-ordered amplitudes, which reduces the latter to "partial amplitudes" with simpler pole structures and factorization properties. Given the $n$-point scalar amplitude, we first extract spinning amplitudes with $n{-}2$ scalars and one gluon by imposing "gauge invariance", and then use a special "no-gluon kinematics" to determine the $(n{+}1)$-point scalar amplitude completely (which in turn contains the $n$-point single-gluon amplitude). Explicit results of up to 8-point scalar amplitudes and up to 6-point single-gluon amplitudes are included as Supplemental Material.

hep-th↗

arXiv2023

Jumpstarting (elliptic) symbol integrations for loop integrals

We derive an algorithm for computing the total differentials of multi-loop integrals expressed as one-fold integrals of multiple polylogarithms, which can involve square roots of polynomials up to degree four and may evaluate to (elliptic) multiple polylogarithms ((e)MPL). This gives simple algebraic rules for computing the $(W{-}1, 1)$-coproduct of the resulting weight-$W$ functions up to period terms, and iterating it gives the symbol without actually performing any integration. In particular, our algorithm generalizes existing MPL integration rules and sidesteps the complicated rationalization procedure in the presence of square roots. We apply our algorithm to conformal double-$D$-gon integrals in $D$ dimensions with generic kinematics and possibly massive circumferential propagators. We directly compute, for the first time, the total differential and symbol (up to period terms) of the $D{=}3$ double-triangle and the $D{=}4$ double-box, which in the special case with massless propagators represent the first appearance of eMPL functions in (two-loop) scattering amplitudes of ${\cal N}{=}6$ Chern-Simons-matter theory and ${\cal N}{=}4$ super-Yang-Mills, respectively.

hep-th↗

arXiv2023

Cutting the traintracks: Cauchy, Schubert and Calabi-Yau

In this note we revisit the maximal-codimension residues, or leading singularities, of four-dimensional $L$-loop traintrack integrals with massive legs, both in Feynman parameter space and in momentum (twistor) space. We identify a class of "half traintracks" as the most general degenerations of traintracks with conventional (0-form) leading singularities, although the integrals themselves still have rigidity $\lfloor\frac{L-1}2\rfloor$ due to lower-loop "full traintrack'' subtopologies. As a warm-up exercise, we derive closed-form expressions for their leading singularities both via (Cauchy's) residues in Feynman parameters, and more geometrically using the so-called Schubert problems in momentum twistor space. For $L$-loop full traintracks, we compute their leading singularities as integrals of $(L{-}1)$-forms, which proves that the rigidity is $L{-}1$ as expected; the form is given by an inverse square root of an irreducible polynomial quartic with respect to each variable, which characterizes an $(L{-}1)$-dim Calabi-Yau manifold (elliptic curve, K3 surface, etc.) for any $L$. We also briefly comment on the implications for the "symbology" of these traintrack integrals.

hep-th↗

arXiv2022

The symbology of Feynman integrals from twistor geometries

We study the symbology of planar Feynman integrals in dimensional regularization by considering geometric configurations in momentum twistor space corresponding to their leading singularities (LS). Cutting propagators in momentum twistor space amounts to intersecting lines associated with loop and external dual momenta, including the special line associated with the point at infinity, which breaks dual conformal symmetry. We show that cross-ratios of intersection points on these lines, especially those on the infinity line, naturally produce symbol letters for Feynman integrals in $D=4-2ε$, which include and generalize their LS. At one loop, we obtain all symbol letters using intersection points from quadruple cuts for integrals up to pentagon kinematics with two massive corners, which agree perfectly with canonical differential equation (CDE) results. We then obtain all two-loop letters, for up to four-mass box and one-mass pentagon kinematics, by considering more intersections arising from two-loop cuts. Finally we comment on how cluster algebras appear from this construction, and importantly how we may extend the method to non-planar integrals.

hep-th↗

arXiv2021

Bootstrapping octagons in reduced kinematics from $A_2$ cluster algebras

Multi-loop scattering amplitudes/null polygonal Wilson loops in ${\mathcal N}=4$ super-Yang-Mills are known to simplify significantly in reduced kinematics, where external legs/edges lie in an $1+1$ dimensional subspace of Minkowski spacetime (or boundary of the $\rm AdS_3$ subspace). Since the edges of a $2n$-gon with even and odd labels go along two different null directions, the kinematics is reduced to two copies of $G(2,n)/T \sim A_{n{-}3}$. In the simplest octagon case, we conjecture that all loop amplitudes and Feynman integrals are given in terms of two overlapping $A_2$ functions (a special case of two-dimensional harmonic polylogarithms): in addition to the letters $v, 1+v, w, 1+w$ of $A_1 \times A_1$, there are two letters $v-w, 1- v w$ mixing the two sectors but they never appear together in the same term; these are the reduced version of four-mass-box algebraic letters. Evidence supporting our conjecture includes all known octagon amplitudes as well as new computations of multi-loop integrals in reduced kinematics. By leveraging this alphabet and conditions on first and last entries, we initiate a bootstrap program in reduced kinematics: within the remarkably simple space of overlapping $A_2$ functions, we easily obtain octagon amplitudes up to two-loop NMHV and three-loop MHV. We also briefly comment on the generalization to $2n$-gons in terms of $A_2$ functions and beyond.

hep-th↗

arXiv2021

The Wilson-loop $d \log$ representation for Feynman integrals

We introduce and study the Wilson-loop ${\rm d}\log$ representation of certain Feynman integrals for scattering amplitudes in ${\cal N}=4$ SYM and beyond, which makes their evaluation completely straightforward. Such a representation was motivated by the dual Wilson loop picture, and it can also be derived by partial Feynman parametrization of loop integrals. We first introduce it for the simplest one-loop examples, the chiral pentagon in four dimensions and the three-mass-easy hexagon in six dimensions, which are represented by two- and three-fold ${\rm d}\log$ integrals that are nicely related to each other. For multi-loop examples, we write the $L$-loop generalized penta-ladders as $2(L{-}1)$-fold ${\rm d}\log$ integrals of some one-loop integral, so that once the latter is known, the integration can be performed in a systematic way. In particular, we write the eight-point penta-ladder as a $2L$-fold ${\rm d}\log$ integral whose symbol can be computed without performing the integration; we also obtain the last entries and the symbol alphabet of these integrals. Similarly we compute and study the symbol of the seven-point double-penta-ladder, which is represented by a $2(L{-}1)$-fold integral of a hexagon; the latter can be written as a two-fold ${\rm d}\log$ integral plus a boundary term. We comment on the relation of our representation to differential equations and resum the ladders by solving certain integral equations.

hep-th↗

arXiv2020

Leveraging Elastic instabilities for Amplified Performance: spine-inspired high-speed and high-force soft robots

Soft machines typically exhibit slow locomotion speed and low manipulation strength because of intrinsic limitations of soft materials. Here, we present a generic design principle that harnesses mechanical instability for a variety of spine-inspired fast and strong soft machines. Unlike most current soft robots that are designed as inherently and unimodally stable, our design leverages tunable snap-through bistability to fully explore the ability of soft robots to rapidly store and release energy within tens of milliseconds. We demonstrate this generic design principle with three high-performance soft machines: High-speed cheetah-like galloping crawlers with locomotion speeds of 2.68 body length/s, high-speed underwater swimmers (0.78 body length/s), and tunable low-to-high-force soft grippers with over 1 to 103 stiffness modulation (maximum load capacity is 11.4 kg). Our study establishes a new generic design paradigm of next-generation high-performance soft robots that are applicable for multifunctionality, different actuation methods, and materials at multiscales.

physics.app-ph↗