arXiv subjects

Haotian Jiang

Publications and source records attributed to Haotian Jiang.

At least 19 recordsLinked to original sources

arXiv2026

Rank-One Matrix Discrepancy and Algorithmic Kadison--Singer

We give a deterministic polynomial-time algorithm that, given rational Hermitian matrices $H_1,\dots,H_N$ of rank at most one, finds signs $s\in\{\pm1\}^N$ with $\|\sum_i s_i H_i\|\le 13\|\sum_i H_i^2\|^{1/2}$. As a corollary, for vectors $v_i$ with $\sum_i v_iv_i^*=I$ and $\|v_i\|^2\leδ$, the signs yield a partition $[N] = S_1 \cup S_2$ such that each part satisfies $\|\sum_{i \in S_j} v_i v_i^* - \frac{I}{2}\| \leq \frac{13}{2}\sqrtδ$ for $j = 1,2$. This gives a deterministic polynomial-time algorithm for the Kadison--Singer problem, in Weaver's equivalent discrepancy-theoretic $\mathsf{KS}_2$ formulation, with a universal constant.

cs.DS↗

arXiv2026

Quasi-Monte Carlo Beyond Hardy-Krause II: $(1 + \varepsilon)n$ Samples Suffice

Numerical integration studies how well one can estimate the integral of a function $f$ over $[0,1)^d$ using $n$ sample points. The two classical methods, Monte Carlo (MC) and quasi-Monte Carlo (QMC), have complementary strengths and weaknesses, and a fundamental question is to design an approach that combines the benefits of both. Recently, building on the transference principle in discrepancy theory, Bansal and Jiang~\cite{BJ25a} gave a randomized QMC method that bridges MC and QMC guarantees using only i.i.d.\ samples. Their method also goes beyond the classical Koksma--Hlawka inequality: it achieves integration error $\widetilde{O}_d(σ_{\mathsf{SO}}(f)/n)$, where the smoothed-out variation $σ_{\mathsf{SO}}(f)$ can be substantially smaller than the Hardy--Krause variation that governs the classical bound. However, their algorithm requires $n^2$ i.i.d.\ samples as input, and this quadratic blowup is inherent to any method based on the transference principle. In this work, we bypass the quadratic blowup: for any constant $\varepsilon > 0$, we show that $(1+\varepsilon)n$ i.i.d.\ samples suffice to both obtain the beyond-Hardy--Krause guarantee of~\cite{BJ25a}, resolving an open problem posed there, and to produce low-discrepancy point sequences. Our algorithms are variants of the online Haar-thinning method of Dwivedi, Feldheim, Gurel-Gurevich, and Ramdas~\cite{DFG+19}.

cs.DS↗

arXiv2026

Performance-Efficiency Tradeoffs in Transformers: An Approximation Theory Perspective

Transformers have achieved remarkable successes across a wide range of applications, yet the theoretical foundation of their model efficiency remains underexplored. In this work, we investigate how the model parameters -- mainly attention heads and head dimensions -- should be allocated across layers to balance expressivity and efficiency. We first provide mathematical analysis on the role of early layers in information extraction from an approximation perspective, with a theoretical characterization on the trade-off between the number of heads and head dimension under a fixed parameter budget. In addition, we uncover and prove the \emph{saturation} behavior of softmax activations: Continuously increasing head dimensions can lead to diminishing returns in learning errors, particularly for long sequences. Supported by both theory and experiments, this saturation pattern suggests that later layers can operate more efficiently with reduced parameters. Combining these insights, we propose principled strategies for allocating attention heads and dimensions across Transformers' layers, shedding light on theoretically-grounded model efficiency of Transformer-based architectures.

cs.LG↗

arXiv2026

A Simpler Analysis of the Bansal-Jiang Quasi Monte-Carlo Algorithm via Haar Wavelets

Numerical integration---approximating the integral of a function $f$ using $n$ point evaluations---is a central task in science and engineering. The two main paradigms for this problem, the Monte Carlo and quasi-Monte Carlo methods, have distinct strengths and limitations, and a fundamental question is to design a method that combines the benefits of both. \smallskip Building on recent algorithmic advances in discrepancy theory, Bansal and Jiang \cite{BJ25a} gave a randomized QMC method that naturally bridges the MC and QMC error guarantees. Their method also achieves a surprising improvement over the classical Koksma--Hlawka inequality for QMC methods: it attains an error bound of $\widetilde{O}(σ_{\mathsf{SO}}(f)/n)$, where $σ_{\mathsf{SO}}(f)$ is a new notion of \emph{smoothed-out variation} that they introduced and showed to be substantially smaller than the Hardy--Krause variation governing the classical bound. \smallskip However, the analysis in \cite{BJ25a} is quite involved: it must carefully exploit the structure of the dyadic decomposition and the randomness of the algorithm inside a sufficiently fine discretization of the Hlawka--Zaremba formula to obtain cancellations among the high-frequency components in the Fourier decomposition of $f$. The contribution of this article is twofold: (1) We give an equivalent characterization of $σ_{\mathsf{SO}}(f)$ in terms of the Haar--Besov seminorm of $f$, relating this new notion of smoothed-out variation to classical quantities. (2) Through this characterization, we provide a conceptually simpler and more direct analysis of the Bansal--Jiang QMC method via Haar decomposition, bypassing the use of the Hlawka--Zaremba formula, Fourier decomposition, and the delicate cancellation arguments of \cite{BJ25a} that heavily exploit the structure of dyadic decomposition.

cs.DS↗

arXiv2026

An Exposition of the $\widetilde{O}(\log^{1/4} n)$ Bound for the Komlós Problem

A conjecture of Komlós states that the combinatorial discrepancy of any matrix $A\in\mathbb R^{m\times n}$ whose columns have Euclidean norm at most one is bounded by a universal constant. We prove that the combinatorial discrepancy of every such matrix is at most $O((\log n)^{1/4}(\log\log n)^{7/4})$. This is the first asymptotic improvement over the $O(\sqrt{\log n})$ bound established by Banaszczyk [Banaszczyk, Random Struct.\ Algorithms, 1998], and it refutes a conjecture of Hajela [Hajela, European J.\ Combin., 1988] that a lower bound of order $Ω(\sqrt{\log n})$ should hold.

math.CO↗

arXiv2026

Numerical Investigation of Sequence Modeling Theory using Controllable Memory Functions

The evolution of sequence modeling architectures, from recurrent neural networks and convolutional models to Transformers and structured state-space models, reflects ongoing efforts to address the diverse temporal dependencies inherent in sequential data. Despite this progress, systematically characterizing the strengths and limitations of these architectures remains a fundamental challenge. In this work, we propose a synthetic benchmarking framework to evaluate how effectively different sequence models capture distinct temporal structures. The core of this approach is to generate synthetic targets, each characterized by a parametric memory function $ρ(s, α)$ and a controllable parameter $α$ that determines the temporal strength. This setup allows us to produce a continuum of tasks that vary in temporal complexity, enabling fine-grained analysis of model behavior with respect to specific memory properties. We focus on four representative memory functions, each corresponding to a distinct class of temporal structures: exponential and polynomial functions for decay dynamics, impulse functions for long-range dependencies, and Airy functions for sparsity patterns. Experiments on several sequence modeling architectures confirm existing theoretical insights and reveal new findings regarding approximation capabilities, optimization dynamics, and architectural trade-offs. These results demonstrate the effectiveness of the proposed method in advancing theoretical understanding and highlight the importance of using controllable targets with clearly defined structures for evaluating sequence modeling architectures.

cs.LG↗

arXiv2026

InfoFlow: A Framework for Multi-Layer Transformer Analysis

While the approximation properties of single-layer Transformer architectures have been studied in recent works, a rigorous theoretical understanding of the multi-layer setting remains limited. In this work, we establish that multi-layer Transformers possess fundamentally different approximation capabilities from single-layer ones: for certain retrieval tasks, any single-layer Transformer requires least $Ω(\varepsilon^{-k})$ parameters to achieve precision $\varepsilon$, where $k$ grows linearly with sequence length $T$, whereas a two-layer Transformer with a single head per layer achieves the same approximation precision with at most $O (\varepsilon^{-1})$ parameters. To understand this separation, we identify two structural mechanisms underlying multi-layer approximation. Specifically, softmax attention can only efficiently retrieve the token attaining the maximum attention score, incurring exponential-in-length parameter cost for $k$-th largest retrieval with $k \geq 2$. Moreover, the parameter cost of decoding coupled information scales with the size of the retrieved token set. Motivated by these findings, we propose InfoFlow, a framework for multi-layer Transformers. The framework tracks an information set of accessible input positions at each token and layer, assigning an explicit approximation rate to each mode of information propagation. This abstraction recovers known approximation bounds, remains consistent with experimental observations on trained networks, and yields concrete predictions in settings where direct theoretical analysis is currently intractable. Our results provide a principled framework for reasoning about the approximation efficiency of multi-layer Transformers.

cs.LG↗

arXiv2026

Near-Optimal Constructive Bounds for $\ell_2$ Prefix Discrepancy and Steinitz Problems via Affine Spectral Independence

A classical result of Steinitz from 1913 \cite{Ste13}, answering an earlier question of Riemann and Lévy (e.g., \cite{Lev05}), states that for any norm $\|\cdot\|$ in $\mathbb{R}^d$ and any set of vectors $v_1, \cdots, v_n \in \R^d$ satisfying $\sum_{i=1}^n v_i = 0$, there exists an ordering $π: [n] \rightarrow [n]$ such that every partial sum along this order is bounded by $O(d)$, i.e., $\big\| \sum_{i=1}^t v_{π(i)} \big\| \leq O(d)$ for all $t \in [n]$. Steinitz's bound is tight up to constants in general, but for the $\ell_2$ norm $\|\cdot\|_2$, it has been conjectured that the best bound is $O(\sqrt{d})$. Almost a century later, a breakthrough work of Banaszczyk \cite{Ban12} gave a bound of $O(\sqrt{d} + \sqrt{\log n})$ for the $\ell_2$ Steinitz problem, matching the conjecture under the mild assumption that $d \geq Ω(\log n)$. Banaszczyk's result is non-constructive, and the previous best algorithmic bound was $O(\sqrt{d \log n})$, due to Bansal and Garg \cite{BG17}. In this work, we give an efficient algorithm that matches the conjectured $O(\sqrt{d})$ bound for the $\ell_2$ Steinitz problem under the slightly worse, yet still polylogarithmic, condition of $d \geq Ω(\log^7 n)$. As in prior work, our result extends to the harder problem of $\ell_2$ prefix discrepancy. We employ the framework of obtaining the desired ordering via a discrete Brownian motion, guided by a semidefinite program (SDP). To obtain our results, we use the new technique of ``Decoupling via Affine Spectral Independence'', proposed by Bansal and Jiang \cite{BJ26} to achieve substantial progress on the Beck-Fiala and Komlós conjectures, together with a ``Global Interval Tree'' data structure that simultaneously controls the deviations for all prefixes.

cs.DS↗

arXiv2026

The Effect of Attention Head Count on Transformer Approximation

Transformer has become the dominant architecture for sequence modeling, yet a detailed understanding of how its structural parameters influence expressive power remains limited. In this work, we study the approximation properties of transformers, with particular emphasis on the role of the number of attention heads. Our analysis begins with the introduction of a generalized $D$-retrieval task, which we prove to be dense in the space of continuous functions, thereby providing the basis for our theoretical framework. We then establish both upper and lower bounds on the parameter complexity required for $ε$-approximation. Specifically, we show that transformers with sufficiently many heads admit efficient approximation, whereas with too few heads, the number of parameters must scale at least as $O(1/ε^{cT})$, for some constant $c$ and sequence length $T$. To the best of our knowledge, this constitutes the first rigorous lower bound of this type in a nonlinear and practically relevant setting. We further examine the single-head case and demonstrate that an embedding dimension of order $O(T)$ allows complete memorization of the input, where approximation is entirely achieved by the feed-forward block. Finally, we validate our theoretical findings with experiments on both synthetic data and real-world tasks, illustrating the practical relevance of our results.

cs.LG↗

arXiv2026

PPGL-Swarm: Integrated Multimodal Risk Stratification and Hereditary Syndrome Detection in Pheochromocytoma and Paraganglioma

Pheochromocytomas and paragangliomas (PPGLs) are rare neuroendocrine tumors, of which 15-25% develop metastatic disease with 5-year survival rates reported as low as 34%. PPGL may indicate hereditary syndromes requiring stricter, syndrome-specific treatment and surveillance, but clinicians often fail to recognize these associations in routine care. Clinical practice uses GAPP score for PPGL grading, but several limitations remain for PPGL diagnosis: (1) GAPP scoring demands a high workload for clinician because it requires the manual evaluation of six independent components; (2) key components such as cellularity and Ki-67 are often evaluated with subjective criteria; (3) several clinically relevant metastatic risk factors are not captured by GAPP, such as SDHB mutations, which have been associated with reported metastatic rates of 35-75%. Agent-driven diagnostic systems appear promising, but most lack traceable reasoning for decision-making and do not incorporate domain-specific knowledge such as PPGL genotype information. To address these limitations, we present PPGL-Swarm, an agentic PPGL diagnostic system that generates a comprehensive report, including automated GAPP scoring (with quantified cellularity and Ki-67), genotype risk alerts, and multimodal report with integrated evidence. The system provides an auditable reasoning trail by decomposing diagnosis into micro-tasks, each assigned to a specialized agent. The gene and table agents use knowledge enhancement to better interpret genotype and laboratory findings, and during training we use reinforcement learning to refine tool selection and task assignment.

cs.CV↗

arXiv2026

Automatic Funny Scene Extraction from Long-form Cinematic Videos

Automatically extracting engaging and high-quality humorous scenes from cinematic titles is pivotal for creating captivating video previews and snackable content, boosting user engagement on streaming platforms. Long-form cinematic titles, with their extended duration and complex narratives, challenge scene localization, while humor's reliance on diverse modalities and its nuanced style add further complexity. This paper introduces an end-to-end system for automatically identifying and ranking humorous scenes from long-form cinematic titles, featuring shot detection, multimodal scene localization, and humor tagging optimized for cinematic content. Key innovations include a novel scene segmentation approach combining visual and textual cues, improved shot representations via guided triplet mining, and a multimodal humor tagging framework leveraging both audio and text. Our system achieves an 18.3% AP improvement over state-of-the-art scene detection on the OVSD dataset and an F1 score of 0.834 for detecting humor in long text. Extensive evaluations across five cinematic titles demonstrate 87% of clips extracted by our pipeline are intended to be funny, while 98% of scenes are accurately localized. With successful generalization to trailers, these results showcase the pipeline's potential to enhance content creation workflows, improve user engagement, and streamline snackable content generation for diverse cinematic media formats.

cs.IR↗

arXiv2025

Sliding multiferroicity in hexagonal stacked CrI3

Developing new multiferroics at the two-dimensional (2D) limit with energy-efficient magnetoelectric coupling can inform the interplay physics of novel orders and advance on-chip high-performance computing applications. Here we apply stacking order engineering to create a new type of 2D multiferroics, namely sliding multiferroics, based on polar hexagonal stacked (H-stacked) CrI3. This new stacking order removes structural inversion symmetry and gives rise to room temperature sliding ferroelectricity, as confirmed by Raman spectroscopy, second harmonic generation spectroscopy and electrical transport measurements. Building upon the gate-dependent reflective magnetic circular dichroism, first-principles calculations, and modeling, sliding ferroelectricity is shown to interplay with an emergent interfacial ferromagnetism via interlayer spin-polarized charge transfer. This coupling mechanism results in non-volatile magnetic switching by as low as 0.4V across the H-stacked CrI3. Our demonstration introduces polar stacking order engineering of 2D magnets as a general approach to create non-volatile 2D multiferroics with efficient magnetoelectric coupling, paving the way for low-power electronics and spintronics at the atomically thin limit.

cond-mat.mtrl-sci↗

arXiv2025

Decoupling via Affine Spectral-Independence: Beck-Fiala and Komlós Bounds Beyond Banaszczyk

The Beck-Fiala Conjecture [Discrete Appl. Math, 1981] asserts that any set system of $n$ elements with degree $k$ has combinatorial discrepancy $O(\sqrt{k})$. A substantial generalization is the Komlós Conjecture, which states that any $m \times n$ matrix with unit length columns has discrepancy $O(1)$. In this work, we resolve the Beck-Fiala Conjecture for $k \geq \log^2 n$. We also give an $\widetilde{O}(\sqrt{k} + \sqrt{\log n})$ bound for $k \leq \log^2 n$, where $\widetilde{O}(\cdot)$ hides $\mathsf{poly}(\log \log n)$ factors. These bounds improve upon the $O(\sqrt{k \log n})$ bound due to Banaszczyk [Random Struct. Algor., 1998]. For the Komlos problem, we give an $\widetilde{O}(\log^{1/4} n)$ bound, improving upon the previous $O(\sqrt{\log n})$ bound [Random Struct. Algor., 1998]. All of our results also admit efficient polynomial-time algorithms. To obtain these results, we exploit a new technique of ``decoupling via affine spectral-independence'' in designing rounding algorithms. In particular, our algorithms obtain the desired colorings via a discrete Brownian motion, guided by a semidefinite program (SDP). Besides standard constraints used in prior works, we add some extra affine spectral-independence constraints, which effectively decouple the evolution of discrepancies across different rows, and allow us to better control how many rows accumulate large discrepancies at any point during the process. This new technique is quite general and may be of independent interest.

math.CO↗

arXiv2025

High-Dimensional Quasi-Monte Carlo via Combinatorial Discrepancy

Monte Carlo (MC) and Quasi-Monte Carlo (QMC) methods are classical approaches for the numerical integration of functions $f$ over $[0,1]^d$. While QMC methods can achieve faster convergence rates than MC in moderate dimensions, their tractability in high dimensions typically relies on additional structure -- such as low effective dimension or carefully chosen coordinate weights -- since worst-case error bounds grow prohibitively large as $d$ increases. In this work, we study the construction of high-dimensional QMC point sets via combinatorial discrepancy, extending the recent QMC method of Bansal and Jiang. We establish error bounds for these constructions in weighted function spaces, and for functions with low effective dimension in both the superposition and truncation sense. We also present numerical experiments to empirically assess the performance of these constructions.

math.NA↗

arXiv2025

An Improved Bound for the Beck-Fiala Conjecture

In 1981, Beck and Fiala [Discrete Appl. Math, 1981] conjectured that given a set system $A \in \{0,1\}^{m \times n}$ with degree at most $k$ (i.e., each column of $A$ has at most $k$ non-zeros), its combinatorial discrepancy $\mathsf{disc}(A) := \min_{x \in \{\pm 1\}^n} \|Ax\|_\infty$ is at most $O(\sqrt{k})$. Previously, the best-known bounds for this conjecture were either $O(k)$, first established by Beck and Fiala [Discrete Appl. Math, 1981], or $O(\sqrt{k \log n})$, first proved by Banaszczyk [Random Struct. Algor., 1998]. We give an algorithmic proof of an improved bound of $O(\sqrt{k \log\log n})$ whenever $k \geq \log^5 n$, thus matching the Beck-Fiala conjecture up to $O(\sqrt{\log \log n})$ for almost the full regime of $k$.

math.CO↗

arXiv2025

The Effect of Depth on the Expressivity of Deep Linear State-Space Models

Deep state-space models (SSMs) have gained increasing popularity in sequence modelling. While there are numerous theoretical investigations of shallow SSMs, how the depth of the SSM affects its expressiveness remains a crucial problem. In this paper, we systematically investigate the role of depth and width in deep linear SSMs, aiming to characterize how they influence the expressive capacity of the architecture. First, we rigorously prove that in the absence of parameter constraints, increasing depth and increasing width are generally equivalent, provided that the parameter count remains within the same order of magnitude. However, under the assumption that the parameter norms are constrained, the effects of depth and width differ significantly. We show that a shallow linear SSM with large parameter norms can be represented by a deep linear SSM with smaller norms using a constructive method. In particular, this demonstrates that deep SSMs are more capable of representing targets with large norms than shallow SSMs under norm constraints. Finally, we derive upper bounds on the minimal depth required for a deep linear SSM to represent a given shallow linear SSM under constrained parameter norms. We also validate our theoretical results with numerical experiments

cs.LG↗

arXiv2025

Pathology Image Restoration via Mixture of Prompts

In digital pathology, acquiring all-in-focus images is essential to high-quality imaging and high-efficient clinical workflow. Traditional scanners achieve this by scanning at multiple focal planes of varying depths and then merging them, which is relatively slow and often struggles with complex tissue defocus. Recent prevailing image restoration technique provides a means to restore high-quality pathology images from scans of single focal planes. However, existing image restoration methods are inadequate, due to intricate defocus patterns in pathology images and their domain-specific semantic complexities. In this work, we devise a two-stage restoration solution cascading a transformer and a diffusion model, to benefit from their powers in preserving image fidelity and perceptual quality, respectively. We particularly propose a novel mixture of prompts for the two-stage solution. Given initial prompt that models defocus in microscopic imaging, we design two prompts that describe the high-level image semantics from pathology foundation model and the fine-grained tissue structures via edge extraction. We demonstrate that, by feeding the prompt mixture to our method, we can restore high-quality pathology images from single-focal-plane scans, implying high potentials of the mixture of prompts to clinical usage. Code will be publicly available at https://github.com/caijd2000/MoP.

cs.CV↗

arXiv2025

LSA: Latent Style Augmentation Towards Stain-Agnostic Cervical Cancer Screening

The deployment of computer-aided diagnosis systems for cervical cancer screening using whole slide images (WSIs) faces critical challenges due to domain shifts caused by staining variations across different scanners and imaging environments. While existing stain augmentation methods improve patch-level robustness, they fail to scale to WSIs due to two key limitations: (1) inconsistent stain patterns when extending patch operations to gigapixel slides, and (2) prohibitive computational/storage costs from offline processing of augmented WSIs.To address this, we propose Latent Style Augmentation (LSA), a framework that performs efficient, online stain augmentation directly on WSI-level latent features. We first introduce WSAug, a WSI-level stain augmentation method ensuring consistent stain across patches within a WSI. Using offline-augmented WSIs by WSAug, we design and train Stain Transformer, which can simulate targeted style in the latent space, efficiently enhancing the robustness of the WSI-level classifier. We validate our method on a multi-scanner WSI dataset for cervical cancer diagnosis. Despite being trained on data from a single scanner, our approach achieves significant performance improvements on out-of-distribution data from other scanners. Code will be available at https://github.com/caijd2000/LSA.

eess.IV↗