Publications (20)
Saturation for product systems of polynomials
Qinqi Wu, Jiaqi Yu
Recently, Qiu, Xu, Ye and Yu proved that for product system of finitely many minimal systems, the maximal -step pro-nilfactor of the system is the topological characteristi…
Fractal Turán-Nazarov Inequality and Observability for Schrödinger Equations
Jiaqi Yu, Shanlin Huang
This paper establishes limitations on observability inequality and unique continuation for Schrödinger equations on fractal sets. We prove that, in contrast to the heat equation,…
A refined saturation theorem for polynomials and applications
Xiangdong Ye, Jiaqi Yu
For a dynamical system , and distinct non-constant integral polynomials vanishing at , the notion of regionally proximal relation along…
Pixel-scale NIR-VIS Spectral Routers Based on 2D Mie-type Metagratings
Yifan Shao, Shuhan Guo, Rui Chen +7
The out-of-band energy loss caused by in-built color filters significantly degrades the signal-to-noise ratio and the dynamic range of conventional image sensors, which has restric…
DREAM: Domain-free Reverse Engineering Attributes of Black-box Model
Rongqing Li, Jiaqi Yu, Changsheng Li +3
Deep learning models are usually black boxes when deployed on machine learning platforms. Prior works have shown that the attributes (, the number of convolutional layers) of…
Application of uncertainty principles for decaying densities to the observability of the Schrödinger equation
Kévin Le Balc'h, Jiaqi Yu
In this article, we study the Schrödinger equation posed in the Euclidean space. We prove observability inequalities for measurable sets that are thick with respect to decaying de…
Towards End-to-End Alignment of User Satisfaction via Questionnaire in Video Recommendation
Na Li, Jiaqi Yu, Minzhi Xie +8
Short-video recommender systems typically optimize ranking models using dense user behavioral signals, such as clicks and watch time. However, these signals are only indirect proxi…
Compact Hadamard Latent Codes for Efficient Spectral Rendering
Jiaqi Yu, Dar'ya Guarnera, Giuseppe Claudio Guarnera
Spectral rendering accurately reproduces wavelength-dependent appearance but is computationally expensive, as shading must be evaluated at many wavelength samples and scales roughl…
TAME: Test-Time Adversarial Prompt Tuning via Mixture-of-Experts for Vision-Language Models
Xin Wang, Yixu Wang, Jiaming Zhang +6
Large-scale pre-trained Vision-Language models (VLMs), such as CLIP, exhibit strong zero-shot generalization, yet remain highly vulnerable to imperceptible adversarial perturbation…
Uncertainty as Remedy: Mitigating Satisfaction Label Bias in Short Video Multi-Objective Ensemble Ranking
Zonghe Shao, Tiantian He, Xiaoxiao Xu +6
The core objective of short video recommendation is to model users' unobservable true satisfaction with recommended videos. As the dominant industrial framework, end-to-end multi-o…
On the observability of the Schrödinger equation in the torus from open sets
Kévin Le Balc'h, Jiaqi Yu
We study the observability of the Schrödinger equation on the -dimensional torus , , from an open subset . Our first main result e…
SentGuard: Sentence-Level Streaming Guardrails for Large Language Models
Jiaqi Yu, Xin Wang, Yixu Wang +4
Large language models increasingly stream long, reasoning-intensive responses in real time, making when to moderate as critical as whether to moderate. Existing guardrails fall int…
LDDR: Linear-DPP-Based Dynamic-Resolution Frame Sampling for Video MLLMs
Jingfeng Chen, Jiawen Qian, Wendi Deng +5
Video understanding in multimodal large language models requires selecting informative frames from long, redundant videos under limited visual-token budgets. Existing methods often…
EPAR: Electromagnetic Pathways to Architectural Reliability in Quantum Processors
Navnil Choudhury, Yizhuo Tan, Jiaqi Yu +2
As superconducting processors scale, understanding how physical layout shapes qubit interactions is essential for architectural reliability. Existing methods offer limited insight…
Saturated theorem along cubes for a measure and applications
Jiahao Qiu, Jiaqi Yu
We show that for a minimal system , the set of saturated points along cubes with respect to its maximal -step pro-nilfactor has a full measure. As an appl…
Offline-Online Curriculum RL for Multimodal Reasoning
Wendi Deng, Hang Du, Guoshun Nan +11
Multimodal large language models exhibit capabilities on reasoning tasks, yet often produce flawed intermediate steps while yielding correct final answers. This behavior undermines…
S2Aligner: Pair-Efficient and Transferable Pre-Training for Sparse Text-Attributed Graphs
Yuhan Wang, Haopeng Zhang, Yibo Ding +6
Pre-training on text-attributed graphs (TAGs) is central to building transferable graph foundation models, where LLM-as-Aligner methods align graph and text representations through…
Heisenberg Uniqueness Pairs and the wave equation
Shanlin Huang, Jiaqi Yu
Given a curve and a set in the plane, the concept of the Heisenberg uniqueness pair was first introduced by Hedenmalm and Motes-Rodr\'ıgez (Ann. of Math. 173(…
An End-to-End Multi-objective Ensemble Ranking Framework for Video Recommendation
Tiantian He, Minzhi Xie, Runtong Li +6
We propose a novel End-to-end Multi-objective Ensemble Ranking framework (EMER) for the multi-objective ensemble ranking module, which is the most critical component of the short v…
DREAM: Domain-agnostic Reverse Engineering Attributes of Black-box Model
Rongqing Li, Jiaqi Yu, Changsheng Li +3
Deep learning models are usually black boxes when deployed on machine learning platforms. Prior works have shown that the attributes (e.g., the number of convolutional layers) of a…