Searcharxiv⌕ Search

arXiv subjects

Jiayin Liu

Publications and source records attributed to Jiayin Liu.

15 recordsLinked to original sources
arXiv2026

Restricted Projections to Hyperplanes in $\mathbb{R}^n$

We study dimensions of sets projected to an $(n-2)$-dimensional family of hyperplanes in $\mathbb{R}^n$ under curvature conditions. Let $n\ge 3$ and $Σ\subset S^{n-1}$ be an $(n-2)$-dimensional $C^2$ manifold such that $Σ$ has non-vanishing geodesic curvature ($n=3$)/sectional curvature $>1$ ($n \ge 4)$. Let $Z \subset \mathbb{R}^{n}$ be analytic with $\dim Z \le n-2$ and $0 < s < \dim Z$. Then \begin{equation*} \dim \{x \in Σ: \dim π_{T_xS^{n-1}}(Z) < s\} \le s \end{equation*} where $π_{T_xS^{n-1}}$ is the orthogonal projection from $\mathbb{R}^n$ to the tangent space $T_xS^{n-1}$. In particular, for $\mathcal{H}^{n-2}$-a.e. $x \in Σ$, $\dim π_{T_xS^{n-1}}(Z) = \dim Z$. When $n=3$ and $\dim Z < 1$, the quantitative estimate improves the one obtained by Gan-Guo-Guth-Harris-Maldague-Wang. For the case $\dim Z > n-2$, if in addition $π_{T_yS^{n-1}}(Z) \le n-2$ for some $y \in S^{n-1}$, we show that $\dim π_{T_xS^{n-1}}(Z) = \min\{\dim Z, n-1\}$ for $\mathcal{H}^{n-2}$-a.e. $x \in Σ$.

math.CA↗
arXiv2026

Physically Plausible Multi-System Trajectory Generation and Symmetry Discovery

From metronomes to celestial bodies, mechanics underpins how the world evolves in time and space. With consideration of this, a number of recent neural network models leverage inductive biases from classical mechanics to encourage model interpretability and ensure forecasted states are physical. However, in general, these models are designed to capture the dynamics of a single system with fixed physical parameters, from state-space measurements of a known configuration space. In this paper we introduce Symplectic Phase Space GAN (SPS-GAN) which can capture the dynamics of multiple systems, and generalize to unseen physical parameters from. Moreover, SPS-GAN does not require prior knowledge of the system configuration space. In fact, SPS-GAN can discover the configuration space structure of the system from arbitrary measurement types (e.g., state-space measurements, video frames). To achieve physically plausible generation, we introduce a novel architecture which embeds a Hamiltonian neural network recurrent module in a conditional GAN backbone. To discover the structure of the configuration space, we optimize the conditional time-series GAN objective with an additional physically motivated term to encourages a sparse representation of the configuration space. We demonstrate the utility of SPS-GAN for trajectory prediction, video generation and symmetry discovery. Our approach captures multiple systems and achieves performance on par with supervised models designed for single systems.

cs.LG↗
arXiv2025

Deep Learning for Double Auction

Auctions are important mechanisms extensively implemented in various markets, e.g., search engines' keyword auctions, antique auctions, etc. Finding an optimal auction mechanism is extremely difficult due to the constraints of imperfect information, incentive compatibility (IC), and individual rationality (IR). In addition to the traditional economic methods, some recently attempted to find the optimal (single) auction using deep learning methods. Unlike those attempts focusing on single auctions, we develop deep learning methods for double auctions, where imperfect information exists on both the demand and supply sides. The previous attempts on single auction cannot directly apply to our contexts and those attempts additionally suffer from limited generalizability, inefficiency in ensuring the constraints, and learning fluctuations. We innovate in designing deep learning models for solving the more complex problem and additionally addressing the previous models' three limitations. Specifically, we achieve generalizability by leveraging a transformer-based architecture to model market participants as sequences for varying market sizes; we utilize the numerical features of the constraints and pre-treat them for a higher learning efficiency; we develop a gradient-conflict-elimination scheme to address the problem of learning fluctuation. Extensive experimental evaluations demonstrate the superiority of our approach to classical and machine learning baselines.

cs.LG↗
arXiv2025

One Dimensional Asymptotic Plateau Problem in $n$-Dimensional Asymptotically Conical Manifolds

Let $(M,g)$ be an asymptotically conical Riemannian manifold having dimension $n\ge 2$, opening angle $α\in (0,π/2) \setminus \{\arcsin \frac{1}{2k+1}\}_{k \in \mathbb{N}}$ and positive asymptotic rate. Under the assumption that the exponential map is proper at each point, we give a solution to the one dimensional asymptotic Plateau problem on $M$. Precisely, for any pair of antipodal points in the ideal boundary $\partial_\infty M = \mathbb S^{n-1}$, we prove the existence of a geodesic line with asymptotic prescribed boundaries and the Morse index $\le n-1$.

math.DG↗
arXiv2025

Backdoor Mitigation by Distance-Driven Detoxification

Backdoor attacks undermine the integrity of machine learning models by allowing attackers to manipulate predictions using poisoned training data. Such attacks lead to targeted misclassification when specific triggers are present, while the model behaves normally under other conditions. This paper considers a post-training backdoor defense task, aiming to detoxify the backdoors in pre-trained models. We begin by analyzing the underlying issues of vanilla fine-tuning and observe that it is often trapped in regions with low loss for both clean and poisoned samples. Motivated by such observations, we propose Distance-Driven Detoxification (D3), an innovative approach that reformulates backdoor defense as a constrained optimization problem. Specifically, D3 promotes the model's departure from the vicinity of its initial weights, effectively reducing the influence of backdoors. Extensive experiments on state-of-the-art (SOTA) backdoor attacks across various model architectures and datasets demonstrate that D3 not only matches but often surpasses the performance of existing SOTA post-training defense techniques.

cs.CR↗
arXiv2025

Revisiting the Auxiliary Data in Backdoor Purification

Backdoor attacks occur when an attacker subtly manipulates machine learning models during the training phase, leading to unintended behaviors when specific triggers are present. To mitigate such emerging threats, a prevalent strategy is to cleanse the victim models by various backdoor purification techniques. Despite notable achievements, current state-of-the-art (SOTA) backdoor purification techniques usually rely on the availability of a small clean dataset, often referred to as auxiliary dataset. However, acquiring an ideal auxiliary dataset poses significant challenges in real-world applications. This study begins by assessing the SOTA backdoor purification techniques across different types of real-world auxiliary datasets. Our findings indicate that the purification effectiveness fluctuates significantly depending on the type of auxiliary dataset used. Specifically, a high-quality in-distribution auxiliary dataset is essential for effective purification, whereas datasets from varied or out-of-distribution sources significantly degrade the defensive performance. Based on this, we propose Guided Input Calibration (GIC), which aims to improve purification efficacy by employing a learnable transformation. Guided by the victim model itself, GIC aligns the characteristics of the auxiliary dataset with those of the original training set. Comprehensive experiments demonstrate that GIC can substantially enhance purification performance across diverse types of auxiliary datasets. The code and data will be available via https://github.com/shawkui/BackdoorBenchER.

cs.CR↗
arXiv2024

On the Hausdorff dimension of circular Furstenberg sets

For $0 \leq s \leq 1$ and $0 \leq t \leq 3$, a set $F \subset \mathbb{R}^{2}$ is called a circular $(s,t)$-Furstenberg set if there exists a family of circles $\mathcal{S}$ of Hausdorff dimension $\dim_{\mathrm{H}} \mathcal{S} \geq t$ such that $$\dim_{\mathrm{H}} (F \cap S) \geq s, \qquad S \in \mathcal{S}.$$ We prove that if $0 \leq t \leq s \leq 1$, then every circular $(s,t)$-Furstenberg set $F \subset \mathbb{R}^{2}$ has Hausdorff dimension $\dim_{\mathrm{H}} F \geq s + t$. The case $s = 1$ follows from earlier work of Wolff on circular Kakeya sets.

math.CA↗
arXiv2023

Restricted Projections to Lines in $\mathbb{R}^{n+1}$

We prove the following restricted projection theorem. Let $n\ge 3$ and $Σ\subset S^{n}$ be an $(n-1)$-dimensional $C^2$ manifold such that $Σ$ has sectional curvature $>1$. Let $Z \subset \mathbb{R}^{n+1}$ be analytic and let $0 < s < \min\{\dim Z, 1\}$. Then \begin{equation*} \dim \{z \in Σ: \dim (Z \cdot z) < s\} \le (n-2)+s = (n-1) + (s-1) < n-1. \end{equation*} In particular, for almost every $z \in Σ$, $\dim (Z \cdot z) = \min\{\dim Z, 1\}$. The core idea, originated from Käenmäki-Orponen-Venieri, is to transfer the restricted projection problem to the study of the dimension lower bound of Furstenberg sets of cinematic family contained in $C^2([0,1]^{n-1})$. This cinematic family of functions with multivariables are extensions of those of one variable by Pramanik-Yang-Zahl and Sogge. Since the Furstenberg sets of cinematic family contain the affine Furstenberg sets as a special case, the dimension lower bound of Furstenberg sets improves the one by Héra, Héra-Keleti-Máthé and D{ą}browski-Orponen-Villa. Moreover, our method to show the restricted projection theorem can also give a new proof for the Mattila's projection theorem in $\mathbb{R}^n$ with $n \ge 3$.

math.CA↗
arXiv2023

Existence of hyperbolic motions to a class of Hamiltonians and generalized $N$-body system via a geometric approach

For the classical $N$-body problem in $\mathbb{R}^d$ with $d\ge2$, Maderna-Venturelli in their remarkable paper [Ann. Math. 2020] proved the existence of hyperbolic motions with any positive energy constant, starting from any configuration and along any non-collision configuration. Their original proof relies on the long time behavior of solutions by Chazy 1922 and Marchal-Saari 1976, on the Hölder estimate for Mañé's potential by Maderna 2012, and on the weak KAM theory. We give a new and completely different proof for the above existence of hyperbolic motions. The central idea is that, via some geometric observation, we build up uniform estimates for Euclidean length and angle of geodesics of Mañé's potential starting from a given configuration and ending at the ray along a given non-collision configuration. Note that we do not need any of the above previous studies used in Maderna-Venturelli's proof. Moreover, our geometric approach works for Hamiltonians $\frac12\|p\|^2-F(x)$, where $F(x)\ge 0$ is lower semicontinuous and decreases very slowly to $0$ faraway from collisions. We therefore obtain the existence of hyperbolic motions to such Hamiltonians with any positive energy constant, starting from any admissible configuration and along any non-collision configuration. Consequently, for several important potentials $F\in C^{2}(Ω)$, we get similar existence of hyperbolic motions to the generalized $N$-body system $\ddot{x} = \nabla_x F(x)$, which is an extension of Maderna-Venturelli [Ann. Math. 2020].

math.DS↗
arXiv2023

Dimension Estimates on Circular $(s,t)$-Furstenberg Sets

In this paper, we show that circular $(s,t)$-Furstenberg sets in $\mathbb R^2$ have Hausdorff dimension at least $$\max\{\frac{t}3+s,(2t+1)s-t\} \text{ for all $0<s,t\le 1$}.$$ This result extends the previous dimension estimates on circular Kakeya sets by Wolff.

math.CA↗
arXiv2023

Mean Parity Fair Regression in RKHS

We study the fair regression problem under the notion of Mean Parity (MP) fairness, which requires the conditional mean of the learned function output to be constant with respect to the sensitive attributes. We address this problem by leveraging reproducing kernel Hilbert space (RKHS) to construct the functional space whose members are guaranteed to satisfy the fairness constraints. The proposed functional space suggests a closed-form solution for the fair regression problem that is naturally compatible with multiple sensitive attributes. Furthermore, by formulating the fairness-accuracy tradeoff as a relaxed fair regression problem, we derive a corresponding regression function that can be implemented efficiently and provides interpretable tradeoffs. More importantly, under some mild assumptions, the proposed method can be applied to regression problems with a covariance-based notion of fairness. Experimental results on benchmark datasets show the proposed methods achieve competitive and even superior performance compared with several state-of-the-art methods.

stat.ML↗
arXiv2023

A Rademacher type theorem for Hamiltonians $H(x,p)$ and application to absolute minimizers

We establish a Rademacher type theorem involving Hamiltonians $H(x,p)$ under very weak conditions in both of Euclidean and Carnot-Carathéodory spaces. In particular,$H(x,p)$ is assumed to be only measurable in the variable $x$, and to be quasiconvex and lower-semicontinuous in the variable $p$. Without the lower-semicontinuity in the variable $p$, we provide a counter example showing the failure of such a Rademacher type theorem. Moreover, by applying such a Rademacher type theorem we build up an existence result of absolute minimizers for the corresponding $L^\infty$-functional. These improve or extend several known results in the literature.

math.CA↗
arXiv2021

On the Dimension of Kakeya Sets in the First Heisenberg Group

We define Kakeya sets in the Heisenberg group and show that the Heisenberg Hausdorff dimension of Kakeya sets in the first Heisenberg group is at least 3. This lower bound is sharp since, under our definition, the $\{xoy\}$-plane is a Kakeya set with Heisenberg Hausdorff dimension 3.

math.CA↗