papers

Publications (35)

cs.LG2022

Example Perplexity

Nevin L. Zhang, Weiyan Xie, Zhi Lin +4

Some examples are easier for humans to classify than others. The same should be true for deep neural networks (DNNs). We use the term example perplexity to refer to the level of di…

cs.LG2016

Deep Adaptive Network: An Efficient Deep Neural Network with Sparse Binary Connections

Xichuan Zhou, Shengli Li, Kai Qin +5

Deep neural networks are state-of-the-art models for understanding the content of images, video and raw input data. However, implementing a deep neural network in embedded systems…

cs.LG2023

A Causal Framework to Unify Common Domain Generalization Approaches

Nevin L. Zhang, Kaican Li, Han Gao +5

Domain generalization (DG) is about learning models that generalize well to new domains that are related to, but different from, the training domain(s). It is a fundamental problem…

cond-mat.quant-gas2025

Beyond Conventional Pairing: Bosonic Quartic Superfluidity, Exotic XY Magnetism, and Phase Criticality

Chenrong Liu, Zhi Lin

We report two unprecedented bosonic quartic superfluid (BQSF) phases in binary boson mixtures with synthetic pair-hopping (SPH) interaction realizable through Floquet engineering,…

cs.LG2024

Consistency Regularization for Domain Generalization with Logit Attribution Matching

Han Gao, Kaican Li, Weiyan Xie +5

Domain generalization (DG) is about training models that generalize well under domain shift. Previous research on DG has been conducted mostly in single-source or multi-source sett…

math.OC2024

Optimal Stirring Strategies for Passive Scalars in a Domain with a General Shape and No-Flux Boundary Condition

Sirui Zhu, Zhi Lin, Liang Li +1

Multiscale metrics such as negative Sobolev norms are effective for quantifying the degree of mixedness of a passive scalar field advected by an incompressible flow in the absence…

physics.atm-clus2022

Phase diagram of the two-component bosonic system with pair hopping in synthetic dimension

Chenrong Liu, Zhi Lin

We systematically study the ground-state phase diagrams and the demixing effect of a two-dimensional two-component bosonic system with pair hopping in synthetic dimension by using…

cond-mat.quant-gas2020

Novel quantum phases of two-component bosons with pair hopping in synthetic dimension

Zhi Lin, Chenrong Liu, Yan Chen

We study two-component (or pseudospin-1/2) bosons with pair hopping interactions in synthetic dimension, for which a feasible experimental scheme on a square optical lattice is als…

cond-mat.stat-mech2011

Quantum phase transitions of ultra-cold Bose system in non-rectangular optical lattices

Zhi Lin, Jun Zhang, Ying Jiang

In this paper, we investigate systematically the Mott-insulator-Superfluid quantum phase transitions for ultracold scalar bosons in triangular, hexagonal, as well as Kagomé optica…

cs.CL2025

HAVE: Head-Adaptive Gating and ValuE Calibration for Hallucination Mitigation in Large Language Models

Xin Tong, Zhi Lin, Jingya Wang +1

Large Language Models (LLMs) often produce hallucinations in retrieval-augmented or long-context generation, even when relevant evidence is present. This stems from two issues: hea…

physics.flu-dyn2016

Stirring by multi-cylinder in potential flow

Zhi Lin, Yuanzhao Zhang

We consider the enhanced mixing due to multiple cylinders organised in schools moving synchronously in a potential flow. Here simple interactions between cylinders are modelled by…

cs.LG2025

MEUV: Achieving Fine-Grained Capability Activation in Large Language Models via Mutually Exclusive Unlock Vectors

Xin Tong, Zhi Lin, Jingya Wang +2

Large language models (LLMs) enforce safety alignment to reliably refuse malicious requests, yet the same blanket safeguards also block legitimate uses in policing, defense, and ot…

cond-mat.quant-gas2012

Visibility of ultra-cold Bose system in triangular optical lattices

Zhi Lin, Jun Zhang, Ying Jiang

In this paper, by treating the hopping parameter in Bose-Hubbard model as a perturbation, with the help of the re-summed Green's function method and cumulants expansion, the moment…

math.NA2025

Fourth- and Higher-Order Semi-Lagrangian Finite Volume Methods for the Two-dimensional Advection Equation on Arbitrarily Complex Domains

Yunxia Sun, Kaiyi Liang, Yuke Zhu +2

To numerically solve the two-dimensional advection equation, we propose a family of fourth- and higher-order semi-Lagrangian finite volume (SLFV) methods that feature (1) fourth-,…

physics.flu-dyn2010

Models and measures of mixing and effective diffusion

Zhi Lin, Katarína Bodová, Charles R. Doering

Mixing a passive scalar field by stirring can be measured in a variety of ways including tracer particle dispersion, via the flux-gradient relationship, or by suppression of scalar…

cond-mat.stat-mech2014

An analytical approach to quantum phase transitions of ultracold Bose systems in bipartite optical lattices: Along the avenue of Green's function

Zhi Lin, Jun Zhang, Yan Chen +1

In this paper, we present a generalized Green's function method which can be used to investigate the quantum phase transitions analytically in a systematic way for ultracold Bose s…

cs.IT2022

Gain without Pain: Recycling Reflected Energy from Wireless Powered RIS-aided Communications

Hao Xie, Bowen Gu, Dong Li +2

In this paper, we investigate and analyze energy recycling for a reconfigurable intelligent surface (RIS)-aided wireless-powered communication network. As opposed to the existing w…

cond-mat.quant-gas2022

Generalized effective-potential Landau theory for a tunable state-dependent hexagonal optical lattice

Sheng Yue, Dan-Yang Chen, Chenrong Liu +2

We analytically study the ground-state phase diagrams of ultracold bosons with various values of the effective magnetic quantum number in a state-dependent hexagonal optical la…

physics.flu-dyn2010

Stirring by squirmers

Zhi Lin, Jean-Luc Thiffeault, Stephen Childress

We analyse a simple 'Stokesian squirmer' model for the enhanced mixing due to swimming micro-organisms. The model is based on a calculation of Thiffeault & Childress [Physics Lette…

cond-mat.quant-gas2026

Grand-canonical phase diagram and chiral-current suppression at flux in a bosonic two-leg ladder

Meng Zhang, Qingyun Xu, Zhi Lin

We investigate the ground-state phase diagram of repulsively interacting bosons on a two-leg ladder threaded by a uniform artificial magnetic flux, using the cluster Gutzwiller mea…

cond-mat.quant-gas2020

Phase diagrams of tunable spin-orbit coupled Bose-Einstein condensates

Zhi Lin

We analytically calculate phase boundaries of tunable spin-orbit coupled BECs with effective two-body interactions by using variational method. Phase diagrams for periodically driv…

cond-mat.quant-gas2018

Analytic calculation of high order corrections to quantum phase transitions of ultracold Bose gases in bipartite superlattices

Zhi Lin, Wanli Liu

We clarify some technical issues in the present generalized effective-potential Landau theory (GEPLT) that makes the GEPLT more consistent and complete. Utilizing this clarified GE…

quant-ph2020

Computable steering criterion for bipartite quantum systems

Guo-Zhu Pan, Jun-Long Zhao, Zhi Lin +3

Quantum steering describes the ability of one observer to nonlocally affect the other observer's state through local measurements, which represents a new form of quantum nonlocal c…

cond-mat.quant-gas2019

Chiral magnetic effect in three-dimensional optical lattices

Zhen Zheng, Zhi Lin, Dan-Wei Zhang +2

Although Weyl semimetals have been extensively studied for exploring rich topological physics, the direct observation of the celebrated chiral magnetic effect (CME) associated with…

quant-ph2019

Emulating topological currents arising from a dipolar parity anomaly in two-dimensional optical lattices

Zhi Lin, Xian-Jia Huang, Dan-Wei Zhang +2

Dipolar parity anomaly can be induced by spatiotemporally weak-dependent energy-momentum separation of paired Dirac points in two-dimensional Dirac semimetals. Here we reveal topol…

cs.CV2025

A mini-batch training strategy for deep subspace clustering networks

Yuxuan Jiang, Chenwei Yu, Zhi Lin +1

Mini-batch training is a cornerstone of modern deep learning, offering computational efficiency and scalability for training complex architectures. However, existing deep subspace…

eess.IV2022

A New Dataset and A Baseline Model for Breast Lesion Detection in Ultrasound Videos

Zhi Lin, Junhao Lin, Lei Zhu +3

Breast lesion detection in ultrasound is critical for breast cancer diagnosis. Existing methods mainly rely on individual 2D ultrasound images or combine unlabeled video and labele…

cs.CV2023

Two-Stage Holistic and Contrastive Explanation of Image Classification

Weiyan Xie, Xiao-Hui Li, Zhi Lin +3

The need to explain the output of a deep neural network classifier is now widely recognized. While previous methods typically explain a single class in the output, we advocate expl…

physics.flu-dyn2010

Optimal stirring strategies for passive scalar mixing

Zhi Lin, Jean-Luc Thiffeault, Charles R. Doering

We address the challenge of optimal incompressible stirring to mix an initially inhomogeneous distribution of passive tracers. As a quantitative measure of mixing we adopt the $H^{…

cond-mat.quant-gas2024

Delta-Learning approach combined with the cluster Gutzwiller approximation for strongly correlated bosonic systems

Zhi Lin, Tong Wang, Sheng Yue

The cluster Gutzwiller method is widely used to study the strongly correlated bosonic systems, owing to its ability to provide a more precise description of quantum fluctuations. H…

cond-mat.str-el2024

Frustrated hopping from orbital decoration of a primitive two-dimensional lattice

Aravind Devarakonda, Christie S. Koay, Daniel G. Chica +18

Materials hosting flat electronic bands are a central focus of condensed matter physics as promising venues for novel electronic ground states. Two-dimensional (2D) geometrically f…

cond-mat.str-el2020

Effective p-wave Fermi-Fermi Interaction Induced by Bosonic Superfluids

Yongzheng Wu, Zheng Yan, Zhi Lin +2

We study the two-dimensional Bose-Fermi mixture on square lattice at finite temperature by using the determinant quantum Monte Carlo method within the weakly interacting regime. He…

cs.AI2024

CPSDBench: A Large Language Model Evaluation Benchmark and Baseline for Chinese Public Security Domain

Xin Tong, Bo Jin, Zhi Lin +3

Large Language Models (LLMs) have demonstrated significant potential and effectiveness across multiple application domains. To assess the performance of mainstream LLMs in public s…

cond-mat.quant-gas2018

Generalized effective-potential Landau theory for the two-dimensional extended Bose-Hubbard model

Zhi Lin, Ming Yang

We analytically study the quantum phase diagrams of ultracold dipolar Bose gases in an optical square lattice at zero temperature by using the generalized effective-potential Landa…

cond-mat.quant-gas2025

Floquet-Engineered Hybrid Topological Orders with Majorana Edge Modes in Number-Conserving Fermionic Quantum Simulators

Zhi Lin, Qi Song, Sheng Yue +3

We develop an experimental protocol based on Floquet-engineered ultracold fermions in optical lattices, enabling the emulation of pair-hopping and competing singlet/triplet pairing…