BZPEERPapers, people, evidence and research paths
Menu

Discover

SearchFind papers, people, topics, institutions and opportunities from one place.GraphSee how papers, authors, topics and evidence connect.

Research tools

HomeContinue from the papers, people and topics that matter now.CollectionsTurn saved papers into reading lists, evidence maps and shareable context.ResearchReconstruct the path that led you from search to paper to question.CollaborationMove from a paper, topic or expert request into a focused thread.PublishCreate full-text research articles with metadata, formulas and versions.

Network

ExpertsFind specialists whose work and reviews explain why they can help.CommunitiesJoin research groups organized around papers, topics and shared evidence.

Opportunities

ListingsBrowse roles, calls and collaborations connected to real research context.

Account

Log inReturn to your saved papers, graph views and active threads.Create accountStart saving papers, building a research trail and joining teams.
Log inCreate account

Source author record

Cheng Jin

Cheng Jin appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record
physics.atom-phphysics.opticsMachine LearningComputer VisionArtificial IntelligenceCryptography and SecurityDistributed, Parallel, and Cluster ComputingNetworking and Internet ArchitectureNeural and Evolutionary Computingphysics.soc-phSocial and Information Networks

Catalog footprint

What is connected

15works
11topics
4close collaborators

Actions

Connect this record

Log in to claim
Open graphBrowse works

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

15 published item(s)

preprint2026arXiv

A Breast Vision Pathology Foundation Model for Real-world Clinical Utility

Pathology foundation models have shown strong retrospective performance, but whether such systems can support clinically relevant use remains unclear. This challenge is particularly important in breast cancer, where pathological assessment serves as the gold standard for diagnosis and guides treatment planning, surgical decision-making and risk stratification across pre-, intra- and post-operative stages. Here we present \textbf{BRAVE}, a breast-adaptive pathology foundation model developed and evaluated using a total resource of 101,638 breast whole-slide images from 32 sources across Asia, Europe and North America. We assessed BRAVE across 34 tasks in 82 cohorts spanning pre-operative biopsy, intra-operative frozen section and post-operative resection, using an evidence chain comprising retrospective benchmarking, clinically challenging scenarios, workflow-oriented clinical impact simulations, prospective observational validation with the thresholds locked in the retrospective cohorts and crossover pathologist-AI interaction studies. Across these settings, BRAVE supported practical roles in the clinical workflow, including safe exclusion of low-risk cases from routine review, AI-assisted second-review rescue of initially missed positives and prioritization of cases for further assessment. In prospective validation across three centres, BRAVE excluded 76.9% of negative biopsy cases (NPV 0.953) and 70.1% of negative frozen-section cases (NPV 0.973), and triaged 78.8% of post-operative subtyping cases as high-confidence clear-cut cases (NPV 1.000). In reader studies, AI assistance improved balanced accuracy from 88.5% to 95.1% (OR 3.14, P<0.001), with better efficiency, confidence and inter-rater agreement. BRAVE-derived scores also independently predicted disease-free survival (adjusted HR 4.79, P<0.001) and overall survival (adjusted HR 8.14, P<0.001).

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies

Deploying and fine-tuning Large Language Models (LLMs) on resource-constrained edge devices requires navigating a strict trade-off between memory footprint and task performance. While Quantization-Aware Fine-tuning has emerged as a viable solution, existing paradigms typically decouple quantization and adapter optimization. This separation overlooks a fundamental theoretical constraint we identify as the \textit{Fidelity-Plasticity Trade-off}: a layer's capacity to adapt to new tasks (Plasticity) is inherently constrained by the information capacity of its frozen weights (Fidelity). Aggressively quantizing semantically critical layers creates an information bottleneck that no amount of adapter rank can recover, while high precision in robust syntactic layers wastes valuable memory. To address this, we introduce \textbf{QR-Adaptor}, a unified framework that jointly optimizes per-layer quantization bit-width and LoRA rank. By formulating resource allocation as a multi-objective search aligned with the model's linguistic hierarchy, our method systematically liberates memory from redundancy-heavy layers to reinvest in capacity-critical ones. Extensive experiments demonstrate that QR-Adaptor establishes a new Pareto frontier: notably, a model fine-tuned under a strict 4-bit memory budget achieves performance rivaling 16-bit baselines, demonstrating that precise resource alignment is as critical as model size.

0 citations0 reviewsNo code linkedOpen access
preprint2024arXiv

EPA: Neural Collapse Inspired Robust Out-of-Distribution Detector

Out-of-distribution (OOD) detection plays a crucial role in ensuring the security of neural networks. Existing works have leveraged the fact that In-distribution (ID) samples form a subspace in the feature space, achieving state-of-the-art (SOTA) performance. However, the comprehensive characteristics of the ID subspace still leave under-explored. Recently, the discovery of Neural Collapse ($\mathcal{NC}$) sheds light on novel properties of the ID subspace. Leveraging insight from $\mathcal{NC}$, we observe that the Principal Angle between the features and the ID feature subspace forms a superior representation for measuring the likelihood of OOD. Building upon this observation, we propose a novel $\mathcal{NC}$-inspired OOD scoring function, named Entropy-enhanced Principal Angle (EPA), which integrates both the global characteristic of the ID subspace and its inner property. We experimentally compare EPA with various SOTA approaches, validating its superior performance and robustness across different network architectures and OOD datasets.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

SIT: A Bionic and Non-Linear Neuron for Spiking Neural Network

Spiking Neural Networks (SNNs) have piqued researchers' interest because of their capacity to process temporal information and low power consumption. However, current state-of-the-art methods limited their biological plausibility and performance because their neurons are generally built on the simple Leaky-Integrate-and-Fire (LIF) model. Due to the high level of dynamic complexity, modern neuron models have seldom been implemented in SNN practice. In this study, we adopt the Phase Plane Analysis (PPA) technique, a technique often utilized in neurodynamics field, to integrate a recent neuron model, namely, the Izhikevich neuron. Based on the findings in the advancement of neuroscience, the Izhikevich neuron model can be biologically plausible while maintaining comparable computational cost with LIF neurons. By utilizing the adopted PPA, we have accomplished putting neurons built with the modified Izhikevich model into SNN practice, dubbed as the Standardized Izhikevich Tonic (SIT) neuron. For performance, we evaluate the suggested technique for image classification tasks in self-built LIF-and-SIT-consisted SNNs, named Hybrid Neural Network (HNN) on static MNIST, Fashion-MNIST, CIFAR-10 datasets and neuromorphic N-MNIST, CIFAR10-DVS, and DVS128 Gesture datasets. The experimental results indicate that the suggested method achieves comparable accuracy while exhibiting more biologically realistic behaviors on nearly all test datasets, demonstrating the efficiency of this novel strategy in bridging the gap between neurodynamics and SNN practice.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Electron-Backscattering-Assisted High Harmonic Generation from Bilayer Nanostructures

In the framework of time-dependent density functional theory, we obtain high-order harmonics of photon energies up to 10 Up from bilayer crystals with an interlayer spacing d = 70 Å. At grazing incidence, a clear double-plateau structure is observed in the harmonic spectrum. The photon energy of the second plateau far beyond atomic-like harmonics can be well explained by the inclusion of backscattering of ionized electrons. Ab initio simulations reveal that the cutoff of the second plateau is continuously extended with an increasing d. Our classical calculations predict that the maximum electronic kinetic energy is linearly dependent on d over a wide range. Moreover, the harmonic yield in the second plateau is significantly enhanced by increases in the wavelength of the driving laser. Owing to the confined spreading of the electronic wave packet, a beneficial wavelength scaling of λ2.85 is obtained. This study therefore establishes a novel and efficient way of producing high-energy light source based on layered nanostructures.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Intelligent social bots uncover the link between user preference and diversity of news consumption

The boom of online social media and microblogging platforms has rapidly alter the way we consume news and exchange opinions. Even though considerable efforts try to recommend various contents to users, loss of information diversity and the polarization of interest groups are still an enormous challenge for industry and academia. Here, we take advantage of benign social bots to design a controlled experiment on Weibo (the largest microblogging platform in China). These software bots can exhibit human-like behavior (e.g., preferring particular content) and simulate the formation of personal social networks and news consumption under two well-accepted sociological hypotheses (i.e., homophily and triadic closure). We deployed 68 bots to Weibo, and each bot ran for at least 2 months and followed 100 to 120 accounts. In total, we observed 5,318 users and recorded about 630,000 messages exposed to these bots. Our results show, even with the same selection behaviors, bots preferring entertainment content are more likely to form polarized communities with their peers, in which about 80\% of the information they consume is of the same type, which is a significant difference for bots preferring sci-tech content. The result suggests that users preference played a more crucial role in limiting themselves access to diverse content by compared with the two well-known drivers (self-selection and pre-selection). Furthermore, our results reveal an ingenious connection between specific content and its propagating sub-structures in the same social network. In the Weibo network, entertainment news favors a unidirectional star-like sub-structure, while sci-tech news spreads on a bidirectional clustering sub-structure. This connection can amplify the diversity effect of user preference. The discovery may have important implications for diffusion dynamics study and recommendation system design.

0 citations0 reviewsNo code linkedOpen access
preprint2012arXiv

Groupwise Constrained Reconstruction for Subspace Clustering

Reconstruction based subspace clustering methods compute a self reconstruction matrix over the samples and use it for spectral clustering to obtain the final clustering result. Their success largely relies on the assumption that the underlying subspaces are independent, which, however, does not always hold in the applications with increasing number of subspaces. In this paper, we propose a novel reconstruction based subspace clustering model without making the subspace independence assumption. In our model, certain properties of the reconstruction matrix are explicitly characterized using the latent cluster indicators, and the affinity matrix used for spectral clustering can be directly built from the posterior of the latent cluster indicators instead of the reconstruction matrix. Experimental results on both synthetic and real-world datasets show that the proposed model can outperform the state-of-the-art methods.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Analysis of effects of macroscopic propagation and multiple molecular orbitals on the minimum in high-order harmonic generation of aligned CO$_{2}$

We report theoretical calculations on the effect of the multiple orbital contribution in high-order harmonic generation (HHG) from aligned CO$_2$ with inclusion of macroscopic propagation of harmonic fields in the medium. Our results show very good agreements with recent experiments for the dynamics of the minimum in HHG spectra as laser intensity or alignment angle changes. Calculations are carried out to check how the position of the minimum in HHG spectra depends on the degrees of molecular alignment, laser focusing conditions, and the effects of alignment-dependent ionization rates of the different molecular orbitals. These analyses help to explain why the minima observed in different experiments may vary.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Generation of isolated attosecond pulses in the far field by spatial filtering with an intense few-cycle mid-infrared laser

We report theoretical calculations of high-order harmonic generation (HHG) of Xe with the inclusion of multi-electron effects and macroscopic propagation of the fundamental and harmonic fields in an ionizing medium. By using the time-frequency analysis we show that the reshaping of the fundamental laser field is responsible for the continuum structure in the HHG spectra. We further suggest a method for obtaining an isolated attosecond pulse (IAP) by using a filter centered on axis to select the harmonics in the far field with different divergence. We also discuss the carrier-envelope-phase dependence of an IAP and the possibility to optimize the yield of the IAP. With the intense few-cycle mid-infrared lasers, this offers a possible method for generating isolated attosecond pulses.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Influence of gas pressure on high-order harmonic generation of Ar and Ne

We study the effect of gas pressure on the generation of high-order harmonics where harmonics due to individual atoms are calculated using the recently developed quantitative rescattering theory, and the propagation of the laser and harmonics in the medium is calculated by solving the Maxwell's wave equation. We illustrate that the simulated spectra are very sensitive to the laser focusing conditions at high laser intensity and high pressure since the fundamental laser field is severely reshaped during the propagation. By comparing the simulated results with several experiments we show that the pressure dependence can be qualitatively explained. The lack of quantitative agreement is tentatively attributed to the failure of the complete knowledge of the experimental conditions.

0 citations0 reviewsNo code linkedOpen access
preprint2011arXiv

Intensity dependence of multiple orbital contributions and shape resonance in high-order harmonic generation of aligned N$_{2}$ molecules}

We report measurements and theoretical simulations of high-order harmonic generation (HHG) in aligned N$_2$ molecules using a 1200-nm intense laser field when the generating pulse is perpendicular to the aligning one. With increasing laser intensity, the minimum in the HHG spectra first shifts its position and then disappears. Theoretical simulations including the macroscopic propagation effects in the medium reproduce these observations and the disappearance of the minimum is attributed to the additional contribution of HHG from inner orbitals. We also predict that the well-known shape resonance in the photoionization spectra of N$_2$ should exist in the HHG spectra. It is most clearly seen when the generating laser is parallel to the aligning one, and disappears gradually as the angle between the two lasers increases. No clear evidence of this shape resonance has been reported so far when using lasers with different wavelengths. Further experimentation is needed to draw conclusions.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

A Practical Localization Algorithm Based on Wireless Sensor Networks

Many localization algorithms and systems have been developed by means of wireless sensor networks for both indoor and outdoor environments. To achieve higher localization accuracy, extra hardware equipments are utilized by most of the existing localization algorithms, which increase the cost and greatly limit the range of location-based applications. In this paper we present a method which can effectively meet different localization accuracy requirements of most indoor and outdoor location services in realistic applications. Our algorithm is composed of two phases: partition phase, in which the target region is split into small grids and localization refinement phase in which a higher accuracy location can be generated by applying a trick algorithm. A realistic demo system using our algorithm has been developed to illustrate its feasibility and availability. The results show that our algorithm can improve the localization accuracy.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

Medium propagation effects in high harmonic generation of Ar and N$_{2}$

We report theoretical calculations of high harmonic generation (HHG) by intense infrared lasers in atomic and molecular targets taking into account the macroscopic propagation of both fundamental and harmonic fields. On the examples of Ar and N$_2$, we demonstrate that these {\it ab initio} calculations are capable of accurately reproducing available experimental results with isotropic and aligned target media. We further present detailed analysis of HHG intensity and phase, under various experimental conditions, in particular, as the wavelength of the driving laser changes. Most importantly, our results strongly support the factorization of HHG at the macroscopic level into a product of a returning electron wave packet and the photorecombination transition dipole, under typical experimental conditions. This implies that the single-atom/molecule structure information can be retrieved from experimentally measured HHG spectra.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

New determination of structure parameters in strong field tunneling ionization theory of molecules

In the strong field molecular tunneling ionization theory of Tong et al. [Phys. Rev. A 66, 033402 (2002)], the ionization rate depends on the asymptotic wavefunction of the molecular orbital from which the electron is removed. The orbital wavefunctions obtained from standard quantum chemistry packages in general are not good enough in the asymptotic region. Here we construct a one-electron model potential for several linear molecules using density functional theory (DFT). We show that the asymptotic wavefunction can be improved with an iteration method and after one iteration accurate asymptotic wavefunctions and structure parameters are determined. With the new parameters we examine the alignment-dependent tunneling ionization probabilities for several molecules and compare with other calculations and with recent measurements, including ionization from inner molecular orbitals.

0 citations0 reviewsNo code linkedOpen access
preprint2010arXiv

Separation of Target Structure and Medium Propagation Effects in High-Harmonic Generation

We calculate high-harmonic generation (HHG) by intense infrared lasers in atoms and molecules with the inclusion of macroscopic propagation of the harmonics in the gas medium. We show that the observed experimental spectra can be accurately reproduced theoretically despite that HHG spectra are sensitive to the experimental conditions. We further demonstrate that the simulated (or experimental) HHG spectra can be factored out as a product of a \macroscopic wave packet" and the photo-recombination transition dipole moment where the former depends on the laser properties and the experimental conditions, while the latter is the property of the target only. The factorization makes it possible to extract target structure from experimental HHG spectra, and for ultrafast dynamic imaging of transient molecules.

0 citations0 reviewsNo code linkedOpen access