BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Lei Feng

Lei Feng contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
Machine Learninghep-phastro-ph.HEComputer VisionArtificial Intelligenceastro-ph.EPBiological PhysicsComputation and Languagecond-mat.mtrl-scicond-mat.quant-gasphysics.app-phphysics.flu-dynphysics.pop-phquant-ph

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
20works
0followers
14topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

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

Published work

20 published item(s)

preprint2026arXiv

COHERENCE: Benchmarking Fine-Grained Image-Text Alignment in Interleaved Multimodal Contexts

In recent years, Multimodal Large Language Models (MLLMs) have achieved remarkable progress on a wide range of multimodal benchmarks. Despite these advances, most existing benchmarks mainly focus on single-image or multi-image comprehension. In real-world scenarios such as document reading, information is often presented as interleaved multimodel contexts. This requires MLLMs not only to recognize the content of individual images, but also to identify relevant textual and visual evidence, establish fine-grained alignments between them, and reason over these aligned signals in interleaved contexts based on contextual evidence. However, there is still a lack of systematic benchmarks for quantifying the fine-grained understanding ability of MLLMs in interleaved image-text contexts. To fill this gap, we propose COHERENCE, a benchmark designed to evaluate the ability of MLLMs to recover fine-grained image-text correspondences in interleaved multimodal contexts. COHERENCE covers interleaved image-text content from four representative domains and contains 6,161 high-quality questions. Moreover, we perform a six-type error analysis, enabling fine-grained attribution of failures in interleaved image-text understanding to the specific capabilities missing in current MLLMs.

0 citations0 reviewsNo code linkedOpen access
preprint2023arXiv

keqing: knowledge-based question answering is a nature chain-of-thought mentor of LLM

Large language models (LLMs) have exhibited remarkable performance on various natural language processing (NLP) tasks, especially for question answering. However, in the face of problems beyond the scope of knowledge, these LLMs tend to talk nonsense with a straight face, where the potential solution could be incorporating an Information Retrieval (IR) module and generating response based on these retrieved knowledge. In this paper, we present a novel framework to assist LLMs, such as ChatGPT, to retrieve question-related structured information on the knowledge graph, and demonstrate that Knowledge-based question answering (Keqing) could be a nature Chain-of-Thought (CoT) mentor to guide the LLM to sequentially find the answer entities of a complex question through interpretable logical chains. Specifically, the workflow of Keqing will execute decomposing a complex question according to predefined templates, retrieving candidate entities on knowledge graph, reasoning answers of sub-questions, and finally generating response with reasoning paths, which greatly improves the reliability of LLM's response. The experimental results on KBQA datasets show that Keqing can achieve competitive performance and illustrate the logic of answering each question.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Direct measurement of vorticity using tracer particles with internal markers

Current experiment techniques for vorticity measurement suffer from limited spatial and temporal resolution to resolve the small-scale eddy dynamics in turbulence. In this study, we develop a new method for direct vorticity measurement in fluid flows based on digital inline holography (DIH). The DIH system utilizes a collimated laser beam to illuminate the tracers with internal markers and a digital sensor to record the generated holograms. The tracers made of the polydimethylsiloxane (PDMS) prepolymer mixed with internal markers are fabricated using a standard microfluidic droplet generator. A rotation measurement algorithm is developed based on the 3D location reconstruction and tracking of the internal markers and is assessed through synthetic holograms to identify the optimal parameter settings and measurement range (e.g., rotation rate from 0.3 to 0.7 rad/frame under numerical aperture of imaging of 0.25). Our proposed method based on DIH is evaluated by a calibration experiment of single tracer rotation, which yields the same optimal measurement range. Using von Kármán swirling flow setup, we further demonstrate the capability of the approach to simultaneously measure the Lagrangian rotation and translation of multiple tracers. Our method can measure vorticity in a small region on the order of 100 $μ$m or less and can be potentially used to quantify the Kolmogorov-scale vorticity field in turbulent flows.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Explanation of nearby SNRs for primary electron excess and proton spectral bump

Several groups have reported a possible excess of primary electrons at high energies with the joint fit of the positron fraction and total electron/positron spectra. With the latest release of high-precision electron/positron spectra measured by AMS-02, we further confirm this excess by fitting $ΔΦ$ $\rm(i.e., Φ_{e^-}-Φ_{e^+})$ data in this work. Then we investigate the contribution of a single nearby supernova remnant to the primary electron excess and find that Monogem can reasonably account for this excess. Moreover, we predict that the electron spectrum may harden again at a few TeVs due to Vela's contribution. DAMPE, which can accurately measure electrons at TeV scale, is expected to provide the robust test of this new spectral feature in the near future. Finally, we fit the proton spectrum data of DAMPE with Monogem or Loop I. We find that both the primary electron excess and the proton spectral bump could be mainly generated by Monogem.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

GearNet: Stepwise Dual Learning for Weakly Supervised Domain Adaptation

This paper studies weakly supervised domain adaptation(WSDA) problem, where we only have access to the source domain with noisy labels, from which we need to transfer useful information to the unlabeled target domain. Although there have been a few studies on this problem, most of them only exploit unidirectional relationships from the source domain to the target domain. In this paper, we propose a universal paradigm called GearNet to exploit bilateral relationships between the two domains. Specifically, we take the two domains as different inputs to train two models alternately, and asymmetrical Kullback-Leibler loss is used for selectively matching the predictions of the two models in the same domain. This interactive learning schema enables implicit label noise canceling and exploits correlations between the source and target domains. Therefore, our GearNet has the great potential to boost the performance of a wide range of existing WSDL methods. Comprehensive experimental results show that the performance of existing methods can be significantly improved by equipping with our GearNet.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Learning with Multiple Complementary Labels

A complementary label (CL) simply indicates an incorrect class of an example, but learning with CLs results in multi-class classifiers that can predict the correct class. Unfortunately, the problem setting only allows a single CL for each example, which notably limits its potential since our labelers may easily identify multiple CLs (MCLs) to one example. In this paper, we propose a novel problem setting to allow MCLs for each example and two ways for learning with MCLs. In the first way, we design two wrappers that decompose MCLs into many single CLs, so that we could use any method for learning with CLs. However, the supervision information that MCLs hold is conceptually diluted after decomposition. Thus, in the second way, we derive an unbiased risk estimator; minimizing it processes each set of MCLs as a whole and possesses an estimation error bound. We further improve the second way into minimizing properly chosen upper bounds. Experiments show that the former way works well for learning with MCLs but the latter is even better.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Mitigating Neural Network Overconfidence with Logit Normalization

Detecting out-of-distribution inputs is critical for safe deployment of machine learning models in the real world. However, neural networks are known to suffer from the overconfidence issue, where they produce abnormally high confidence for both in- and out-of-distribution inputs. In this work, we show that this issue can be mitigated through Logit Normalization (LogitNorm) -- a simple fix to the cross-entropy loss -- by enforcing a constant vector norm on the logits in training. Our method is motivated by the analysis that the norm of the logit keeps increasing during training, leading to overconfident output. Our key idea behind LogitNorm is thus to decouple the influence of output's norm during network optimization. Trained with LogitNorm, neural networks produce highly distinguishable confidence scores between in- and out-of-distribution data. Extensive experiments demonstrate the superiority of LogitNorm, reducing the average FPR95 by up to 42.30% on common benchmarks.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Nebula-Relay Hypothesis: The Chirality of Biological Molecules in Molecular Clouds

The chiral puzzle of biological molecules is thought to be closely related to the origin of life and is still a mystery so far. Previously, we proposed a new model on the origin of life, Nebula-Relay hypothesis, which assumed that the life on Earth originated at the planetary system of the Sun's predecessor star and then filled in the pre-solar nebula after its death. As primitive lives existed in the pre-solar nebula for a long time, did the chiral biomolecules form during this period? We explore such a possibility in this work and find that the ultra-low temperature environment of molecular clouds is beneficial to generating the chiral polymer chain of biological molecules.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Open-Sampling: Exploring Out-of-Distribution data for Re-balancing Long-tailed datasets

Deep neural networks usually perform poorly when the training dataset suffers from extreme class imbalance. Recent studies found that directly training with out-of-distribution data (i.e., open-set samples) in a semi-supervised manner would harm the generalization performance. In this work, we theoretically show that out-of-distribution data can still be leveraged to augment the minority classes from a Bayesian perspective. Based on this motivation, we propose a novel method called Open-sampling, which utilizes open-set noisy labels to re-balance the class priors of the training dataset. For each open-set instance, the label is sampled from our pre-defined distribution that is complementary to the distribution of original class priors. We empirically show that Open-sampling not only re-balances the class priors but also encourages the neural network to learn separable representations. Extensive experiments demonstrate that our proposed method significantly outperforms existing data re-balancing methods and can boost the performance of existing state-of-the-art methods.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Pointwise Binary Classification with Pairwise Confidence Comparisons

To alleviate the data requirement for training effective binary classifiers in binary classification, many weakly supervised learning settings have been proposed. Among them, some consider using pairwise but not pointwise labels, when pointwise labels are not accessible due to privacy, confidentiality, or security reasons. However, as a pairwise label denotes whether or not two data points share a pointwise label, it cannot be easily collected if either point is equally likely to be positive or negative. Thus, in this paper, we propose a novel setting called pairwise comparison (Pcomp) classification, where we have only pairs of unlabeled data that we know one is more likely to be positive than the other. Firstly, we give a Pcomp data generation process, derive an unbiased risk estimator (URE) with theoretical guarantee, and further improve URE using correction functions. Secondly, we link Pcomp classification to noisy-label learning to develop a progressive URE and improve it by imposing consistency regularization. Finally, we demonstrate by experiments the effectiveness of our methods, which suggests Pcomp is a valuable and practically useful type of pairwise supervision besides the pairwise label.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Constraining $U(1)_{L_μ-L_τ}$ charged dark matter model for muon $g-2$ anomaly with AMS-02 electron and positron data

Very recently, the Fermi-Lab reported the new experimental combined results on the magnetic momentum of muon with a 4.2$σ$ discrepancy compared with the expectation of the Standard Model \cite{Fermi_Lab}. A new light gauge boson $X$ in the $L_μ-L_τ$ model provides a good explanation for the $g-2$ anomaly. A Dirac fermion dark matter with a large $L_μ-L_τ$ charge can explain both the $g-2$ anomaly and the dark matter relic density \cite{Asai_2021}. In this work, we focus on the case that the mass of the dark matter is larger than the mass of muon (i.e. $m_Ψ > m_μ$) for which the channel $ΨΨ\rightarrow μ^- μ^+$ opens. Although the cross section $(σv)_{μ^{-}μ^{+}}$ is smaller by a factor of $1/q_Ψ^2$ ($q_Ψ$ represents the $L_μ-L_τ$ charge of the dark matter) compared with the channel $ΨΨ\rightarrow XX \rightarrow νν\barν\barν$, the resulting secondary electrons and positrons could imprint on their spectra above GeV energies due to the reacceleration effect of cosmic ray propagation. We use the AMS-02 measurements of electrons and positrons to constrain the annihilation cross section of the channel $ΨΨ\rightarrow μ^{-}μ^{+}$, which rules out part of the parameter space of the large $L_μ-L_τ$ charged dark matter model to account for the muon $g-2$ anomaly.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Learning from Similarity-Confidence Data

Weakly supervised learning has drawn considerable attention recently to reduce the expensive time and labor consumption of labeling massive data. In this paper, we investigate a novel weakly supervised learning problem of learning from similarity-confidence (Sconf) data, where we aim to learn an effective binary classifier from only unlabeled data pairs equipped with confidence that illustrates their degree of similarity (two examples are similar if they belong to the same class). To solve this problem, we propose an unbiased estimator of the classification risk that can be calculated from only Sconf data and show that the estimation error bound achieves the optimal convergence rate. To alleviate potential overfitting when flexible models are used, we further employ a risk correction scheme on the proposed risk estimator. Experimental results demonstrate the effectiveness of the proposed methods.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Anomalous bremsstrahlung and the structure of cosmic ray electron-positron fluxes at the GeV-TeV energy range

We reveal that the energy spectra of electrons-positrons in primary cosmic rays measured at atmosphere top have double structures: an excess component $Φ^s_{e^+}(E)=Φ^s_{e^-}(E)$ around $400 GeV$, which origins from a strong $e^+e^-$-source and the distorted background $Φ^0_{e^-}(E)$. We supposed that the difference between AMS-CALET and Fermi-LAT-DAMPE data origins from the energy loss of the fluxes due to the anomalous bremsstrahlung effect at a special window. The evolution of spectra under anomalous bremsstrahlung effect satisfies an improved electromagnetic cascade equation. The above spectra are parameterized and they can be regarded as the subjects exploring new physics. We suggest to check the previous applications of the Bethe-Heitler formula in the study of the propagation of high energy electrons and photons.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Combating noisy labels by agreement: A joint training method with co-regularization

Deep Learning with noisy labels is a practically challenging problem in weakly supervised learning. The state-of-the-art approaches "Decoupling" and "Co-teaching+" claim that the "disagreement" strategy is crucial for alleviating the problem of learning with noisy labels. In this paper, we start from a different perspective and propose a robust learning paradigm called JoCoR, which aims to reduce the diversity of two networks during training. Specifically, we first use two networks to make predictions on the same mini-batch data and calculate a joint loss with Co-Regularization for each training example. Then we select small-loss examples to update the parameters of both two networks simultaneously. Trained by the joint loss, these two networks would be more and more similar due to the effect of Co-Regularization. Extensive experimental results on corrupted data from benchmark datasets including MNIST, CIFAR-10, CIFAR-100 and Clothing1M demonstrate that JoCoR is superior to many state-of-the-art approaches for learning with noisy labels.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Interpretation of the cosmic ray positron and electron excesses with an annihilating-decaying dark matter scenario

The precise measurements of energy spectra of cosmic ray positrons and/or electrons by recent experiments show clear excesses above 10 GeV. Moreover, a potential sharp spectral feature was suggested by the Dark Matter Particle Explorer (DAMPE) data. These results inspire quite a number of discussions on the connection with either the annihilation/decay of dark matter (DM) or the astrophysical origins. Here we discuss a DM scenario in which DM particles could annihilate and decay into standard model particle pairs simultaneously. In this model, the peak structure is due to the DM annihilation in a nearby subhalo and the broad positron/electron excesses are due to the decay of DM in the Milky Way. This model can reasonably explain the DAMPE and AMS-02 data of the total $e^+e^-$ spectra and the positron fraction, with model parameters being consistent with existing constraints. A simple realization of such a DM model is the spin-1 vector DM model.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Plasma dark matter and electronic recoil events in XENON1T

Dark matter might be in the form of a dark plasma in the Milky Way halo. Specifically, we consider here a hidden sector consisting of a light `dark electron' and a much heavier `dark proton', each charged under an unbroken $U(1)'$ gauge symmetry. These self-interacting dark sector particles can also interact with ordinary matter via the kinetic mixing interaction, and lead to a signal in dark matter direct detection experiments. Indeed, keV electron recoils can arise quite naturally in such models from dark electron scattering off loosely bound atomic electrons. Here we examine the recently reported XENON1T excess in the context of such a plasma dark matter model. We find that the observed excess can be explained if kinetic mixing is in the approximate range: $10^{-12} \lesssim ε\lesssim 10^{-10}$. The allowed parameter space is consistent with astrophysical and cosmological constraints and consistent also with other direct detection experiments.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Progressive Identification of True Labels for Partial-Label Learning

Partial-label learning (PLL) is a typical weakly supervised learning problem, where each training instance is equipped with a set of candidate labels among which only one is the true label. Most existing methods elaborately designed learning objectives as constrained optimizations that must be solved in specific manners, making their computational complexity a bottleneck for scaling up to big data. The goal of this paper is to propose a novel framework of PLL with flexibility on the model and optimization algorithm. More specifically, we propose a novel estimator of the classification risk, theoretically analyze the classifier-consistency, and establish an estimation error bound. Then we propose a progressive identification algorithm for approximately minimizing the proposed risk estimator, where the update of the model and identification of true labels are conducted in a seamless manner. The resulting algorithm is model-independent and loss-independent, and compatible with stochastic optimization. Thorough experiments demonstrate it sets the new state of the art.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Two-Path Phonon-Interference Resonance Induces a Stop Band in Silicon Crystal Matrix by Embedded Nanoparticles Array

In this work, we report a new stop-band formation mechanism by performing the atomistic Green's function calculation and the wave-packet molecular dynamics simulation for a system with germanium-nanoparticle array embedded in a crystalline silicon matrix. When only a single nanoparticle is embedded, the local resonance, induced through destructive interference between two different phonon wave paths, gives rise to several sharp and significant transmittance dips. On the other hand, when the number of embedded nanoparticles further increases to ten, a stop band with complete phonon reflection is formed due to the two-path resonance Bragg-like phonon interference. The wave packet simulations further uncover that the stop band originates from the collective phonon resonances at the embedded nanoparticles. Compared with the traditional stop-band formation mechanism that is the single-path Bragg reflection, the resonance mechanism has a significant advantage in not requiring the strict periodicity in the embedded nanoparticles array. We also demonstrate that the stop band can significantly suppress thermal conductance in the low-frequency regime. Our work provides a robust, scalable. and easily modulable stop-band formation mechanism. which opens a new degree of freedom for phononics-related heat control.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Pattern formation in a driven Bose-Einstein Condensate

Pattern formation is ubiquitous in nature from morphogenesis and cloud formation to galaxy filamentation. More often than not, patterns arise in a medium driven far from equilibrium due to the interplay of dynamical instability and nonlinear wave mixing. We report, based on momentum and real space pattern recognition, formation of density patterns with two- (D$_2$), four- (D$_4$) and six-fold (D$_6$) symmetries in Bose-Einstein condensates (BECs) with atomic interactions driven at two frequencies. The symmetry of the pattern is controlled by the ratio of the frequencies. The D$_6$ density waves, in particular, arise from a resonant wave mixing process that coherently correlates and enhances the excitations that respect the symmetry.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Searching for the possible signal of the photon-axionlike particle oscillation in the combined GeV and TeV spectra of supernova remnants

The conversion between photons and axionlike particles (ALPs) in the Milky Way magnetic field could result in the detectable oscillation phenomena in $γ$-ray spectra of Galactic sources. In this work, the GeV (Fermi-LAT) and TeV (MAGIC/VERITAS/H.E.S.S.) data of three bright supernova remnants (SNRs, ie. IC443, W51C and W49B) have been adopted together to search such the oscillation effect. Different from our previous analysis of the sole Fermi-LAT data of IC443, we do not find any reliable signal for the photon-ALP oscillation in the joint broadband spectrum of each SNR. The reason for the inconsistence is that in this work we use the latest revision (P8R3) of Fermi-LAT data, updated diffuse emission templates and the new version of the source catalog (4FGL), which lead to some modification of the GeV spectrum of IC443. Then we set constraints on ALP parameters based on the combined analysis of all the three sources. Though these constraints are somewhat weaker than limits from the CAST experiment and globular clusters, they are supportive of and complementary to these other results.

0 citations0 reviewsNo code linkedOpen access