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

Zhenyu Zhang

Zhenyu Zhang contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
Computer VisionArtificial IntelligenceMachine LearningComputation and Languagehep-phcond-mat.mes-hallcond-mat.mtrl-sciInformation Theorymath.ITComputationcond-mat.str-elCryptography and SecurityDistributed, Parallel, and Cluster Computingeess.IVMethodologyMultimediaPopulations and EvolutionSoftware Engineering

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
31works
0followers
18topics
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

31 published item(s)

preprint2026arXiv

A Survey on Failure Analysis and Fault Injection in AI Systems

The rapid advancement of Artificial Intelligence (AI) has led to its integration into various areas, especially with Large Language Models (LLMs) significantly enhancing capabilities in Artificial Intelligence Generated Content (AIGC). However, the complexity of AI systems has also exposed their vulnerabilities, necessitating robust methods for failure analysis (FA) and fault injection (FI) to ensure resilience and reliability. Despite the importance of these techniques, there lacks a comprehensive review of FA and FI methodologies in AI systems. This study fills this gap by presenting a detailed survey of existing FA and FI approaches across six layers of AI systems. We systematically analyze 160 papers and repositories to answer three research questions including (1) what are the prevalent failures in AI systems, (2) what types of faults can current FI tools simulate, (3) what gaps exist between the simulated faults and real-world failures. Our findings reveal a taxonomy of AI system failures, assess the capabilities of existing FI tools, and highlight discrepancies between real-world and simulated failures. Moreover, this survey contributes to the field by providing a framework for fault diagnosis, evaluating the state-of-the-art in FI, and identifying areas for improvement in FI techniques to enhance the resilience of AI systems.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Interactive Evaluation Requires a Design Science

AI evaluation is undergoing a structural change. Large language models (LLMs) are increasingly deployed as systems that act over time through tools, environments, users, and other agents, while many evaluation practices still inherit assumptions from response-centered benchmarks (e.g., fixed inputs, isolated outputs, and outcome judgments that can be made from a single response). The field has begun to build interactive benchmarks, but the resulting landscape is fragmented: benchmarks differ in what interaction artifacts they admit, how trajectories are scored, and what claims their results support. This position paper argues that interactive evaluation should be treated as a principled evaluation paradigm, not merely a new family of agent benchmarks. Simply adopting previous evaluation paradigms does not suffice. We define evaluation as an autonomous mapping from evidence to judgments, and show that interactive evaluation changes both sides of this mapping: the evidence becomes interaction-generated trajectories, while the evaluation procedure must assess process, recoverability, coordination, robustness, and system-level performance. Building on this definition, we propose a two-axis taxonomy, derive design principles and reporting standards, examine representative scenarios, and analyze how longstanding evaluation challenges reappear at the trajectory level.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Muses: Designing, Composing, Generating Nonexistent Fantasy 3D Creatures without Training

We present Muses, the first training-free method for fantastic 3D creature generation in a feed-forward paradigm. Previous methods, which rely on part-aware optimization, manual assembly, or 2D image generation, often produce unrealistic or incoherent 3D assets due to the challenges of intricate part-level manipulation and limited out-of-domain generation. In contrast, Muses leverages the 3D skeleton, a fundamental representation of biological forms, to explicitly and rationally compose diverse elements. This skeletal foundation formalizes 3D content creation as a structure-aware pipeline of design, composition, and generation. Muses begins by constructing a creatively composed 3D skeleton with coherent layout and scale through graph-constrained reasoning. This skeleton then guides a voxel-based assembly process within a structured latent space, integrating regions from different objects. Finally, image-guided appearance modeling under skeletal conditions is applied to generate a style-consistent and harmonious texture for the assembled shape. Extensive experiments establish Muses' state-of-the-art performance in terms of visual fidelity and alignment with textual descriptions, and potential on flexible 3D object editing. Project page: https://luhexiao.github.io/Muses.github.io/.

0 citations0 reviewsNo code linkedOpen access
preprint2025arXiv

Fantastic Reasoning Behaviors and Where to Find Them: Unsupervised Discovery of the Reasoning Process

Despite the growing reasoning capabilities of recent large language models (LLMs), their internal mechanisms during the reasoning process remain underexplored. Prior approaches often rely on human-defined concepts (e.g., overthinking, reflection) at the word level to analyze reasoning in a supervised manner. However, such methods are limited, as it is infeasible to capture the full spectrum of potential reasoning behaviors, many of which are difficult to define in token space. In this work, we propose an unsupervised framework (namely, RISE: Reasoning behavior Interpretability via Sparse auto-Encoder) for discovering reasoning vectors, which we define as directions in the activation space that encode distinct reasoning behaviors. By segmenting chain-of-thought traces into sentence-level 'steps' and training sparse auto-encoders (SAEs) on step-level activations, we uncover disentangled features corresponding to interpretable behaviors such as reflection and backtracking. Visualization and clustering analyses show that these behaviors occupy separable regions in the decoder column space. Moreover, targeted interventions on SAE-derived vectors can controllably amplify or suppress specific reasoning behaviors, altering inference trajectories without retraining. Beyond behavior-specific disentanglement, SAEs capture structural properties such as response length, revealing clusters of long versus short reasoning traces. More interestingly, SAEs enable the discovery of novel behaviors beyond human supervision. We demonstrate the ability to control response confidence by identifying confidence-related vectors in the SAE decoder space. These findings underscore the potential of unsupervised latent discovery for both interpreting and controllably steering reasoning in LLMs.

0 citations0 reviewsNo code linkedOpen access
preprint2025arXiv

OpenGround: Active Cognition-based Reasoning for Open-World 3D Visual Grounding

3D visual grounding aims to locate objects based on natural language descriptions in 3D scenes. Existing methods rely on a pre-defined Object Lookup Table (OLT) to query Visual Language Models (VLMs) for reasoning about object locations, which limits the applications in scenarios with undefined or unforeseen targets. To address this problem, we present OpenGround, a novel zero-shot framework for open-world 3D visual grounding. Central to OpenGround is the Active Cognition-based Reasoning (ACR) module, which is designed to overcome the fundamental limitation of pre-defined OLTs by progressively augmenting the cognitive scope of VLMs. The ACR module performs human-like perception of the target via a cognitive task chain and actively reasons about contextually relevant objects, thereby extending VLM cognition through a dynamically updated OLT. This allows OpenGround to function with both pre-defined and open-world categories. We also propose a new dataset named OpenTarget, which contains over 7000 object-description pairs to evaluate our method in open-world scenarios. Extensive experiments demonstrate that OpenGround achieves competitive performance on Nr3D, state-of-the-art on ScanRefer, and delivers a substantial 17.6% improvement on OpenTarget. Project Page at https://why-102.github.io/openground.io/.

0 citations0 reviewsNo code linkedOpen access
preprint2023arXiv

Sam-Guided Enhanced Fine-Grained Encoding with Mixed Semantic Learning for Medical Image Captioning

With the development of multimodality and large language models, the deep learning-based technique for medical image captioning holds the potential to offer valuable diagnostic recommendations. However, current generic text and image pre-trained models do not yield satisfactory results when it comes to describing intricate details within medical images. In this paper, we present a novel medical image captioning method guided by the segment anything model (SAM) to enable enhanced encoding with both general and detailed feature extraction. In addition, our approach employs a distinctive pre-training strategy with mixed semantic learning to simultaneously capture both the overall information and finer details within medical images. We demonstrate the effectiveness of this approach, as it outperforms the pre-trained BLIP2 model on various evaluation metrics for generating descriptions of medical images.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Accelerating Bayesian inference of dependency between complex biological traits

Inferring dependencies between complex biological traits while accounting for evolutionary relationships between specimens is of great scientific interest yet remains infeasible when trait and specimen counts grow large. The state-of-the-art approach uses a phylogenetic multivariate probit model to accommodate binary and continuous traits via a latent variable framework, and utilizes an efficient bouncy particle sampler (BPS) to tackle the computational bottleneck -- integrating many latent variables from a high-dimensional truncated normal distribution. This approach breaks down as the number of specimens grows and fails to reliably characterize conditional dependencies between traits. Here, we propose an inference pipeline for phylogenetic probit models that greatly outperforms BPS. The novelty lies in 1) a combination of the recent Zigzag Hamiltonian Monte Carlo (Zigzag-HMC) with linear-time gradient evaluations and 2) a joint sampling scheme for highly correlated latent variables and correlation matrix elements. In an application exploring HIV-1 evolution from 535 viruses, the inference requires joint sampling from an 11,235-dimensional truncated normal and a 24-dimensional covariance matrix. Our method yields a 5-fold speedup compared to BPS and makes it possible to learn partial correlations between candidate viral mutations and virulence. Computational speedup now enables us to tackle even larger problems: we study the evolution of influenza H1N1 glycosylations on around 900 viruses. For broader applicability, we extend the phylogenetic probit model to incorporate categorical traits, and demonstrate its use to study Aquilegia flower and pollinator co-evolution.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

ASFD: Automatic and Scalable Face Detector

Along with current multi-scale based detectors, Feature Aggregation and Enhancement (FAE) modules have shown superior performance gains for cutting-edge object detection. However, these hand-crafted FAE modules show inconsistent improvements on face detection, which is mainly due to the significant distribution difference between its training and applying corpus, COCO vs. WIDER Face. To tackle this problem, we essentially analyse the effect of data distribution, and consequently propose to search an effective FAE architecture, termed AutoFAE by a differentiable architecture search, which outperforms all existing FAE modules in face detection with a considerable margin. Upon the found AutoFAE and existing backbones, a supernet is further built and trained, which automatically obtains a family of detectors under the different complexity constraints. Extensive experiments conducted on popular benchmarks, WIDER Face and FDDB, demonstrate the state-of-the-art performance-efficiency trade-off for the proposed automatic and scalable face detector (ASFD) family. In particular, our strong ASFD-D6 outperforms the best competitor with AP 96.7/96.2/92.1 on WIDER Face test, and the lightweight ASFD-D0 costs about 3.1 ms, more than 320 FPS, on the V100 GPU with VGA-resolution images.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Capacity Bounds for the Two-User IM/DD Interference Channel

This paper studies the capacity of the two-user intensity-modulation/direct-detection (IM/DD) interference channel (IC), which is relevant in the context of multi-user optical wireless communications. Despite some known single-letter capacity characterizations for general discrete-memoryless ICs, a computable capacity expression for the IM/DD IC is missing. In this paper, we provide tight and easily computable inner and outer bounds for a general two-user IM/DD IC under peak and average optical intensity constraints. The bounds enable characterizing the asymptotic sum-rate capacity in the strong and weak interference regimes, as well as the generalized degrees of freedom (GDoF) in the symmetric case. Using the obtained bounds, the GDoF of the IM/DD IC is shown to have a `W' shape similar to the Gaussian IC with power constraints. The obtained bounds are also evaluated numerically in different interference regimes to show their tightness, and used to study the performance of on-chip and indoor OWC systems.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Data-Efficient Double-Win Lottery Tickets from Robust Pre-training

Pre-training serves as a broadly adopted starting point for transfer learning on various downstream tasks. Recent investigations of lottery tickets hypothesis (LTH) demonstrate such enormous pre-trained models can be replaced by extremely sparse subnetworks (a.k.a. matching subnetworks) without sacrificing transferability. However, practical security-crucial applications usually pose more challenging requirements beyond standard transfer, which also demand these subnetworks to overcome adversarial vulnerability. In this paper, we formulate a more rigorous concept, Double-Win Lottery Tickets, in which a located subnetwork from a pre-trained model can be independently transferred on diverse downstream tasks, to reach BOTH the same standard and robust generalization, under BOTH standard and adversarial training regimes, as the full pre-trained model can do. We comprehensively examine various pre-training mechanisms and find that robust pre-training tends to craft sparser double-win lottery tickets with superior performance over the standard counterparts. For example, on downstream CIFAR-10/100 datasets, we identify double-win matching subnetworks with the standard, fast adversarial, and adversarial pre-training from ImageNet, at 89.26%/73.79%, 89.26%/79.03%, and 91.41%/83.22% sparsity, respectively. Furthermore, we observe the obtained double-win lottery tickets can be more data-efficient to transfer, under practical data-limited (e.g., 1% and 10%) downstream schemes. Our results show that the benefits from robust pre-training are amplified by the lottery ticket scheme, as well as the data-limited transfer setting. Codes are available at https://github.com/VITA-Group/Double-Win-LTH.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Electromagnetic Dalitz Decays of $D_{(s)}^\ast$ Mesons

Rare electromagnetic decays of charmed mesons are useful laboratories to explore the structure of hadronic states and the interactions between photon and charmed mesons, to test the chiral perturbation theory in flavor sector and to search for new physics including dark photons. In this paper, we calculate the relative branching ratios of electromagnetic Dalitz decays $D_{(s)}^\ast\to D_{(s)}\ell^+\ell^-$ to their corresponding radiative decays $D_{(s)}^\ast\to D_{(s)}γ$, the dileptonic invariant mass spectra and the leptonic angular distributions with transition form factor in Vector-Meson Dominance model, where $D_{(s)}^\ast$ represents $D^\ast(2007)^0$, $D^\ast(2010)^\pm$, $D^\ast(2640)^\pm$, $D_s^{\ast\pm}$, $D_{s1}^\ast(2700)^\pm$ and $D_{s1}^\ast(2860)^\pm$.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Josephson-Coulomb drag effect between graphene and LaAlO3/SrTiO3 interfacial superconductor

Coulomb drag refers to the phenomenon that a charge current in one electronic circuit induces a responsive current in a neighboring circuit solely through Coulomb interactions. For conventional interactions between fermionic particles such as electrons, the as-induced drag current in the passive layer is orders of magnitude weaker than the active current due to strong dielectric screening effect between the two. Here we propose a 'super' Coulomb drag effect between an active normal conductor and a passive superconductor of Josephson junction arrays, whereby the passive current can greatly exceed the active. The drag force originates from the interactions between the substantially enhanced dynamical quantum fluctuations of the superconducting phases in the passive layer and the normal electrons in the active layer. We demonstrate this effect in the devices composed of monolayer graphene and LaAlO3/SrTiO3 heterointerface, an inherently non-uniform superconductor described by Josephson junction arrays. Remarkable drag signal is observed in the superconducting transition regime of the LaAlO3/SrTiO3 interface, with its sign independent of the carrier type in the graphene layer. The estimated passive-to-active ratio can reach about 0.3 at the optimal gate voltage and the temperature dependence follows that of the typical Josephson energy between superconducting puddles. Strikingly, the ratio ought to be as large as 10^5 at zero temperature by theoretical extrapolation. From engineering perspective, our device may work as current or voltage transformers, and the drag mechanism lays the foundation for synchronizing Josephson-junction-array-based terahertz radiators.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Label Anchored Contrastive Learning for Language Understanding

Contrastive learning (CL) has achieved astonishing progress in computer vision, speech, and natural language processing fields recently with self-supervised learning. However, CL approach to the supervised setting is not fully explored, especially for the natural language understanding classification task. Intuitively, the class label itself has the intrinsic ability to perform hard positive/negative mining, which is crucial for CL. Motivated by this, we propose a novel label anchored contrastive learning approach (denoted as LaCon) for language understanding. Specifically, three contrastive objectives are devised, including a multi-head instance-centered contrastive loss (ICL), a label-centered contrastive loss (LCL), and a label embedding regularizer (LER). Our approach does not require any specialized network architecture or any extra data augmentation, thus it can be easily plugged into existing powerful pre-trained language models. Compared to the state-of-the-art baselines, LaCon obtains up to 4.1% improvement on the popular datasets of GLUE and CLUE benchmarks. Besides, LaCon also demonstrates significant advantages under the few-shot and data imbalance settings, which obtains up to 9.4% improvement on the FewGLUE and FewCLUE benchmarking tasks.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Layout-Aware Information Extraction for Document-Grounded Dialogue: Dataset, Method and Demonstration

Building document-grounded dialogue systems have received growing interest as documents convey a wealth of human knowledge and commonly exist in enterprises. Wherein, how to comprehend and retrieve information from documents is a challenging research problem. Previous work ignores the visual property of documents and treats them as plain text, resulting in incomplete modality. In this paper, we propose a Layout-aware document-level Information Extraction dataset, LIE, to facilitate the study of extracting both structural and semantic knowledge from visually rich documents (VRDs), so as to generate accurate responses in dialogue systems. LIE contains 62k annotations of three extraction tasks from 4,061 pages in product and official documents, becoming the largest VRD-based information extraction dataset to the best of our knowledge. We also develop benchmark methods that extend the token-based language model to consider layout features like humans. Empirical results show that layout is critical for VRD-based extraction, and system demonstration also verifies that the extracted knowledge can help locate the answers that users care about.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Linearity Grafting: Relaxed Neuron Pruning Helps Certifiable Robustness

Certifiable robustness is a highly desirable property for adopting deep neural networks (DNNs) in safety-critical scenarios, but often demands tedious computations to establish. The main hurdle lies in the massive amount of non-linearity in large DNNs. To trade off the DNN expressiveness (which calls for more non-linearity) and robustness certification scalability (which prefers more linearity), we propose a novel solution to strategically manipulate neurons, by "grafting" appropriate levels of linearity. The core of our proposal is to first linearize insignificant ReLU neurons, to eliminate the non-linear components that are both redundant for DNN performance and harmful to its certification. We then optimize the associated slopes and intercepts of the replaced linear activations for restoring model performance while maintaining certifiability. Hence, typical neuron pruning could be viewed as a special case of grafting a linear function of the fixed zero slopes and intercept, that might overly restrict the network flexibility and sacrifice its performance. Extensive experiments on multiple datasets and network backbones show that our linearity grafting can (1) effectively tighten certified bounds; (2) achieve competitive certifiable robustness without certified robust training (i.e., over 30% improvements on CIFAR-10 models); and (3) scale up complete verification to large adversarially trained models with 17M parameters. Codes are available at https://github.com/VITA-Group/Linearity-Grafting.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Multi-Modal Masked Pre-Training for Monocular Panoramic Depth Completion

In this paper, we formulate a potentially valuable panoramic depth completion (PDC) task as panoramic 3D cameras often produce 360° depth with missing data in complex scenes. Its goal is to recover dense panoramic depths from raw sparse ones and panoramic RGB images. To deal with the PDC task, we train a deep network that takes both depth and image as inputs for the dense panoramic depth recovery. However, it needs to face a challenging optimization problem of the network parameters due to its non-convex objective function. To address this problem, we propose a simple yet effective approach termed M{^3}PT: multi-modal masked pre-training. Specifically, during pre-training, we simultaneously cover up patches of the panoramic RGB image and sparse depth by shared random mask, then reconstruct the sparse depth in the masked regions. To our best knowledge, it is the first time that we show the effectiveness of masked pre-training in a multi-modal vision task, instead of the single-modal task resolved by masked autoencoders (MAE). Different from MAE where fine-tuning completely discards the decoder part of pre-training, there is no architectural difference between the pre-training and fine-tuning stages in our M$^{3}$PT as they only differ in the prediction density, which potentially makes the transfer learning more convenient and effective. Extensive experiments verify the effectiveness of M{^3}PT on three panoramic datasets. Notably, we improve the state-of-the-art baselines by averagely 26.2% in RMSE, 51.7% in MRE, 49.7% in MAE, and 37.5% in RMSElog on three benchmark datasets.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

NTIRE 2021 Challenge on Quality Enhancement of Compressed Video: Methods and Results

This paper reviews the first NTIRE challenge on quality enhancement of compressed video, with a focus on the proposed methods and results. In this challenge, the new Large-scale Diverse Video (LDV) dataset is employed. The challenge has three tracks. Tracks 1 and 2 aim at enhancing the videos compressed by HEVC at a fixed QP, while Track 3 is designed for enhancing the videos compressed by x265 at a fixed bit-rate. Besides, the quality enhancement of Tracks 1 and 3 targets at improving the fidelity (PSNR), and Track 2 targets at enhancing the perceptual quality. The three tracks totally attract 482 registrations. In the test phase, 12 teams, 8 teams and 11 teams submitted the final results of Tracks 1, 2 and 3, respectively. The proposed methods and solutions gauge the state-of-the-art of video quality enhancement. The homepage of the challenge: https://github.com/RenYang-home/NTIRE21_VEnh

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Quarantine: Sparsity Can Uncover the Trojan Attack Trigger for Free

Trojan attacks threaten deep neural networks (DNNs) by poisoning them to behave normally on most samples, yet to produce manipulated results for inputs attached with a particular trigger. Several works attempt to detect whether a given DNN has been injected with a specific trigger during the training. In a parallel line of research, the lottery ticket hypothesis reveals the existence of sparse subnetworks which are capable of reaching competitive performance as the dense network after independent training. Connecting these two dots, we investigate the problem of Trojan DNN detection from the brand new lens of sparsity, even when no clean training data is available. Our crucial observation is that the Trojan features are significantly more stable to network pruning than benign features. Leveraging that, we propose a novel Trojan network detection regime: first locating a "winning Trojan lottery ticket" which preserves nearly full Trojan information yet only chance-level performance on clean inputs; then recovering the trigger embedded in this already isolated subnetwork. Extensive experiments on various datasets, i.e., CIFAR-10, CIFAR-100, and ImageNet, with different network architectures, i.e., VGG-16, ResNet-18, ResNet-20s, and DenseNet-100 demonstrate the effectiveness of our proposal. Codes are available at https://github.com/VITA-Group/Backdoor-LTH.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

RigNet: Repetitive Image Guided Network for Depth Completion

Depth completion deals with the problem of recovering dense depth maps from sparse ones, where color images are often used to facilitate this task. Recent approaches mainly focus on image guided learning frameworks to predict dense depth. However, blurry guidance in the image and unclear structure in the depth still impede the performance of the image guided frameworks. To tackle these problems, we explore a repetitive design in our image guided network to gradually and sufficiently recover depth values. Specifically, the repetition is embodied in both the image guidance branch and depth generation branch. In the former branch, we design a repetitive hourglass network to extract discriminative image features of complex environments, which can provide powerful contextual instruction for depth prediction. In the latter branch, we introduce a repetitive guidance module based on dynamic convolution, in which an efficient convolution factorization is proposed to simultaneously reduce its complexity and progressively model high-frequency structures. Extensive experiments show that our method achieves superior or competitive results on KITTI benchmark and NYUv2 dataset.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Sparsity Winning Twice: Better Robust Generalization from More Efficient Training

Recent studies demonstrate that deep networks, even robustified by the state-of-the-art adversarial training (AT), still suffer from large robust generalization gaps, in addition to the much more expensive training costs than standard training. In this paper, we investigate this intriguing problem from a new perspective, i.e., injecting appropriate forms of sparsity during adversarial training. We introduce two alternatives for sparse adversarial training: (i) static sparsity, by leveraging recent results from the lottery ticket hypothesis to identify critical sparse subnetworks arising from the early training; (ii) dynamic sparsity, by allowing the sparse subnetwork to adaptively adjust its connectivity pattern (while sticking to the same sparsity ratio) throughout training. We find both static and dynamic sparse methods to yield win-win: substantially shrinking the robust generalization gap and alleviating the robust overfitting, meanwhile significantly saving training and inference FLOPs. Extensive experiments validate our proposals with multiple network architectures on diverse datasets, including CIFAR-10/100 and Tiny-ImageNet. For example, our methods reduce robust generalization gap and overfitting by 34.44% and 4.02%, with comparable robust/standard accuracy boosts and 87.83%/87.82% training/inference FLOPs savings on CIFAR-100 with ResNet-18. Besides, our approaches can be organically combined with existing regularizers, establishing new state-of-the-art results in AT. Codes are available in https://github.com/VITA-Group/Sparsity-Win-Robust-Generalization.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Study of exotic hadrons with machine learning

We analyzed the invariant mass spectrum of near-threshold exotic states for one-channel candidates with a deep neural network. It can extract the scattering length and effective range, which would shed light on the nature of given states, from the experimental mass spectrum. As an application, the mass spectrum of the $X(3872)$ and the $T_{cc}^+$ are studied. The obtained scattering lengths, effective ranges, and most relevant thresholds are consistent with those from fitting to the experimental data. The advantage of the neural network is that it is more stable than the fitting, especially for low-statistic data. The network, which provides another way to analyze the experimental data, can also be applied to other one-channel near-threshold exotic candidates.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

The Principle of Diversity: Training Stronger Vision Transformers Calls for Reducing All Levels of Redundancy

Vision transformers (ViTs) have gained increasing popularity as they are commonly believed to own higher modeling capacity and representation flexibility, than traditional convolutional networks. However, it is questionable whether such potential has been fully unleashed in practice, as the learned ViTs often suffer from over-smoothening, yielding likely redundant models. Recent works made preliminary attempts to identify and alleviate such redundancy, e.g., via regularizing embedding similarity or re-injecting convolution-like structures. However, a "head-to-toe assessment" regarding the extent of redundancy in ViTs, and how much we could gain by thoroughly mitigating such, has been absent for this field. This paper, for the first time, systematically studies the ubiquitous existence of redundancy at all three levels: patch embedding, attention map, and weight space. In view of them, we advocate a principle of diversity for training ViTs, by presenting corresponding regularizers that encourage the representation diversity and coverage at each of those levels, that enabling capturing more discriminative information. Extensive experiments on ImageNet with a number of ViT backbones validate the effectiveness of our proposals, largely eliminating the observed ViT redundancy and significantly boosting the model generalization. For example, our diversified DeiT obtains 0.70%~1.76% accuracy boosts on ImageNet with highly reduced similarity. Our codes are fully available in https://github.com/VITA-Group/Diverse-ViT.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

JUNO Physics and Detector

The Jiangmen Underground Neutrino Observatory (JUNO) is a 20 kton LS detector at 700-m underground. An excellent energy resolution and a large fiducial volume offer exciting opportunities for addressing many important topics in neutrino and astro-particle physics. With 6 years of data, the neutrino mass ordering can be determined at 3-4 sigma and three oscillation parameters can be measured to a precision of 0.6% or better by detecting reactor antineutrinos. With 10 years of data, DSNB could be observed at 3-sigma; a lower limit of the proton lifetime of 8.34e33 years (90% C.L.) can be set by searching for p->nu_bar K^+; detection of solar neutrinos would shed new light on the solar metallicity problem and examine the vacuum-matter transition region. A core-collapse supernova at 10 kpc would lead to ~5000 IBD and ~2000 (300) all-flavor neutrino-proton (electron) scattering events. Geo-neutrinos can be detected with a rate of ~400 events/year. We also summarize the final design of the JUNO detector and the key R&D achievements. All 20-inch PMTs have been tested. The average photon detection efficiency is 28.9% for the 15,000 MCP PMTs and 28.1% for the 5,000 dynode PMTs, higher than the JUNO requirement of 27%. Together with the >20 m attenuation length of LS, we expect a yield of 1345 p.e. per MeV and an effective energy resolution of 3.02%/\sqrt{E (MeV)}$ in simulations. The underwater electronics is designed to have a loss rate <0.5% in 6 years. With degassing membranes and a micro-bubble system, the radon concentration in the 35-kton water pool could be lowered to <10 mBq/m^3. Acrylic panels of radiopurity <0.5 ppt U/Th are produced. The 20-kton LS will be purified onsite. Singles in the fiducial volume can be controlled to ~10 Hz. The JUNO experiment also features a double calorimeter system with 25,600 3-inch PMTs, a LS testing facility OSIRIS, and a near detector TAO.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Magnetic moment preservation and emergent Kondo resonance of Co-phthalocyanine on semimetallic Sb(111)

Magnetic molecules on surfaces have been widely investigated to reveal delicate interfacial couplings and for potential technological applications. In these endeavors, one prevailing challenge is how to preserve or recover the molecular spins, especially on highly metallic substrates that can readily quench the magnetic moments of the ad-molecules. Here we use scanning tunneling microscopy/spectroscopy to exploit the semimetallic nature of antimony and observe, surprisingly yet pleasantly, that the spin of Co-phthalocyanine is well preserved on Sb(111), as unambiguously evidenced by the emergent strong Kondo resonance across the molecule. Our first-principles calculations further confirm that the optimal density of states near the Fermi level of the semimetal is a decisive factor, weakening the overall interfacial coupling, while still ensuring sufficiently effective electron-spin scattering in the many-body system. Beyond isolated ad-molecules, we discover that each of the magnetic moments in a molecular dimer or a densely packed island is distinctly preserved as well, rendering such molecular magnets immense potentials for ultra-high density memory devices.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Artificial Intelligence for High-Throughput Discovery of Topological Insulators: the Example of Alloyed Tetradymites

Significant advances have been made in predicting new topological materials using high-throughput empirical descriptors or symmetry-based indicators. To date, these approaches have been applied to materials in existing databases, and are severely limited to systems with well-defined symmetries, leaving a much larger materials space unexplored. Using tetradymites as a prototypical class of examples, we uncover a novel two-dimensional descriptor by applying an artificial intelligence (AI) based approach for fast and reliable identification of the topological characters of a drastically expanded range of materials, without prior determination of their specific symmetries and detailed band structures. By leveraging this descriptor that contains only the atomic number and electronegativity of the constituent species, we have readily scanned a huge number of alloys in the tetradymite family. Strikingly, nearly half of which are identified to be topological insulators, revealing a much larger territory of the topological materials world. The present work also attests the increasingly important role of such AI-based approaches in modern materials discovery.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Feasibility and physics potential of detecting $^8$B solar neutrinos at JUNO

The Jiangmen Underground Neutrino Observatory~(JUNO) features a 20~kt multi-purpose underground liquid scintillator sphere as its main detector. Some of JUNO&#39;s features make it an excellent experiment for $^8$B solar neutrino measurements, such as its low-energy threshold, its high energy resolution compared to water Cherenkov detectors, and its much large target mass compared to previous liquid scintillator detectors. In this paper we present a comprehensive assessment of JUNO&#39;s potential for detecting $^8$B solar neutrinos via the neutrino-electron elastic scattering process. A reduced 2~MeV threshold on the recoil electron energy is found to be achievable assuming the intrinsic radioactive background $^{238}$U and $^{232}$Th in the liquid scintillator can be controlled to 10$^{-17}$~g/g. With ten years of data taking, about 60,000 signal and 30,000 background events are expected. This large sample will enable an examination of the distortion of the recoil electron spectrum that is dominated by the neutrino flavor transformation in the dense solar matter, which will shed new light on the tension between the measured electron spectra and the predictions of the standard three-flavor neutrino oscillation framework. If $Δm^{2}_{21}=4.8\times10^{-5}~(7.5\times10^{-5})$~eV$^{2}$, JUNO can provide evidence of neutrino oscillation in the Earth at the about 3$σ$~(2$σ$) level by measuring the non-zero signal rate variation with respect to the solar zenith angle. Moveover, JUNO can simultaneously measure $Δm^2_{21}$ using $^8$B solar neutrinos to a precision of 20\% or better depending on the central value and to sub-percent precision using reactor antineutrinos. A comparison of these two measurements from the same detector will help elucidate the current tension between the value of $Δm^2_{21}$ reported by solar neutrino experiments and the KamLAND experiment.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

HIN: Hierarchical Inference Network for Document-Level Relation Extraction

Document-level RE requires reading, inferring and aggregating over multiple sentences. From our point of view, it is necessary for document-level RE to take advantage of multi-granularity inference information: entity level, sentence level and document level. Thus, how to obtain and aggregate the inference information with different granularity is challenging for document-level RE, which has not been considered by previous work. In this paper, we propose a Hierarchical Inference Network (HIN) to make full use of the abundant information from entity level, sentence level and document level. Translation constraint and bilinear transformation are applied to target entity pair in multiple subspaces to get entity-level inference information. Next, we model the inference between entity-level information and sentence representation to achieve sentence-level inference information. Finally, a hierarchical aggregation approach is adopted to obtain the document-level inference information. In this way, our model can effectively aggregate inference information from these three different granularities. Experimental results show that our method achieves state-of-the-art performance on the large-scale DocRED dataset. We also demonstrate that using BERT representations can further substantially boost the performance.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Joint Extraction of Entities and Relations Based on a Novel Decomposition Strategy

Joint extraction of entities and relations aims to detect entity pairs along with their relations using a single model. Prior work typically solves this task in the extract-then-classify or unified labeling manner. However, these methods either suffer from the redundant entity pairs, or ignore the important inner structure in the process of extracting entities and relations. To address these limitations, in this paper, we first decompose the joint extraction task into two interrelated subtasks, namely HE extraction and TER extraction. The former subtask is to distinguish all head-entities that may be involved with target relations, and the latter is to identify corresponding tail-entities and relations for each extracted head-entity. Next, these two subtasks are further deconstructed into several sequence labeling problems based on our proposed span-based tagging scheme, which are conveniently solved by a hierarchical boundary tagger and a multi-span decoding algorithm. Owing to the reasonable decomposition strategy, our model can fully capture the semantic interdependency between different steps, as well as reduce noise from irrelevant entity pairs. Experimental results show that our method outperforms previous work by 5.2%, 5.9% and 21.5% (F1 score), achieving a new state-of-the-art on three public datasets

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Residual Clipping Noise in Multi-layer Optical OFDM: Modeling, Analysis, and Application

Optical orthogonal frequency division multiplexing (O-OFDM) schemes are variations of OFDM schemes which produce non-negative signals. Asymmetrically-clipped O-OFDM (ACO-OFDM) is a single-layer O-OFDM scheme, whose spectral efficiency can be enhanced by adopting multiple ACO-OFDM layers or a combination of ACO-OFDM and other O-OFDM schemes. However, since symbol detection in such enhanced ACO-OFDM (eACO-OFDM) is done iteratively, erroneous detection leads to residual clipping noise (RCN) which can degrade performance in practice. Thus, it is important to develop an accurate model for RCN which can be used to design RCN-aware eACO-OFDM schemes. To this end, this paper provides a mathematical analysis of RCN leading to an accurate model of RCN power. The obtained model is used to analyze the performance of various eACO-OFDM schemes. It is shown that the model provides an accurate evaluation of symbol error rate (SER), which would be underestimated if RCN is ignored. Moreover, the model is shown to be useful for designing an RCN-aware resource allocation that increases the robustness of the system in terms of meeting a target SER, compared to an RCN-unaware design.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

TAO Conceptual Design Report: A Precision Measurement of the Reactor Antineutrino Spectrum with Sub-percent Energy Resolution

The Taishan Antineutrino Observatory (TAO, also known as JUNO-TAO) is a satellite experiment of the Jiangmen Underground Neutrino Observatory (JUNO). A ton-level liquid scintillator detector will be placed at about 30 m from a core of the Taishan Nuclear Power Plant. The reactor antineutrino spectrum will be measured with sub-percent energy resolution, to provide a reference spectrum for future reactor neutrino experiments, and to provide a benchmark measurement to test nuclear databases. A spherical acrylic vessel containing 2.8 ton gadolinium-doped liquid scintillator will be viewed by 10 m^2 Silicon Photomultipliers (SiPMs) of >50% photon detection efficiency with almost full coverage. The photoelectron yield is about 4500 per MeV, an order higher than any existing large-scale liquid scintillator detectors. The detector operates at -50 degree C to lower the dark noise of SiPMs to an acceptable level. The detector will measure about 2000 reactor antineutrinos per day, and is designed to be well shielded from cosmogenic backgrounds and ambient radioactivities to have about 10% background-to-signal ratio. The experiment is expected to start operation in 2022.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Large-area, periodic, and tunable intrinsic pseudo-magnetic fields in low-angle twisted bilayer graphene

A properly strained graphene monolayer or bilayer is expected to harbour periodic pseudo-magnetic fields with high symmetry, yet to date, a convincing demonstration of such pseudo-magnetic fields has been lacking, especially for bilayer graphene. Here, we report the first definitive experimental proof for the existence of large-area, periodic pseudo-magnetic fields, as manifested by vortex lattices in commensurability with the moiré patterns of low-angle twisted bilayer graphene. The pseudo-magnetic fields are strong enough to confine the massive Dirac electrons into circularly localized pseudo-Landau levels, as observed by scanning tunneling microscopy/spectroscopy, and also corroborated by tight-binding calculations. We further demonstrate that the geometry, amplitude, and periodicity of the pseudo-magnetic field can be fine-tuned by both the rotation angle and heterostrain applied to the system. Collectively, the present study substantially enriches twisted bilayer graphene as a powerful enabling platform for exploration of new and exotic physical phenomena, including quantum valley Hall effects and quantum anomalous Hall effects.

0 citations0 reviewsNo code linkedOpen access