Source author record

Shuo Yang

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

ResearcherUnclaimed source record

Catalog footprint

What is connected

78works

29topics

4close collaborators

Actions

Connect this record

Open graph Browse works

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.

preprint2026arXiv

Active Tabular Augmentation via Policy-Guided Diffusion Inpainting

Generative tabular augmentation is appealing in data-scarce domains, yet the prevailing focus on distributional fidelity does not reliably translate into better downstream models. We formalize a fidelity-utility gap: common generative objectives prioritize distributional plausibility, whereas augmentation succeeds only when injected samples reduce the current learner's held-out evaluation loss. This gap motivates learning not just how to generate, but what to generate and when to inject as training evolves. We propose TAP (Tabular Augmentation Policy), which couples diffusion inpainting with a lightweight, learner-conditioned policy to steer generation toward high-utility regions and controls safe injection via explicit gating and conservative windowed commitment. Under severe data scarcity, TAP consistently outperforms strong generative baselines on seven real-world datasets, improving classification accuracy by up to 15.6 percentage points and reducing regression RMSE by up to 32%.

preprint2026arXiv

AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin

Fine-tuning large language models (LLMs) improves performance but introduces critical safety vulnerabilities: even minimal harmful data can severely compromise safety measures. We observe that perturbations orthogonal to the alignment direction - defined by weight differences between aligned (safe) and unaligned models - rapidly compromise model safety. In contrast, updates along the alignment direction largely preserve it, revealing the parameter space as a "narrow safety basin". To address this, we propose AsFT (Anchoring Safety in Fine-Tuning) to maintain safety by explicitly constraining update directions during fine-tuning. By penalizing updates orthogonal to the alignment direction, AsFT effectively constrains the model within the "narrow safety basin," thus preserving its inherent safety. Extensive experiments on multiple datasets and models show that AsFT reduces harmful behaviors by up to 7.60%, improves task performance by 3.44%, and consistently outperforms existing methods across multiple tasks.

preprint2026arXiv

Belief-Guided Inference Control for Large Language Model Services via Verifiable Observations

In black-box large language model (LLM) services, response reliability is often only partially observable at decision time, while stronger inference pathways incur substantial computational cost, inducing a budgeted sequential decision problem: for each request, the system should decide whether the default low-cost response is sufficiently reliable or whether additional computation should be allocated to improve response quality. In this paper, we propose \textbf{Ver}ifiable \textbf{O}bservations for Risk-aware \textbf{I}nference \textbf{C}ontrol (\textsc{Veroic}), a framework for adaptive inference control in black-box LLM settings, which formulates request-time control as a \textit{partially observable Markov decision process} to capture partial observability and sequential budget coupling. It constructs a lightweight verifiable observation channel from the input-output pair by aggregating heterogeneous quality signals into a belief state over latent response reliability, which is then used by a budget-aware policy to decide whether to return the default output or trigger a higher-cost inference pathway. Experiments on diverse tasks show that \textsc{Veroic} achieves improved quality-cost trade-offs, stronger risk estimation and calibration, and more robust long-horizon inference control than competitive baselines.

preprint2026arXiv

Bian Que: An Agentic Framework with Flexible Skill Arrangement for Online System Operations

Operating and maintaining (O&M) large-scale online engine systems (eg, search, recommendation and advertising) demands substantial human effort for release monitoring, alert response, and root cause analysis. Despite the inherent suitability of LLM-based agents for such operational scenarios, the critical bottleneck impeding their practical deployment lies not in reasoning, but in orchestration capability - specifically, the precise selection of relevant data (encompassing metrics, logs, and change events) and applicable knowledge (including handbook-defined rules and empirically derived practitioner experience) tailored to each individual operational event. Feeding all signals indiscriminately causes dilution and hallucination, while manually curating the event-to-(data, knowledge) mapping is intractable under dozens of daily releases. Here we present Bian Que, an agentic operating framework with three contributions: (i) The unified operational paradigm, which abstracts routine daily O&M actions into three canonical patterns: release interception, proactive inspection, and alert root cause analysis; (ii) The flexible Skill Arrangement, each predefined Skill explicitly defines the requisite data and operational knowledge for each specific context. Such Skills can be automatically generated and updated by LLM agents, and can also be iteratively optimized by on-call engineers via natural language instructions. (iii) The unified self-evolving mechanism, where each correction signal enables two parallel evolutionary pathways: distilling event memory into knowledge, and targeted refinement of Skills. Deployed on the e-commerce search engine of KuaiShou, Bian Que reduces alert volume by 75%, achieves 80% root-cause analysis accuracy, cuts mean time to resolution by over 50%, and attains a 99.0% pass rate on offline evaluations. Codes are at https://github.com/benchen4395/BianQue_Assistant.

preprint2026arXiv

Exploring Data-Free LoRA Transferability for Video Diffusion Models

Video diffusion models leveraging step distillation or causal distillation have achieved remarkable performance. However, adapting existing LoRAs to these variants remains a critical challenge due to weight space mismatches. We observe that direct application leads to style degradation and structural collapse, yet the underlying mechanisms remain poorly understood. To fill this gap, we delve into the weight space and identify that the incompatibility stems from spectral interference within shared functional clusters defined over singular subspaces. Specifically, our analysis reveals that while both paradigms respect spectral rigidity, they establish conflicting routing pathways that clash through constructive overload or destructive cancellation. To address this issue, we propose Cluster-Aware Spectral Arbitration (CASA), a data-free framework that dynamically arbitrates between safeguarding the target's manifold and restoring LoRA alignment based on spectral density. Extensive experiments demonstrate that CASA effectively mitigates artifacts and revives LoRA functionality. Our code is available at https://github.com/Noahwangyuchen/CASA

preprint2026arXiv

LLM Agents Enable User-Governed Personalization Beyond Platform Boundaries

Personalization today is fundamentally platform-centric: services build user representations from the behavioral fragments they observe. Yet no platform can construct a complete picture of the user, as competitive incentives, legal constraints, user privacy concerns, and epistemic limits create persistent data barriers. This paper argues for a shift from platform-centric personalization to user-governed personalization, where only the user can integrate fragmented contexts across platforms and the offline world. The key asymmetry lies in data access: only users can aggregate their own cross-platform and offline information. Large language model (LLM) agents make such integration practically feasible for the first time by enabling reasoning over heterogeneous personal data and transforming users' cross-context information into actionable personalization capabilities. We provide proof-of-concept evidence that users equipped with cross-platform data exports and an off-the-shelf LLM agent can outperform single-platform personalization baselines. We conclude by outlining a research agenda for building scalable user-governed personalization systems.

preprint2026arXiv

Matrix Manifold Neural Networks++

Deep neural networks (DNNs) on Riemannian manifolds have garnered increasing interest in various applied areas. For instance, DNNs on spherical and hyperbolic manifolds have been designed to solve a wide range of computer vision and nature language processing tasks. One of the key factors that contribute to the success of these networks is that spherical and hyperbolic manifolds have the rich algebraic structures of gyrogroups and gyrovector spaces. This enables principled and effective generalizations of the most successful DNNs to these manifolds. Recently, some works have shown that many concepts in the theory of gyrogroups and gyrovector spaces can also be generalized to matrix manifolds such as Symmetric Positive Definite (SPD) and Grassmann manifolds. As a result, some building blocks for SPD and Grassmann neural networks, e.g., isometric models and multinomial logistic regression (MLR) can be derived in a way that is fully analogous to their spherical and hyperbolic counterparts. Building upon these works, we design fully-connected (FC) and convolutional layers for SPD neural networks. We also develop MLR on Symmetric Positive Semi-definite (SPSD) manifolds, and propose a method for performing backpropagation with the Grassmann logarithmic map in the projector perspective. We demonstrate the effectiveness of the proposed approach in the human action recognition and node classification tasks.

preprint2026arXiv

Neural Networks on Symmetric Spaces of Noncompact Type

Recent works have demonstrated promising performances of neural networks on hyperbolic spaces and symmetric positive definite (SPD) manifolds. These spaces belong to a family of Riemannian manifolds referred to as symmetric spaces of noncompact type. In this paper, we propose a novel approach for developing neural networks on such spaces. Our approach relies on a unified formulation of the distance from a point to a hyperplane on the considered spaces. We show that some existing formulations of the point-to-hyperplane distance can be recovered by our approach under specific settings. Furthermore, we derive a closed-form expression for the point-to-hyperplane distance in higher-rank symmetric spaces of noncompact type equipped with G-invariant Riemannian metrics. The derived distance then serves as a tool to design fully-connected (FC) layers and an attention mechanism for neural networks on the considered spaces. Our approach is validated on challenging benchmarks for image classification, electroencephalogram (EEG) signal classification, image generation, and natural language inference.

preprint2026arXiv

OCR-Memory: Optical Context Retrieval for Long-Horizon Agent Memory

Autonomous LLM agents increasingly operate in long-horizon, interactive settings where success depends on reusing experience accumulated over extended histories. However, existing agent memory systems are fundamentally constrained by text-context budgets: storing or revisiting raw trajectories is prohibitively token-expensive, while summarization and text-only retrieval trade token savings for information loss and fragmented evidence. To address this limitation, we propose Optical Context Retrieval Memory (OCR-Memory), a memory framework that leverages the visual modality as a high-density representation of agent experience, enabling retention of arbitrarily long histories with minimal prompt overhead at retrieval time. Specifically, OCR-Memory renders historical trajectories into images annotated with unique visual identifiers. OCR-Memory retrieves stored experience via a \emph{locate-and-transcribe} paradigm that selects relevant regions through visual anchors and retrieves the corresponding verbatim text, avoiding free-form generation and reducing hallucination. Experiments on long-horizon agent benchmarks show consistent gains under strict context limits, demonstrating that optical encoding increases effective memory capacity while preserving faithful evidence recovery.

preprint2026arXiv

Reasoning Portability: Guiding Continual Learning for MLLMs in the RLVR Era

Vision-Language Models in Continual Learning (VLM-CL) aim to continuously adapt to new multimodal tasks while retaining prior knowledge. The emerging paradigm that couples Multimodal Large Language Models (MLLMs) with Reinforcement Learning with Verifiable Rewards (RLVR) calls for a new pattern to guide continual adaptation. Advances in reasoning capability now make it feasible to impose constraints at the reasoning level. We formalize portability, a sample-level measure of how reusable the previous policy's behavior is on a new task, and empirically show that reasoning-level signals remain reliable on out-of-distribution samples while answer-level signals do not. We instantiate this as Reasoning Portability (RP) and propose Reasoning-based Dynamic Balance Continual Learning (RDB-CL), which modulates the per-sample Kullback-Leibler regularization in RLVR according to RP: a tight anchor preserves reusable reasoning on high-RP samples, while a relaxed anchor on low-RP samples permits exploration of new reasoning pathways. Experiments show that RDB-CL consistently outperforms baselines, improving Last accuracy by +12.0% over the vanilla RLVR baseline.

preprint2026arXiv

Searching for long-lived axion-like particles via displaced vertices at the HL-LHC

Axion-like particles (ALPs) are well-motivated extensions of the standard model (SM) that appear in numerous new physics scenarios. In this paper, we concentrate on searches for long-lived ALPs predicted by the photophobic scenario at the HL-LHC with the center-of-mass energy $\sqrt{s}=14$ TeV and the integrated luminosity $\mathcal{L}=$ $3$ ab$^{-1}$. We consider the process $pp \to γa$ with the ALP $a$ decaying into a pair of displaced charged leptons and perform a detailed analysis of two types of signals: $π^+ π^- γE\mkern-10.5 mu/_T $ and $\ell^+ \ell^- γE\mkern-10.5 mu/_T $. For the $π^+ π^- γE\mkern-10.5 mu/_T $ signal, we find that the prospective sensitivities of the HL-LHC can reach $g_{aWW}\in [8.72 \times 10^{-3}, 6.42 \times 10^{-2}]$ TeV$^{-1}$ for the ALP mass $m_a \in [4, 10]$ GeV. While for the $\ell^+ \ell^- γE\mkern-10.5 mu/_T $ signal, the HL-LHC can probe a broader parameter space, with the sensitivities covering $m_a \in [4, 10]$ GeV and $g_{aWW} \in [4.17 \times 10^{-3}, 2.00 \times 10^{-1}]$ TeV$^{-1}$. These long-lived searches complement some previous prompt decay studies from LEP and LHC experiments, extending the parameter space explored by the LHCb collaboration. Our results show that the HL-LHC has significant potential to probe long-lived ALPs via their displaced vertex signatures.

preprint2026arXiv

SuperFace: Preference-Aligned Facial Expression Estimation Beyond Pseudo Supervision

Accurate facial estimation is crucial for realistic digital human animation, and ARKit blendshape coefficients offer an interpretable representation by mapping facial motions to semantic animation controls. However, learning high-quality ARKit coefficient prediction remains limited by the absence of reliable ground-truth supervision. Existing methods typically rely on capture software such as Live Link Face to provide pseudo labels, which may contain noisy activations, biased coefficient magnitudes, and missing or inaccurate facial actions. Consequently, models trained with supervised learning tend to reproduce imperfect pseudo labels rather than optimize for perceptual expression fidelity. In this paper, we propose SuperFace, a preference-driven framework that moves ARKit facial expression estimation from pseudo-label imitation toward human-aligned perceptual optimization. Instead of treating software-estimated coefficients as fixed ground truth, SuperFace uses them only as an initialization and further improves coefficient prediction through human preference feedback on rendered facial expressions. By aligning the model with perceptual judgments rather than numerical pseudo labels, SuperFace enables more visually faithful and expressive facial animation. Experiments show that SuperFace improves expression fidelity over Live Link Face supervision, demonstrating the effectiveness of preference-driven optimization for semantic facial action prediction.

preprint2026arXiv

Talk2Move: Reinforcement Learning for Text-Instructed Object-Level Geometric Transformation in Scenes

We introduce Talk2Move, a reinforcement learning (RL) based diffusion framework for text-instructed spatial transformation of objects within scenes. Spatially manipulating objects in a scene through natural language poses a challenge for multimodal generation systems. While existing text-based manipulation methods can adjust appearance or style, they struggle to perform object-level geometric transformations-such as translating, rotating, or resizing objects-due to scarce paired supervision and pixel-level optimization limits. Talk2Move employs Group Relative Policy Optimization (GRPO) to explore geometric actions through diverse rollouts generated from input images and lightweight textual variations, removing the need for costly paired data. A spatial reward guided model aligns geometric transformations with linguistic description, while off-policy step evaluation and active step sampling improve learning efficiency by focusing on informative transformation stages. Furthermore, we design object-centric spatial rewards that evaluate displacement, rotation, and scaling behaviors directly, enabling interpretable and coherent transformations. Experiments on curated benchmarks demonstrate that Talk2Move achieves precise, consistent, and semantically faithful object transformations, outperforming existing text-guided editing approaches in both spatial accuracy and scene coherence.

preprint2023arXiv

Differentiable Safe Controller Design through Control Barrier Functions

Learning-based controllers, such as neural network (NN) controllers, can show high empirical performance but lack formal safety guarantees. To address this issue, control barrier functions (CBFs) have been applied as a safety filter to monitor and modify the outputs of learning-based controllers in order to guarantee the safety of the closed-loop system. However, such modification can be myopic with unpredictable long-term effects. In this work, we propose a safe-by-construction NN controller which employs differentiable CBF-based safety layers, and investigate the performance of safe-by-construction NN controllers in learning-based control. Specifically, two formulations of controllers are compared: one is projection-based and the other relies on our proposed set-theoretic parameterization. Both methods demonstrate improved closed-loop performance over using CBF as a separate safety filter in numerical experiments.

preprint2023arXiv

Fast Contact-Implicit Model-Predictive Control

We present a general approach for controlling robotic systems that make and break contact with their environments. Contact-implicit model predictive control (CI-MPC) generalizes linear MPC to contact-rich settings by utilizing a bi-level planning formulation with lower-level contact dynamics formulated as time-varying linear complementarity problems (LCPs) computed using strategic Taylor approximations about a reference trajectory. These dynamics enable the upper-level planning problem to reason about contact timing and forces, and generate entirely new contact-mode sequences online. To achieve reliable and fast numerical convergence, we devise a structure-exploiting interior-point solver for these LCP contact dynamics and a custom trajectory optimizer for the tracking problem. We demonstrate real-time solution rates for CI-MPC and the ability to generate and track non-periodic behaviours in hardware experiments on a quadrupedal robot. We also show that the controller is robust to model mismatch and can respond to disturbances by discovering and exploiting new contact modes across a variety of robotic systems in simulation, including a pushbot, planar hopper, planar quadruped, and planar biped.

preprint2022arXiv

An Effective Graph Learning based Approach for Temporal Link Prediction: The First Place of WSDM Cup 2022

Temporal link prediction, as one of the most crucial work in temporal graphs, has attracted lots of attention from the research area. The WSDM Cup 2022 seeks for solutions that predict the existence probabilities of edges within time spans over temporal graph. This paper introduces the solution of AntGraph, which wins the 1st place in the competition. We first analysis the theoretical upper-bound of the performance by removing temporal information, which implies that only structure and attribute information on the graph could achieve great performance. Based on this hypothesis, then we introduce several well-designed features. Finally, experiments conducted on the competition datasets show the superiority of our proposal, which achieved AUC score of 0.666 on dataset A and 0.902 on dataset B, the ablation studies also prove the efficiency of each feature. Code is publicly available at https://github.com/im0qianqian/WSDM2022TGP-AntGraph.

preprint2022arXiv

An Efficient Multitask Neural Network for Face Alignment, Head Pose Estimation and Face Tracking

While Convolutional Neural Networks (CNNs) have significantly boosted the performance of face related algorithms, maintaining accuracy and efficiency simultaneously in practical use remains challenging. The state-of-the-art methods employ deeper networks for better performance, which makes it less practical for mobile applications because of more parameters and higher computational complexity. Therefore, we propose an efficient multitask neural network, Alignment & Tracking & Pose Network (ATPN) for face alignment, face tracking and head pose estimation. Specifically, to achieve better performance with fewer layers for face alignment, we introduce a shortcut connection between shallow-layer and deep-layer features. We find the shallow-layer features are highly correspond to facial boundaries that can provide the structural information of face and it is crucial for face alignment. Moreover, we generate a cheap heatmap based on the face alignment result and fuse it with features to improve the performance of the other two tasks. Based on the heatmap, the network can utilize both geometric information of landmarks and appearance information for head pose estimation. The heatmap also provides attention clues for face tracking. The face tracking task also saves us the face detection procedure for each frame, which also significantly boost the real-time capability for video-based tasks. We experimentally validate ATPN on four benchmark datasets, WFLW, 300VW, WIDER Face and 300W-LP. The experimental results demonstrate that it achieves better performance with much less parameters and lower computational complexity compared to other light models.

preprint2022arXiv

CAFE: Learning to Condense Dataset by Aligning Features

Dataset condensation aims at reducing the network training effort through condensing a cumbersome training set into a compact synthetic one. State-of-the-art approaches largely rely on learning the synthetic data by matching the gradients between the real and synthetic data batches. Despite the intuitive motivation and promising results, such gradient-based methods, by nature, easily overfit to a biased set of samples that produce dominant gradients, and thus lack global supervision of data distribution. In this paper, we propose a novel scheme to Condense dataset by Aligning FEatures (CAFE), which explicitly attempts to preserve the real-feature distribution as well as the discriminant power of the resulting synthetic set, lending itself to strong generalization capability to various architectures. At the heart of our approach is an effective strategy to align features from the real and synthetic data across various scales, while accounting for the classification of real samples. Our scheme is further backed up by a novel dynamic bi-level optimization, which adaptively adjusts parameter updates to prevent over-/under-fitting. We validate the proposed CAFE across various datasets, and demonstrate that it generally outperforms the state of the art: on the SVHN dataset, for example, the performance gain is up to 11%. Extensive experiments and analyses verify the effectiveness and necessity of proposed designs.

preprint2022arXiv

Enhanced quantum sensing with room-temperature solid-state masers

Quantum sensing with solid-state systems finds broad applications in diverse areas ranging from material and biomedical sciences to fundamental physics. Several solid-state spin sensors have been developed, facilitating the ultra-sensitive detection of physical quantities such as magnetic and electric fields and temperature. Exploiting collective behaviour of non-interacting spins holds the promise of pushing the detection limit to even lower levels, while to date, those levels are scarcely reached due to the broadened linewidth and inefficient readout of solid-state spin ensembles. Here, we experimentally demonstrate that such drawbacks can be overcome by newly reborn maser technology at room temperature in the solid state. Owing to maser action, we observe a 4-fold reduction in the inhomogeneously broadened linewidth of a molecular spin ensemble, which is narrower than the same measured from single spins at cryogenic temperatures. The maser-based readout applied to magnetometry showcases a signal-to-noise ratio (SNR) of 30 dB for single shots. This technique would be a significant addition to the toolbox for boosting the sensitivity of solid-state ensemble spin sensors.

preprint2022arXiv

Entity-aware and Motion-aware Transformers for Language-driven Action Localization in Videos

Language-driven action localization in videos is a challenging task that involves not only visual-linguistic matching but also action boundary prediction. Recent progress has been achieved through aligning language query to video segments, but estimating precise boundaries is still under-explored. In this paper, we propose entity-aware and motion-aware Transformers that progressively localizes actions in videos by first coarsely locating clips with entity queries and then finely predicting exact boundaries in a shrunken temporal region with motion queries. The entity-aware Transformer incorporates the textual entities into visual representation learning via cross-modal and cross-frame attentions to facilitate attending action-related video clips. The motion-aware Transformer captures fine-grained motion changes at multiple temporal scales via integrating long short-term memory into the self-attention module to further improve the precision of action boundary prediction. Extensive experiments on the Charades-STA and TACoS datasets demonstrate that our method achieves better performance than existing methods.

preprint2022arXiv

Estimating Instance-dependent Bayes-label Transition Matrix using a Deep Neural Network

In label-noise learning, estimating the transition matrix is a hot topic as the matrix plays an important role in building statistically consistent classifiers. Traditionally, the transition from clean labels to noisy labels (i.e., clean-label transition matrix (CLTM)) has been widely exploited to learn a clean label classifier by employing the noisy data. Motivated by that classifiers mostly output Bayes optimal labels for prediction, in this paper, we study to directly model the transition from Bayes optimal labels to noisy labels (i.e., Bayes-label transition matrix (BLTM)) and learn a classifier to predict Bayes optimal labels. Note that given only noisy data, it is ill-posed to estimate either the CLTM or the BLTM. But favorably, Bayes optimal labels have less uncertainty compared with the clean labels, i.e., the class posteriors of Bayes optimal labels are one-hot vectors while those of clean labels are not. This enables two advantages to estimate the BLTM, i.e., (a) a set of examples with theoretically guaranteed Bayes optimal labels can be collected out of noisy data; (b) the feasible solution space is much smaller. By exploiting the advantages, we estimate the BLTM parametrically by employing a deep neural network, leading to better generalization and superior classification performance.

preprint2022arXiv

Gaia: Graph Neural Network with Temporal Shift aware Attention for Gross Merchandise Value Forecast in E-commerce

E-commerce has gone a long way in empowering merchants through the internet. In order to store the goods efficiently and arrange the marketing resource properly, it is important for them to make the accurate gross merchandise value (GMV) prediction. However, it's nontrivial to make accurate prediction with the deficiency of digitized data. In this article, we present a solution to better forecast GMV inside Alipay app. Thanks to graph neural networks (GNN) which has great ability to correlate different entities to enrich information, we propose Gaia, a graph neural network (GNN) model with temporal shift aware attention. Gaia leverages the relevant e-seller' sales information and learn neighbor correlation based on temporal dependencies. By testing on Alipay's real dataset and comparing with other baselines, Gaia has shown the best performance. And Gaia is deployed in the simulated online environment, which also achieves great improvement compared with baselines.

preprint2022arXiv

Linear Bandit Algorithms with Sublinear Time Complexity

We propose two linear bandits algorithms with per-step complexity sublinear in the number of arms $K$. The algorithms are designed for applications where the arm set is extremely large and slowly changing. Our key realization is that choosing an arm reduces to a maximum inner product search (MIPS) problem, which can be solved approximately without breaking regret guarantees. Existing approximate MIPS solvers run in sublinear time. We extend those solvers and present theoretical guarantees for online learning problems, where adaptivity (i.e., a later step depends on the feedback in previous steps) becomes a unique challenge. We then explicitly characterize the tradeoff between the per-step complexity and regret. For sufficiently large $K$, our algorithms have sublinear per-step complexity and $\tilde O(\sqrt{T})$ regret. Empirically, we evaluate our proposed algorithms in a synthetic environment and a real-world online movie recommendation problem. Our proposed algorithms can deliver a more than 72 times speedup compared to the linear time baselines while retaining similar regret.

preprint2022arXiv

MirrorAlign: A Super Lightweight Unsupervised Word Alignment Model via Cross-Lingual Contrastive Learning

Word alignment is essential for the downstream cross-lingual language understanding and generation tasks. Recently, the performance of the neural word alignment models has exceeded that of statistical models. However, they heavily rely on sophisticated translation models. In this study, we propose a super lightweight unsupervised word alignment model named MirrorAlign, in which bidirectional symmetric attention trained with a contrastive learning objective is introduced, and an agreement loss is employed to bind the attention maps, such that the alignments follow mirror-like symmetry hypothesis. Experimental results on several public benchmarks demonstrate that our model achieves competitive, if not better, performance compared to the state of the art in word alignment while significantly reducing the training and decoding time on average. Further ablation analysis and case studies show the superiority of our proposed MirrorAlign. Notably, we recognize our model as a pioneer attempt to unify bilingual word embedding and word alignments. Encouragingly, our approach achieves {16.4X speedup} against GIZA++, and {50X parameter compression} compared with the Transformer-based alignment methods. We release our code to facilitate the community: https://github.com/moore3930/MirrorAlign.

preprint2022arXiv

Objects in Semantic Topology

A more realistic object detection paradigm, Open-World Object Detection, has arisen increasing research interests in the community recently. A qualified open-world object detector can not only identify objects of known categories, but also discover unknown objects, and incrementally learn to categorize them when their annotations progressively arrive. Previous works rely on independent modules to recognize unknown categories and perform incremental learning, respectively. In this paper, we provide a unified perspective: Semantic Topology. During the life-long learning of an open-world object detector, all object instances from the same category are assigned to their corresponding pre-defined node in the semantic topology, including the `unknown' category. This constraint builds up discriminative feature representations and consistent relationships among objects, thus enabling the detector to distinguish unknown objects out of the known categories, as well as making learned features of known objects undistorted when learning new categories incrementally. Extensive experiments demonstrate that semantic topology, either randomly-generated or derived from a well-trained language model, could outperform the current state-of-the-art open-world object detectors by a large margin, e.g., the absolute open-set error is reduced from 7832 to 2546, exhibiting the inherent superiority of semantic topology on open-world object detection.

preprint2022arXiv

Prospects for detecting axion-like particles via the decay $Z\rightarrow af\bar{f}$ at future $Z$ factories

We investigate the prospects for detecting axion-like particles (ALPs, dubbed as "a") via the decay $Z\rightarrow a f\bar{f}$ at future $Z$ factories. Considering the decay channels $a\rightarrow μ^+ μ^-$ and $a\rightarrow b \bar{b}$ , four types of signals $μ^+ μ^- /E$, $b b /E$, $e^+ e^- μ^+ μ^-$ and $e^+ e^- b b$ are explored. We demonstrate that these channels are promising for detecting ALPs at $Z$ factories and obtain the sensitivity bounds on the couplings $g_{aZZ}$ and $g_{aγZ}$.

preprint2022arXiv

Reliable Label Correction is a Good Booster When Learning with Extremely Noisy Labels

Learning with noisy labels has aroused much research interest since data annotations, especially for large-scale datasets, may be inevitably imperfect. Recent approaches resort to a semi-supervised learning problem by dividing training samples into clean and noisy sets. This paradigm, however, is prone to significant degeneration under heavy label noise, as the number of clean samples is too small for conventional methods to behave well. In this paper, we introduce a novel framework, termed as LC-Booster, to explicitly tackle learning under extreme noise. The core idea of LC-Booster is to incorporate label correction into the sample selection, so that more purified samples, through the reliable label correction, can be utilized for training, thereby alleviating the confirmation bias. Experiments show that LC-Booster advances state-of-the-art results on several noisy-label benchmarks, including CIFAR-10, CIFAR-100, Clothing1M and WebVision. Remarkably, under the extreme 90\% noise ratio, LC-Booster achieves 92.9\% and 48.4\% accuracy on CIFAR-10 and CIFAR-100, surpassing state-of-the-art methods by a large margin.

preprint2022arXiv

Sample Efficiency of Data Augmentation Consistency Regularization

Data augmentation is popular in the training of large neural networks; currently, however, there is no clear theoretical comparison between different algorithmic choices on how to use augmented data. In this paper, we take a step in this direction - we first present a simple and novel analysis for linear regression with label invariant augmentations, demonstrating that data augmentation consistency (DAC) is intrinsically more efficient than empirical risk minimization on augmented data (DA-ERM). The analysis is then extended to misspecified augmentations (i.e., augmentations that change the labels), which again demonstrates the merit of DAC over DA-ERM. Further, we extend our analysis to non-linear models (e.g., neural networks) and present generalization bounds. Finally, we perform experiments that make a clean and apples-to-apples comparison (i.e., with no extra modeling or data tweaks) between DAC and DA-ERM using CIFAR-100 and WideResNet; these together demonstrate the superior efficacy of DAC.

preprint2021arXiv

CelebA-Spoof Challenge 2020 on Face Anti-Spoofing: Methods and Results

As facial interaction systems are prevalently deployed, security and reliability of these systems become a critical issue, with substantial research efforts devoted. Among them, face anti-spoofing emerges as an important area, whose objective is to identify whether a presented face is live or spoof. Recently, a large-scale face anti-spoofing dataset, CelebA-Spoof which comprised of 625,537 pictures of 10,177 subjects has been released. It is the largest face anti-spoofing dataset in terms of the numbers of the data and the subjects. This paper reports methods and results in the CelebA-Spoof Challenge 2020 on Face AntiSpoofing which employs the CelebA-Spoof dataset. The model evaluation is conducted online on the hidden test set. A total of 134 participants registered for the competition, and 19 teams made valid submissions. We will analyze the top ranked solutions and present some discussion on future work directions.

preprint2021arXiv

Combinatorial Bandits without Total Order for Arms

We consider the combinatorial bandits problem, where at each time step, the online learner selects a size-$k$ subset $s$ from the arms set $\mathcal{A}$, where $\left|\mathcal{A}\right| = n$, and observes a stochastic reward of each arm in the selected set $s$. The goal of the online learner is to minimize the regret, induced by not selecting $s^*$ which maximizes the expected total reward. Specifically, we focus on a challenging setting where 1) the reward distribution of an arm depends on the set $s$ it is part of, and crucially 2) there is \textit{no total order} for the arms in $\mathcal{A}$. In this paper, we formally present a reward model that captures set-dependent reward distribution and assumes no total order for arms. Correspondingly, we propose an Upper Confidence Bound (UCB) algorithm that maintains UCB for each individual arm and selects the arms with top-$k$ UCB. We develop a novel regret analysis and show an $O\left(\frac{k^2 n \log T}ε\right)$ gap-dependent regret bound as well as an $O\left(k^2\sqrt{n T \log T}\right)$ gap-independent regret bound. We also provide a lower bound for the proposed reward model, which shows our proposed algorithm is near-optimal for any constant $k$. Empirical results on various reward models demonstrate the broad applicability of our algorithm.

preprint2021arXiv

DeeperForensics Challenge 2020 on Real-World Face Forgery Detection: Methods and Results

This paper reports methods and results in the DeeperForensics Challenge 2020 on real-world face forgery detection. The challenge employs the DeeperForensics-1.0 dataset, one of the most extensive publicly available real-world face forgery detection datasets, with 60,000 videos constituted by a total of 17.6 million frames. The model evaluation is conducted online on a high-quality hidden test set with multiple sources and diverse distortions. A total of 115 participants registered for the competition, and 25 teams made valid submissions. We will summarize the winning solutions and present some discussions on potential research directions.

preprint2021arXiv

Single-View 3D Object Reconstruction from Shape Priors in Memory

Existing methods for single-view 3D object reconstruction directly learn to transform image features into 3D representations. However, these methods are vulnerable to images containing noisy backgrounds and heavy occlusions because the extracted image features do not contain enough information to reconstruct high-quality 3D shapes. Humans routinely use incomplete or noisy visual cues from an image to retrieve similar 3D shapes from their memory and reconstruct the 3D shape of an object. Inspired by this, we propose a novel method, named Mem3D, that explicitly constructs shape priors to supplement the missing information in the image. Specifically, the shape priors are in the forms of "image-voxel" pairs in the memory network, which is stored by a well-designed writing strategy during training. We also propose a voxel triplet loss function that helps to retrieve the precise 3D shapes that are highly related to the input image from shape priors. The LSTM-based shape encoder is introduced to extract information from the retrieved 3D shapes, which are useful in recovering the 3D shape of an object that is heavily occluded or in complex environments. Experimental results demonstrate that Mem3D significantly improves reconstruction quality and performs favorably against state-of-the-art methods on the ShapeNet and Pix3D datasets.

preprint2021arXiv

The ANTARES Astronomical Time-Domain Event Broker

We describe the Arizona-NOIRLab Temporal Analysis and Response to Events System (ANTARES), a software instrument designed to process large-scale streams of astronomical time-domain alerts. With the advent of large-format CCDs on wide-field imaging telescopes, time-domain surveys now routinely discover tens of thousands of new events each night, more than can be evaluated by astronomers alone. The ANTARES event broker will process alerts, annotating them with catalog associations and filtering them to distinguish customizable subsets of events. We describe the data model of the system, the overall architecture, annotation, implementation of filters, system outputs, provenance tracking, system performance, and the user interface.

preprint2020arXiv

Adaptive Semantic-Visual Tree for Hierarchical Embeddings

Merchandise categories inherently form a semantic hierarchy with different levels of concept abstraction, especially for fine-grained categories. This hierarchy encodes rich correlations among various categories across different levels, which can effectively regularize the semantic space and thus make predictions less ambiguous. However, previous studies of fine-grained image retrieval primarily focus on semantic similarities or visual similarities. In a real application, merely using visual similarity may not satisfy the need of consumers to search merchandise with real-life images, e.g., given a red coat as a query image, we might get a red suit in recall results only based on visual similarity since they are visually similar. But the users actually want a coat rather than suit even the coat is with different color or texture attributes. We introduce this new problem based on photoshopping in real practice. That's why semantic information are integrated to regularize the margins to make "semantic" prior to "visual". To solve this new problem, we propose a hierarchical adaptive semantic-visual tree (ASVT) to depict the architecture of merchandise categories, which evaluates semantic similarities between different semantic levels and visual similarities within the same semantic class simultaneously. The semantic information satisfies the demand of consumers for similar merchandise with the query while the visual information optimizes the correlations within the semantic class. At each level, we set different margins based on the semantic hierarchy and incorporate them as prior information to learn a fine-grained feature embedding. To evaluate our framework, we propose a new dataset named JDProduct, with hierarchical labels collected from actual image queries and official merchandise images on an online shopping application. Extensive experimental results on the public CARS196 and CUB-

preprint2020arXiv

An Automated Framework for Board-level Trojan Benchmarking

Economic and operational advantages have led the supply chain of printed circuit boards (PCBs) to incorporate various untrusted entities. Any of the untrusted entities are capable of introducing malicious alterations to facilitate a functional failure or leakage of secret information during field operation. While researchers have been investigating the threat of malicious modification within the scale of individual microelectronic components, the possibility of a board-level malicious manipulation has essentially been unexplored. In the absence of standard benchmarking solutions, prospective countermeasures for PCB trust assurance are likely to utilize homegrown representation of the attacks that undermines their evaluation and does not provide scope for comparison with other techniques. In this paper, we have developed the first-ever benchmarking solution to facilitate an unbiased and comparable evaluation of countermeasures applicable to PCB trust assurance. Based on a taxonomy tailored for PCB-level alterations, we have developed high-level Trojan models. From these models, we have generated a custom pool of board-level Trojan designs of varied complexity and functionality. We have also developed a tool-flow for automatically inserting these Trojans into various PCB designs and generate the Trojan benchmarks (i.e., PCB designs with Trojan). The tool-based Trojan insertion facilitate a comprehensive evaluation against large number of diverse Trojan implementations and application of data mining for trust verification. Finally, with experimental measurements from a fabricated PCB, we analyze the stealthiness of the Trojan designs.

preprint2020arXiv

Construction and classification of point group symmetry protected topological phases in 2D interacting fermionic systems

The construction and classification of symmetry-protected topological (SPT) phases in interacting bosonic and fermionic systems have been intensively studied in the past few years. Very recently, a complete classification and construction of space group SPT phases were also proposed for interacting bosonic systems. In this paper, we attempt to generalize this classification and construction scheme systematically into interacting fermion systems. In particular, we construct and classify point group SPT phases for 2D interacting fermion systems via lower-dimensional block-state decorations. We discover several intriguing fermionic SPT states that can only be realized in interacting fermion systems (i.e., not in free-fermion or bosonic SPT systems). Moreover, we also verify the recently conjectured crystalline equivalence principle for 2D interacting fermion systems. Finally, the potential experimental realization of these new classes of point group SPT phases in 2D correlated superconductors is addressed.

preprint2020arXiv

High-Order Information Matters: Learning Relation and Topology for Occluded Person Re-Identification

Occluded person re-identification (ReID) aims to match occluded person images to holistic ones across dis-joint cameras. In this paper, we propose a novel framework by learning high-order relation and topology information for discriminative features and robust alignment. At first, we use a CNN backbone and a key-points estimation model to extract semantic local features. Even so, occluded images still suffer from occlusion and outliers. Then, we view the local features of an image as nodes of a graph and propose an adaptive direction graph convolutional (ADGC)layer to pass relation information between nodes. The proposed ADGC layer can automatically suppress the message-passing of meaningless features by dynamically learning di-rection and degree of linkage. When aligning two groups of local features from two images, we view it as a graph matching problem and propose a cross-graph embedded-alignment (CGEA) layer to jointly learn and embed topology information to local features, and straightly predict similarity score. The proposed CGEA layer not only take full use of alignment learned by graph matching but also re-place sensitive one-to-one matching with a robust soft one. Finally, extensive experiments on occluded, partial, and holistic ReID tasks show the effectiveness of our proposed method. Specifically, our framework significantly outperforms state-of-the-art by6.5%mAP scores on Occluded-Duke dataset.

preprint2020arXiv

Lattice model constructions for gapless domain walls between topological phases

Domain walls between different topological phases are one of the most interesting phenomena that reveal the non-trivial bulk properties of topological phases. Very recently, gapped domain walls between different topological phases have been intensively studied. In this paper, we systematically construct a large class of lattice models for gapless domain walls between twisted and untwisted gauge theories with arbitrary finite group $G$. As simple examples, we numerically study several finite groups(including both Abelian and non-Abelian finite group such as $S_3$) in $2$D using the state-of-the-art loop optimization of tensor network renormalization algorithm. We also propose a physical mechanism for understanding the gapless nature of these particular domain wall models. Finally, by taking advantage of the classification and construction of twisted gauge theories using group cohomology theory, we generalize such constructions into arbitrary dimensions, which might provide us a systematical way to understand gapless domain walls and topological quantum phase transitions.

preprint2020arXiv

Predicting quantum many-body dynamics with transferable neural networks

Machine learning (ML) architectures such as convolutional neural networks (CNNs) have garnered considerable recent attention in the study of quantum many-body systems. However, advanced ML approaches such as transfer learning have seldom been applied to such contexts. Here we demonstrate that a simple recurrent unit (SRU) based efficient and transferable sequence learning framework is capable of learning and accurately predicting the time evolution of one-dimensional (1D) Ising model with simultaneous transverse and parallel magnetic fields, as quantitatively corroborated by relative entropy measurements and magnetization between the predicted and exact state distributions. At a cost of constant computational complexity, a larger many-body state evolution was predicted in an autoregressive way from just one initial state, without any guidance or knowledge of any Hamiltonian. Our work paves the way for future applications of advanced ML methods in quantum many-body dynamics only with knowledge from a smaller system.

preprint2019arXiv

Alpha Decay to Doubly Magic Core in Quartetting Wave Function Approach

We present a microscopic calculation of $α$-cluster formation in heavy nuclei $^{104}$Te ($α$+$^{100}$Sn), $^{212}$Po ($α$+$^{208}$Pb) and their neighbors $^{102}$Sn, $^{102}$Te, $^{210}$Pb and $^{210}$Po by using the quartetting wave function approach. Improving the local density approximation, the shell structure of the core nucleus is considered, and the center-of-mass (c.o.m.) effective potential for the quartet is obtained self-consistently from the shell model wavefunctions. The $α$-cluster formation and decay probabilities are obtained by solving the bound-state of the c.o.m. motion of the quartet and the scattering state of the formed $α$-cluster in the Gurvitz approach. Striking shell effects on the $α$-cluster formation probabilities are analyzed for magic numbers 50, 82 and 126. The computed $α$-decay half-lives of these special nuclei are compared with the newest experimental data.

preprint2019arXiv

Computing the quasipotential for highly dissipative and chaotic SDEs. An application to stochastic Lorenz'63

The study of noise-driven transitions occurring rarely on the time-scale of systems modeled by SDEs is of crucial importance for understanding such phenomena as genetic switches in living organisms and magnetization switches of the Earth. For a gradient SDE, the predictions for transition times and paths between its metastable states are done using the potential function. For a nongradient SDE, one needs to decompose its forcing into a gradient of the so-called quasipotential and a rotational component, which cannot be done analytically in general. We propose a methodology for computing the quasipotential for highly dissipative and chaotic systems built on the example of Lorenz'63 with an added stochastic term. It is based on the ordered line integral method, a Dijkstra-like quasipotential solver, and combines 3D computations in whole regions, a dimensional reduction technique, and 2D computations on radial meshes on manifolds or their unions. Our collection of source codes is available on M. Cameron's web page and on GitHub.

preprint2019arXiv

Dynamical symmetry in quantum dissipative models

We show that the dynamical symmetry exists in dissipative quantum many-body systems. Under constraints on both Hamiltonian and dissipation parts, the time evolution of particular observables can be symmetric between repulsive and attractive interactions in the Hubbard model, or symmetric between ferromagnetic and anti-ferromagnetic interactions in the Ising model with external fields. We present a theorem to determine the existence of the dynamical symmetry in dissipative systems. This theorem is also responsible for the symmetry of steady states, even without the constraint on the initial state. We demonstrate the applications of our theorem with numerical simulations using tensor network algorithms.

preprint2019arXiv

Loop update for infinite projected entangled-pair states in two spatial dimensions

We propose an improved approach to carry out the imaginary time evolution of infinite projected entangled-pair states (iPEPS), especially for systems with criticality. A cyclic optimal truncation is introduced to update the tensors along a closed loop, aiming to remove the redundant internal correlations. We demonstrate the algorithm by considering an elaborate evolution based on simple update on a small plaquette. This scheme can also be applied to a full update strategy. We demonstrate their performances on simulating the ground states of the spin-$1/2$ anti-ferromagnetic Heisenberg model and the transverse field Ising model on a square lattice.

preprint2016arXiv

ANTARES: Progress towards building a `Broker' of time-domain alerts

The Arizona-NOAO Temporal Analysis and Response to Events System (ANTARES) is a joint effort of NOAO and the Department of Computer Science at the University of Arizona to build prototype software to process alerts from time-domain surveys, especially LSST, to identify those alerts that must be followed up immediately. Value is added by annotating incoming alerts with existing information from previous surveys and compilations across the electromagnetic spectrum and from the history of past alerts. Comparison against a knowledge repository of properties and features of known or predicted kinds of variable phenomena is used for categorization. The architecture and algorithms being employed are described.

preprint2016arXiv

Application of Statistical Relational Learning to Hybrid Recommendation Systems

Recommendation systems usually involve exploiting the relations among known features and content that describe items (content-based filtering) or the overlap of similar users who interacted with or rated the target item (collaborative filtering). To combine these two filtering approaches, current model-based hybrid recommendation systems typically require extensive feature engineering to construct a user profile. Statistical Relational Learning (SRL) provides a straightforward way to combine the two approaches. However, due to the large scale of the data used in real world recommendation systems, little research exists on applying SRL models to hybrid recommendation systems, and essentially none of that research has been applied on real big-data-scale systems. In this paper, we proposed a way to adapt the state-of-the-art in SRL learning approaches to construct a real hybrid recommendation system. Furthermore, in order to satisfy a common requirement in recommendation systems (i.e. that false positives are more undesirable and therefore penalized more harshly than false negatives), our approach can also allow tuning the trade-off between the precision and recall of the system in a principled way. Our experimental results demonstrate the efficiency of our proposed approach as well as its improved performance on recommendation precision.

preprint2016arXiv

Search for anomalous couplings via single top quark production in association with a photon at LHC

Considering the experimental constraints given by CMS collaboration at $\sqrt{s} = 8$ TeV on the strength of top quark flavour-changing neutral-current (FCNC) couplings $tqγ$ and $tqg$, we investigate the production of top quark in association with a photon and carry out a full simulation for the signals $\ellνbγ$ and $jjbγ$ at 14 TeV LHC. In our numerical analysis, the contributions of single top production with a photon radiation off the top decay products are also included. The discovery potential for anomalous couplings $tqγ$ and $tqg$ with an integrated luminosity of 100 fb$^{-1}$ are examined in detail.

preprint2015arXiv

Criticality in Translation-Invariant Parafermion Chains

In this work we numerically study critical phases in translation-invariant $\mathbb{Z}_N$ parafermion chains with both nearest- and next-nearest-neighbor hopping terms. The model can be mapped to a $\mathbb{Z}_N$ spin model with nearest-neighbor couplings via a generalized Jordan-Wigner transformation and translation invariance ensures that the spin model is always self-dual. We first study the low-energy spectrum of chains with only nearest-neighbor coupling, which are mapped onto standard self-dual $\mathbb{Z}_N$ clock models. For $3\leq N\leq 6$ we match the numerical results to the known conformal field theory(CFT) identification. We then analyze in detail the phase diagram of a $N=3$ chain with both nearest and next-nearest neighbor hopping and six critical phases with central charges being $4/5$, 1 or 2 are found. We find continuous phase transitions between $c=1$ and $c=2$ phases, while the phase transition between $c=4/5$ and $c=1$ is conjectured to be of Kosterlitz-Thouless type.

preprint2015arXiv

DeepID-Net: Deformable Deep Convolutional Neural Networks for Object Detection

In this paper, we propose deformable deep convolutional neural networks for generic object detection. This new deep learning object detection framework has innovations in multiple aspects. In the proposed new deep architecture, a new deformation constrained pooling (def-pooling) layer models the deformation of object parts with geometric constraint and penalty. A new pre-training strategy is proposed to learn feature representations more suitable for the object detection task and with good generalization capability. By changing the net structures, training strategies, adding and removing some key components in the detection pipeline, a set of models with large diversity are obtained, which significantly improves the effectiveness of model averaging. The proposed approach improves the mean averaged precision obtained by RCNN \cite{girshick2014rich}, which was the state-of-the-art, from 31\% to 50.3\% on the ILSVRC2014 detection test set. It also outperforms the winner of ILSVRC2014, GoogLeNet, by 6.1\%. Detailed component-wise analysis is also provided through extensive experimental evaluation, which provide a global view for people to understand the deep learning object detection pipeline.

preprint2015arXiv

From Facial Parts Responses to Face Detection: A Deep Learning Approach

In this paper, we propose a novel deep convolutional network (DCN) that achieves outstanding performance on FDDB, PASCAL Face, and AFW. Specifically, our method achieves a high recall rate of 90.99% on the challenging FDDB benchmark, outperforming the state-of-the-art method by a large margin of 2.91%. Importantly, we consider finding faces from a new perspective through scoring facial parts responses by their spatial structure and arrangement. The scoring mechanism is carefully formulated considering challenging cases where faces are only partially visible. This consideration allows our network to detect faces under severe occlusion and unconstrained pose variation, which are the main difficulty and bottleneck of most existing face detection approaches. We show that despite the use of DCN, our network can achieve practical runtime speed.

preprint2015arXiv

Observation of Anderson localization in ultrathin films of three-dimensional topological insulators

Anderson localization, the absence of diffusive transport in disordered systems, has been manifested as hopping transport in numerous electronic systems, whereas in recently discovered topological insulators it has not been directly observed. Here we report experimental demonstration of transition from diffusive transport in the weak antilocalization regime to variable range hopping transport in the Anderson localization regime with ultrathin (Bi$_{1-x}$Sb$_x$)$_2$Te$_3$ films. As disorder becomes stronger, negative magnetoconductivity due to the weak antilocalization is gradually suppressed, and eventually positive magnetoconductivity emerges when the electron system becomes strongly localized. This works reveals the critical role of disorder in the quantum transport properties of ultrathin topological insulator films, in which theories have predicted rich physics related to topological phase transitions.

preprint2015arXiv

Proximity effect between a topological insulator and a magnetic insulator with large perpendicular anisotropy

We report that thin films of a prototype topological insulator, Bi$_{2}$Se$_{3}$, can be epitaxially grown onto the (0001) surface of BaFe$_{12}$O$_{19}$(BaM), a magnetic insulator with high Curie temperature and large perpendicular anisotropy. In the Bi$_2$Se$_3$ thin films grown on non-magnetic substrates, classic weak antilocalization (WAL) is manifested as cusp-shaped positive magnetoresistance (MR) in perpendicular magnetic fields and parabola-shaped positive MR in parallel fields, whereas in Bi$_{2}$Se$_{3}$/BaM heterostructures the low field MR is parabola-shaped, which is positive in perpendicular fields and negative in parallel fields. The magnetic field and temperature dependence of the MR is explained as a consequence of the suppression of WAL due to strong magnetic interactions at the Bi$_{2}$Se$_{3}$/BaM interface.

preprint2015arXiv

The signatures of the quintuplet leptons at the LHC

We investigate production and detection prospects of the quintuplet heavy leptons at the LHC in the context of a new model which is proposed as a viable and testable solution to the neutrino masses problem. We classify the signals, carry out a full simulation on the signals and the relevant backgrounds at the 14 TeV LHC. After applying suitable kinematic cuts, the background events are substantially suppressed. The signals of the heavy leptons might be detected at the 14 TeV LHC.

preprint2015arXiv

WIDER FACE: A Face Detection Benchmark

Face detection is one of the most studied topics in the computer vision community. Much of the progresses have been made by the availability of face detection benchmark datasets. We show that there is a gap between current face detection performance and the real world requirements. To facilitate future face detection research, we introduce the WIDER FACE dataset, which is 10 times larger than existing datasets. The dataset contains rich annotations, including occlusions, poses, event categories, and face bounding boxes. Faces in the proposed dataset are extremely challenging due to large variations in scale, pose and occlusion, as shown in Fig. 1. Furthermore, we show that WIDER FACE dataset is an effective training source for face detection. We benchmark several representative detection systems, providing an overview of state-of-the-art performance and propose a solution to deal with large scale variation. Finally, we discuss common failure cases that worth to be further investigated. Dataset can be downloaded at: mmlab.ie.cuhk.edu.hk/projects/WIDERFace

preprint2014arXiv

Associated production of the heavy charged gauge boson ${W_{H}}$ and a top quark at LHC

In the context of topflavor seesaw model, we study the production of the heavy charged gauge boson ${W_{H}}$ associated with a top quark at the LHC. Focusing on the searching channel $pp\rightarrow tW_H\rightarrow t\bar{t}b \rightarrow lνjjbbb$, we carry out a full simulation of the signal and the relevant standard model backgrounds. The kinematical distributions of final states are presented. It is found that the backgrounds can be significantly suppressed by sets of kinematic cuts, and the signal of the heavy charged boson might be detected at the LHC with $\sqrt{s}=14$ TeV. With a integrated luminosity of $\LL= $ 100 $fb^{-1}$, a $8.3 σ$ signal significance can be achieved for $m_{W_H}=1.6$ TeV.

preprint2014arXiv

Chiral projected entangled-pair state with topological order

We show that projected entangled-pair states (PEPS) can describe chiral topologically ordered phases. For that, we construct a simple PEPS for spin-1/2 particles in a two-dimensional lattice. We reveal a symmetry in the local projector of the PEPS that gives rise to the global topological character. We also extract characteristic quantities of the edge conformal field theory using the bulk-boundary correspondence.

preprint2014arXiv

DeepID-Net: multi-stage and deformable deep convolutional neural networks for object detection

In this paper, we propose multi-stage and deformable deep convolutional neural networks for object detection. This new deep learning object detection diagram has innovations in multiple aspects. In the proposed new deep architecture, a new deformation constrained pooling (def-pooling) layer models the deformation of object parts with geometric constraint and penalty. With the proposed multi-stage training strategy, multiple classifiers are jointly optimized to process samples at different difficulty levels. A new pre-training strategy is proposed to learn feature representations more suitable for the object detection task and with good generalization capability. By changing the net structures, training strategies, adding and removing some key components in the detection pipeline, a set of models with large diversity are obtained, which significantly improves the effectiveness of modeling averaging. The proposed approach ranked \#2 in ILSVRC 2014. It improves the mean averaged precision obtained by RCNN, which is the state-of-the-art of object detection, from $31\%$ to $45\%$. Detailed component-wise analysis is also provided through extensive experimental evaluation.

preprint2014arXiv

Determination of charged Higgs couplings at the LHC

We review the study of the charged Higgs and top quark associated production at the LHC with the presence of an additional scalar doublet. Top quark spin effects are related to the Higgs fermion couplings through this process. The angular distributions with respect to top quark spin turn out to be distinctive observables to study the $Htb$ interaction in different models.

preprint2014arXiv

Hadronic $b^\prime$ search at the LHC with top and W taggers

We study the sensitivity of a down type quark $b^{\prime}$ via process $pp\rightarrow b^{\prime}\bar{b^{\prime}} \rightarrow tW^-\bar{t}W^+$ using jet substructure methods at the LHC with the collision energy $\sqrt{s}=14$ TeV. We consider the case that the $b^\prime$ is heavy (say from 800 GeV to 1500 GeV) and concentrate on the feasibility of the full hadronic mode. Both top tagger (the HEP top tagger) and W tagger (the CMS W-tagging) are used to reconstruct all objects in the final states. In order to suppress huge SM background events and take into account various cases with different number of boosted objects, we propose a comprehensive reconstruction procedure so as to extract the most crucial observables of the signal events. When $b^\prime$ mass is 1 TeV, it is found that with a 200 fb$^{-1}$ dataset, the LHC may be able to detect the $b^\prime$ with a significance up to $10$ or better. With a 3000 $fb^{-1}$ dataset, the LHC may be able to probe the $b^\prime$ with a mass around up to 2 TeV, only by using the hadronic mode.

preprint2014arXiv

Topology and Criticality in Resonating Affleck-Kennedy-Lieb-Tasaki loop Spin Liquid States

We exploit a natural Projected Entangled-Pair State (PEPS) representation for the resonating Affleck-Kennedy-Lieb-Tasaki loop (RAL) state. By taking advantage of PEPS-based analytical and numerical methods, we characterize the RAL states on various two-dimensional lattices. On square and honeycomb lattices, these states are critical since the dimer-dimer correlations decay as a power law. On kagome lattice, the RAL state has exponentially decaying correlation functions, supporting the scenario of a gapped spin liquid. We provide further evidence that the RAL state on the kagome lattice is a $\mathbb{Z}_2$ spin liquid, by identifying the four topological sectors and computing the topological entropy. Furthermore, we construct a one-parameter family of PEPS states interpolating between the RAL state and a short-range Resonating Valence Bond state and find a critical point, consistent with the fact that the two states belong to two different phases. We also perform a variational study of the spin-1 kagome Heisenberg model using this one-parameter PEPS.

preprint2013arXiv

Charged Higgs Production in Association With W^{\pm} at Large Hadron Colliders

Many new physics models beyond the standard model, such as the littlest higgs models and the left right twin higgs models, predict the existence of the large charged higgs couplings H^-q\bar b and H^+b\bar q, where q=t or the new vector-like heavy quark T; On the other hand, some new physics models like the littlest higgs also predict the gauge-higgs couplings. Such couplings may have rich collider phenomenology. We focus our attention on these couplings induced by the littlest higgs models and the left right twin higgs models models and consider their contributions to the production cross section for W^\pm H^\mp production at the large hadron colliders. We find that the cross sections, in the littlest higgs models, on the parton level gg \to W^\pm H^\mp and q\bar q \to W^\pm H^\mp (q=u,d,s,c,b) may reach tens of several dozen femtobarns in reasonable parameters space at the collision energy of 14 TeV and that the total cross section can even reach a few hundred femtobarns in certain favored space. While in the left right twin higgs models, the production rates are basically one order lower than these in littlest higgs. Therefore, due to the large cross sections of that in the littlest higgs, it may be possible to probe the charged higgs via this process in a large parameter space.

preprint2013arXiv

Production and decay of for the 125 GeV Higgs boson in the littlest Higgs model with T-parity

Motivated by recent search results for the standard model (SM) Higgs boson at the Large Hadron Collider (LHC), we revisit the Higgs phenomenology in the littlest Higgs model with T-parity (LHT). We present the signal strength modifier $μ$ respectively for the main search channels $qq' \rightarrow jjh\rightarrow jjγγ$, $qq'\rightarrow Vh\rightarrow Vγγ$, $qq'\rightarrow Vh \rightarrow Vbb$, $gg\rightarrow h \rightarrow γγ$, and $gg \rightarrow h \rightarrow VV$ in the LHT model. It is found that an enhancement factor of $1.09-1.56$ in $qq' \rightarrow jjh\rightarrow jjγγ$ channel can be obtained for this model in Case B with parameter $f$ in the range 1000 GeV$\sim 500$ GeV. However, the rates for $b\bar{b}$, $τ\barτ$ are significantly suppressed relative to the SM predictions which are still consistent with the current sensitivity. It is hoped that will be further tested with larger integrated luminosity at the LHC.

preprint2013arXiv

Top quark spin and $Htb$ interaction in charged Higgs and top quark associated production at LHC

We study the charged Higgs production at LHC via its associated production with top quark. The kinematic cuts are optimized to suppress the background processes so that the reconstruction of the charged Higgs and top quark is possible. The angular distributions with respect to top quark spin are explored to study the $Htb$ interaction at LHC.

preprint2012arXiv

LHCb $\bigtriangleup A_{CP}$ of $D$ meson and R-Parity Violation

LHCb collaboration has recently announced a measurement of the difference of time-integrated CP asymmetries between $D\rightarrow K^+K^-$ and $D \rightarrow π^+π^-$. This result provides the evidence of large direct CP violation in $D$ meson and reveals some important implications on underlying new physics. It is shown that the direct CP violation in $D$ meson can be enhanced by R-parity violating supersymmetry, while CP violations in $K$ and $B$ mesons are suppressed by this new physics, which is in consistence with previous experiments. Constraints on the model parameters and some consequences are also discussed.

preprint2012arXiv

Quantum phases of disordered flatband lattice fractional quantum Hall systems

By numerical exact diagonalization techniques, we obtain the quantum phase diagram of the lattice fractional quantum Hall (FQH) systems in the presence of quenched disorder. By implementing an array of local potential traps representing the disorder, we show that the system undergoes a series of quantum phase transitions as the disorder and/or the interaction is tuned. As the strength of potential traps is increased, the FQH state turns into a compressible liquid, and then into a topologically trivial insulator. We use numerically calculated energy gap, quantum degeneracy, Chern number, entanglement spectrum, and fidelity to identify various quantum phases. The connection to continuum FQH effects is also discussed.

preprint2012arXiv

Scalar Septuplet Dark Matter and Enhanced $h\rightarrow γγ$ Decay Rate

Inspired by recent results on the Higgs search from ATLAS and CMS, we extend the SM with complex septuplet scalars. The lightest neutral component of the septuplets is a natural cold dark Matter candidate and the charged components can contribute to the $h\to γγ$ decay rate, providing a significant enhancement factor. The dark matter phenomenology and possible collider signatures of the model are investigated. We find a dark matter candidate with mass around 70 GeV consistent with astrophysical and direct detection constraints. In the meanwhile, the enhancement factor of $h\to γγ$ decay rate can be in the range 1.5\sim 2.

preprint2012arXiv

Searching for Heavy Charged Higgs Boson with Jet Substructure at the LHC

We study the heavy charged Higgs boson (from 800 GeV to 1500 GeV in this study) in production associated with a top quark at the LHC with the collision energy $\sqrt{s}=14$ TeV. Such a heavy charged Higgs boson can dominantly decay into a top quark and a bottom quark due to its large Yukawa couplings, like in MSSM. To suppress background events and to confirm the signal, we reconstruct the mass bumps of the heavy charged Higgs boson and the associated top quark. For this purpose, we propose a hybrid-R reconstruction method which utilizes the top tagging technique, a jet substructure technique developed for highly boosted massive particles. By using the full hadronic mode of $p p \to H^{\pm} t \to t tb$ as a test field, we find that this method can greatly reduce the combinatorics in the full reconstruction and can successfully reduce background events down to a controlled level. The sensitivity of LHC to the heavy charged Higgs boson with two $b$ taggings is studied and a $9.5σ$ significance can be achieved when $m_{H^\pm} =1 \textrm{TeV}$.

preprint2012arXiv

Supersymmetric Extension of the Minimal Dark Matter Model

The minimal dark matter model is given a supersymmetric extension. A super SU(2)L quintuplet is introduced with its fermionic neutral component still being the dark matter, the dark matter particle mass is about 19.7 GeV. Mass splitting among the quintplet due to supersymmetry particles is found to be negligibly small compared to the electroweak corrections. Other properties of this supersymmetry model are studied, it has the solutions to the PAMELA and Fermi-LAT anomaly, the predictions in higher energies need further experimental data to verify.

preprint2012arXiv

Topological flat band models with arbitrary Chern numbers

We report the theoretical discovery of a systematic scheme to produce topological flat bands (TFBs) with arbitrary Chern numbers. We find that generically a multi-orbital high Chern number TFB model can be constructed by considering multi-layer Chern number C=1 TFB models with enhanced translational symmetry. A series of models are presented as examples, including a two-band model on a triangular lattice with a Chern number C=3 and an $N$-band square lattice model with $C=N$ for an arbitrary integer $N$. In all these models, the flatness ratio for the TFBs is larger than 30 and increases with increasing Chern number. In the presence of appropriate inter-particle interactions, these models are likely to lead to the formation of novel Abelian and Non-Abelian fractional Chern insulators. As a simple example, we test the C=2 model with hardcore bosons at 1/3 filling and an intriguing fractional quantum Hall state is observed.

preprint2011arXiv

Generic Hubbard model description of semiconductor quantum dot spin qubits

We introduce a Hubbard model as the simple quantum generalization of the classical capacitance circuit model to study semiconductor quantum-dot spin qubits. We prove theoretically that our model is equivalent to the usual capacitance circuit model in the absence of quantum fluctuations. However, our model naturally includes quantum effects such as hopping and spin exchange. The parameters of the generalized Hubbard model can either be directly read off from the experimental plot of the stability diagram or be calculated from the microscopic theory, establishing a quantitative connection between the two. We show that, while the main topology of the charge stability diagram is determined by the ratio between inter-site and on-site Coulomb repulsion, fine details of the stability diagram reveal information about quantum effects. Extracting quantum information from experiments using our Hubbard model approach is simple, but would require the measurement resolution to increase by an order of magnitude.

preprint2011arXiv

Hubbard model description of silicon spin qubits: charge stability diagram and tunnel coupling in Si double quantum dots

We apply the recently introduced Hubbard model approach to quantitatively describe the experimental charge stability diagram and tunnel coupling of silicon double quantum dot systems. The results calculated from both the generalized Hubbard model and the microscopic theory are compared with existing experimental data, and excellent agreement between theory and experiment is found. The central approximation of our theory is a reduction of the full multi-electron multi-band system to an effective two-electron model, which is numerically tractable. In the microscopic theory we utilize the Hund-Mulliken approximation to the electron wave functions and compare the results calculated with two different forms of confinement potentials (biquadratic and Gaussian). We discuss the implications of our work for future studies.

preprint2011arXiv

Low-noise conditional operation of singlet-triplet coupled quantum dot qubits

We theoretically study the influence of charge noise on a controlled phase gate, implemented using two proximal double quantum dots coupled electrostatically. Using the configuration interaction method, we present a full description of the conditional control scheme and quantitatively calculate the gate error arising from charge fluctuations. Our key finding is that the existence of noise-immune sweet spots depends on not only the energy detuning but also the device geometry. The conditions for sweet spots with minimal charge noise are predicted analytically and verified numerically. Going beyond the simple sweet-spot concept we demonstrate the existence of other optimal situations for fast and low-noise singlet-triplet two-qubit gates.

preprint2011arXiv

Quantum theory of the charge stability diagram of semiconductor double quantum dot systems

We complete our recently introduced theoretical framework treating the double quantum dot system with a generalized form of Hubbard model. The effects of all quantum parameters involved in our model on the charge stability diagram are discussed in detail. A general formulation of the microscopic theory is presented, and truncating at one orbital per site, we study the implication of different choices of the model confinement potential on the Hubbard parameters as well as the charge stability diagram. We calculate the charge stability diagram keeping three orbitals per site and find that the effect of additional higher-lying orbitals on the subspace with lowest-energy orbitals only can be regarded as a small renormalization of Hubbard parameters, thereby justifying our practice of keeping only the lowest-orbital in all other calculations. The role of the harmonic oscillator frequency in the implementation of the Gaussian model potential is discussed, and the effect of an external magnetic field is identified to be similar to choosing a more localized electron wave function in microscopic calculations. The full matrix form of the Hamiltonian including all possible exchange terms, and several peculiar charge stability diagrams due to unphysical parameters are presented in the appendix, thus emphasizing the critical importance of a reliable microscopic model in obtaining the system parameters defining the Hamiltonian.

preprint2010arXiv

Associated production of Z boson and a pair of new quarks at the LHC

The associated production of $Z$ boson and a pair of new quarks at the Large Hadron Collider (LHC) is studied. The cross sections for both sequential fermions and vector-like fermions are presented. It is found that for sequential fermions the cross sections can reach $1\sim 10^2$ fb for heavy quark mass $m_Q$ from 1000 GeV to 200 GeV. For vector-like quarks, the cross sections are suppressed by mixing parameter $\sinθ_L$. Focusing on process $pp \to b'b'$, we investigate the possibility of detecting the $6l+2j$ signal. For a $b'$ with light mass and a large branching ratio of $b'\to bZ$, it is found that only several signal events (parton level) can be produced with 1000 fb$^{-1}$ integrated luminosity. Although the signal events are rare, all the final states are produced centrally and multi lepton final states are clear at hadron collider, which could be easily detected.

preprint2010arXiv

Productions of heavy charged leptons via gluon fusion at LHC: A revisit

Heavy charged lepton productions via gluon fusion at the LHC are revisited. Full loop calculations are adopted with an updated parton distribution function and electroweak data. Including contribution from new generation quarks in the loop, pair production of the sequential heavy lepton via gluon fusion at the LHC dominates over that via the Drell-Yan mechanism in some heavy lepton mass range. Exotic lepton single production of vector-like lepton extended models is also calculated. In the later case, the gluon fusion mechanism via the Higgs exchange is emphasized. Our numerical results for both pair and single production of heavy leptons are smaller than previous studies especially for a large heavy lepton mass as a result of full loop calculation and due to the mixing angles.

preprint2009arXiv

Plans for PANDA Online Computing

The PANDA experiment will not use any hardware trigger, i.e. all raw data are streaming in the data acquisition with a bandwidth of ~280 GB/s. The PANDA Online System is designed to perform data reduction by a factor of ~800 by reconstruction algorithms programmed in VHDL (Very High Speed Integrated Circuit Hardware Description Language) on FPGAs (Field Programmable Gate Arrays).

preprint2008arXiv

Competitions of magnetism and superconductivity in FeAs-based materials

Using the numerical unrestricted Hartree-Fock approach, we study the ground state of a two-orbital model describing newly discovered FeAs-based superconductors. We observe the competition of a $(0, π)$ mode spin-density wave and the superconductivity as the doping concentration changes. There might be a small region in the electron-doping side where the magnetism and superconductivity coexist. The superconducting pairing is found to be spin singlet, orbital even, and mixed s$_{xy}$ + d$_{x^{2}-y^{2}}$ wave (even parity).

preprint2008arXiv

Observable Topological Effects of Mobius Molecular Devices

We study the topological properties of quantum states for the spinless particle hopping in a Mobius ladder. This system can be regarded as a molecular device possibly engineered from the aromatic Mobius annulenes, which enjoys a pseudo-spin orbital interaction described by a non-Abelian gauge structure. It results from the nontrivial topology of configuration space, and results in various observable effects, such as optical spectral splitting. The transmission spectrum through the Mobius molecular device is calculated to demonstrate a topological effect as a destructive interferences in the conduction band. The induced interaction also leads to an entanglement between the transverse and longitudinal modes for any locally factorized state.

preprint2006arXiv

Hadronic production of the heavy gauge bosons ($Z_{H}$, $B_{H}$, $W_{H}^{\pm}$) associated with one QCD jet

This paper has been withdrawn by the author, due some mistakes in this paper.

Institution

Affiliation not imported yet

This author record came from a source that does not expose affiliation metadata. Once the author claims the profile or we enrich the record from another provider, this section will link to the concrete institution.

Source provenance

Where this author record came from

arxivconfidence 95%

external id: arxiv:2605.10315:author:2:shuo-yang

Imported May 20, 2026Synced May 21, 2026

arxivconfidence 95%

external id: arxiv:2605.18903:author:3:shuo-yang

Imported May 20, 2026Synced May 21, 2026

arxivconfidence 95%

external id: arxiv:2604.27536:author:5:shuo-yang