arXiv subjects

Hao Zhang

Publications and source records attributed to Hao Zhang.

At least 919 records · Page 51Linked to original sources

arXiv2022

Investigating Higgs self-interaction through di-Higgs plus jet production

The Higgs self coupling measurement is quite essential for determining the shape of the Higgs potential and nature of the Higgs boson. We propose the di-Higgs plus jet final states at hadron colliders to increase the discovery sensitivity of the Higgs self coupling at the low invariant mass region. Our simulation indicates that the allowed region of the Higgs self coupling would be further narrowed from $[-1.5,6.7]$ from the most recent ATLAS report down to $[0.5, 1.7]$. Furthermore, we find negative Higgs self couplings would be disfavored beyond $2σ$ confidence level at a future 100TeV collider with the help of this signal.

hep-ph↗

arXiv2022

Language Model Decomposition: Quantifying the Dependency and Correlation of Language Models

Pre-trained language models (LMs), such as BERT (Devlin et al., 2018) and its variants, have led to significant improvements on various NLP tasks in past years. However, a theoretical framework for studying their relationships is still missing. In this paper, we fill this gap by investigating the linear dependency between pre-trained LMs. The linear dependency of LMs is defined analogously to the linear dependency of vectors. We propose Language Model Decomposition (LMD) to represent a LM using a linear combination of other LMs as basis, and derive the closed-form solution. A goodness-of-fit metric for LMD similar to the coefficient of determination is defined and used to measure the linear dependency of a set of LMs. In experiments, we find that BERT and eleven (11) BERT-like LMs are 91% linearly dependent. This observation suggests that current state-of-the-art (SOTA) LMs are highly "correlated". To further advance SOTA we need more diverse and novel LMs that are less dependent on existing LMs.

cs.CL↗

arXiv2022

UKP-SQuARE v2: Explainability and Adversarial Attacks for Trustworthy QA

Question Answering (QA) systems are increasingly deployed in applications where they support real-world decisions. However, state-of-the-art models rely on deep neural networks, which are difficult to interpret by humans. Inherently interpretable models or post hoc explainability methods can help users to comprehend how a model arrives at its prediction and, if successful, increase their trust in the system. Furthermore, researchers can leverage these insights to develop new methods that are more accurate and less biased. In this paper, we introduce SQuARE v2, the new version of SQuARE, to provide an explainability infrastructure for comparing models based on methods such as saliency maps and graph-based explanations. While saliency maps are useful to inspect the importance of each input token for the model's prediction, graph-based explanations from external Knowledge Graphs enable the users to verify the reasoning behind the model prediction. In addition, we provide multiple adversarial attacks to compare the robustness of QA models. With these explainability methods and adversarial attacks, we aim to ease the research on trustworthy QA models. SQuARE is available on https://square.ukp-lab.de.

cs.CL↗

arXiv2022

Physical Interaction: Reconstructing Hand-object Interactions with Physics

Single view-based reconstruction of hand-object interaction is challenging due to the severe observation missing caused by occlusions. This paper proposes a physics-based method to better solve the ambiguities in the reconstruction. It first proposes a force-based dynamic model of the in-hand object, which not only recovers the unobserved contacts but also solves for plausible contact forces. Next, a confidence-based slide prevention scheme is proposed, which combines both the kinematic confidences and the contact forces to jointly model static and sliding contact motion. Qualitative and quantitative experiments show that the proposed technique reconstructs both physically plausible and more accurate hand-object interaction and estimates plausible contact forces in real-time with a single RGBD sensor.

cs.CV↗

arXiv2022

Plateau regions for zero-bias peaks within 5% of the quantized conductance value $2e^2/h$

Probing an isolated Majorana zero mode is predicted to reveal a tunneling conductance quantized at $2e^2/h$ at zero temperature. Experimentally, a zero-bias peak (ZBP) is expected and its height should remain robust against relevant parameter tuning, forming a quantized plateau. Here, we report the observation of large ZBPs in a thin InAs-Al hybrid nanowire device. The ZBP height can stick close to $2e^2/h$, mostly within $5\%$ tolerance, by sweeping gate voltages and magnetic field. We further map out the phase diagram and identify two plateau regions in the phase space. Despite the presence of disorder and quantum dots, our result constitutes a step forward towards establishing Majorana zero modes.

cond-mat.mes-hall↗

arXiv2022

AMP: Automatically Finding Model Parallel Strategies with Heterogeneity Awareness

Scaling up model sizes can lead to fundamentally new capabilities in many machine learning (ML) tasks. However, training big models requires strong distributed system expertise to carefully design model-parallel execution strategies that suit the model architectures and cluster setups. In this paper, we develop AMP, a framework that automatically derives such strategies. AMP identifies a valid space of model parallelism strategies and efficiently searches the space for high-performed strategies, by leveraging a cost model designed to capture the heterogeneity of the model and cluster specifications. Unlike existing methods, AMP is specifically tailored to support complex models composed of uneven layers and cluster setups with more heterogeneous accelerators and bandwidth. We evaluate AMP on popular models and cluster setups from public clouds and show that AMP returns parallel strategies that match the expert-tuned strategies on typical cluster setups. On heterogeneous clusters or models with heterogeneous architectures, AMP finds strategies with 1.54x and 1.77x higher throughput than state-of-the-art model-parallel systems, respectively.

cs.LG↗

arXiv2022

Application of Deep Learning on Single-Cell RNA-sequencing Data Analysis: A Review

Single-cell RNA-sequencing (scRNA-seq) has become a routinely used technique to quantify the gene expression profile of thousands of single cells simultaneously. Analysis of scRNA-seq data plays an important role in the study of cell states and phenotypes, and has helped elucidate biological processes, such as those occurring during development of complex organisms and improved our understanding of disease states, such as cancer, diabetes, and COVID, among others. Deep learning, a recent advance of artificial intelligence that has been used to address many problems involving large datasets, has also emerged as a promising tool for scRNA-seq data analysis, as it has a capacity to extract informative, compact features from noisy, heterogeneous, and high-dimensional scRNA-seq data to improve downstream analysis. The present review aims at surveying recently developed deep learning techniques in scRNA-seq data analysis, identifying key steps within the scRNA-seq data analysis pipeline that have been advanced by deep learning, and explaining the benefits of deep learning over more conventional analysis tools. Finally, we summarize the challenges in current deep learning approaches faced within scRNA-seq data and discuss potential directions for improvements in deep algorithms for scRNA-seq data analysis.

q-bio.GN↗

arXiv2022

The $p$-adic Gelfand-Kapranov-Zelevinsky hypergeometric complex

To a torus action on a complex vector space, Gelfand, Kapranov and Zelevinsky introduce a system of differential equations, which are now called the GKZ hypergeometric system. Its solutions are GKZ hypergeometric functions. We study the $p$-adic counterpart of the GKZ hypergeometric system. The $p$-adic GKZ hypergeometric complex is a twisted relative de Rham complex of over-convergent differential forms with logarithmic poles. It is an over-holonomic object in the derived category of arithmetic $\mathcal D$-modules with Frobenius structures. Traces of Frobenius on fibers at Techmüller points of the GKZ hypergeometric complex define the hypergeometric function over the finite field introduced by Gelfand and Graev. Over the non-degenerate locus, the GKZ hypergeometric complex defines an over-convergent $F$-isocrystal. It is the crystalline companion of the $\ell$-adic GKZ hypergeometric sheaf that we constructed before. Our method is a combination of Dwork's theory and the theory of arithmetic $\mathcal D$-modules of Berthelot.

math.AG↗

arXiv2022

A quantum spin liquid phase in the Kitaev-Hubbard model

The quantum spin liquid (QSL) state has been searched intensively in Kitaev-like materials, such as the Iridium oxides $A_2$IrO$_3$ and $α$-RuCl$_3$. The half-filled Kitaev-Hubbard model with bond dependent hopping terms is used to describe the Kitaev-like materials, which is calculated using the state-of-the-art fermionic projected entangled pair states (fPEPS) method. We find a QSL phase near the Mott insulator transition, which has a strong first-order transition to the semi-metal phase with the decrease of Hubbard $U$. We suggest that a promising routine to find the QSL is to find the Iridium oxides that are near the Mott insulator transitions.

cond-mat.str-el↗

arXiv2022

Congruences for modular forms and applications to crank functions

In this paper, motivated by the work of Mahlburg, we find congruences for a large class of modular forms. Moreover, we generalize the generating function of the Andrews-Garvan-Dyson crank on partition and establish several new infinite families of congruences. In this framework, we showed that both the birank of an ordered pair of partitions introduced by Hammond and Lewis, and $k$-crank of $k$-colored partition introduced by Fu and Tang process the same as the partition function and crank.

math.NT↗

arXiv2022

Squeezed-Light-Enhanced Dispersive Gyroscope based Optical Microcavities

Optical gyroscope based on the Sagnac effect have excellent potential in the application of high-sensitivity inertial rotation sensors. In this paper, we demonstrate that for an optical resonance gyroscope with normal dispersion, the measurement sensitivity can be increased by two orders of magnitude through coupling into a squeezed vacuum light, which is different from that in the classical situation. When the system is operated under critical anomalous dispersion condition, injecting a squeezed vacuum light allows the measurement sensitivity beyond the corresponding standard quantum limit by five orders of magnitude, with a minimum value of 3.8*10^-5 Hz. This work offers a promising possibility for developing optical gyroscopes that combine high sensitivity with tiny size.

physics.optics↗

arXiv2022

Distribution of Dirichlet $L$-functions

In this article, we study the distribution of values of Dirichlet $L$-functions, the distribution of values of the random models for Dirichlet $L$-functions, and the discrepancy between these two kinds of distributions. For each question, we consider the cases of $\frac12<\RE s<1$ and $\RE s=1$ separately.

math.NT↗

arXiv2022

Learning Reconstructability for Drone Aerial Path Planning

We introduce the first learning-based reconstructability predictor to improve view and path planning for large-scale 3D urban scene acquisition using unmanned drones. In contrast to previous heuristic approaches, our method learns a model that explicitly predicts how well a 3D urban scene will be reconstructed from a set of viewpoints. To make such a model trainable and simultaneously applicable to drone path planning, we simulate the proxy-based 3D scene reconstruction during training to set up the prediction. Specifically, the neural network we design is trained to predict the scene reconstructability as a function of the proxy geometry, a set of viewpoints, and optionally a series of scene images acquired in flight. To reconstruct a new urban scene, we first build the 3D scene proxy, then rely on the predicted reconstruction quality and uncertainty measures by our network, based off of the proxy geometry, to guide the drone path planning. We demonstrate that our data-driven reconstructability predictions are more closely correlated to the true reconstruction quality than prior heuristic measures. Further, our learned predictor can be easily integrated into existing path planners to yield improvements. Finally, we devise a new iterative view planning framework, based on the learned reconstructability, and show superior performance of the new planner when reconstructing both synthetic and real scenes.

cs.GR↗

arXiv2022

Parameterization of Cross-Token Relations with Relative Positional Encoding for Vision MLP

Vision multi-layer perceptrons (MLPs) have shown promising performance in computer vision tasks, and become the main competitor of CNNs and vision Transformers. They use token-mixing layers to capture cross-token interactions, as opposed to the multi-head self-attention mechanism used by Transformers. However, the heavily parameterized token-mixing layers naturally lack mechanisms to capture local information and multi-granular non-local relations, thus their discriminative power is restrained. To tackle this issue, we propose a new positional spacial gating unit (PoSGU). It exploits the attention formulations used in the classical relative positional encoding (RPE), to efficiently encode the cross-token relations for token mixing. It can successfully reduce the current quadratic parameter complexity $O(N^2)$ of vision MLPs to $O(N)$ and $O(1)$. We experiment with two RPE mechanisms, and further propose a group-wise extension to improve their expressive power with the accomplishment of multi-granular contexts. These then serve as the key building blocks of a new type of vision MLP, referred to as PosMLP. We evaluate the effectiveness of the proposed approach by conducting thorough experiments, demonstrating an improved or comparable performance with reduced parameter complexity. For instance, for a model trained on ImageNet1K, we achieve a performance improvement from 72.14\% to 74.02\% and a learnable parameter reduction from $19.4M$ to $18.2M$. Code could be found at https://github.com/Zhicaiwww/PosMLP.

cs.CV↗

arXiv2022

Optomechanical compensatory cooling mechanism with exceptional points

The ground state cooling of Brillouin scattering optomechanical system is limited by defects in practical sample. In this paper, we propose a new compensatory cooling mechanism for Brillouin scattering optomechanical system with exceptional points (EPs). By using the EPs both in optical and mechanical modes, the limited cooling process is compensated effectively. The dual-EPs system, which is discovered in this work for the first time, can be induced by two defects with specific relative angles and has function of not only actively manipulating the coupling strength of optical modes but also the Brillouin phonon modes. Our results provide new tools to manipulate the optomechanical interaction in multi-mode systems and open the possibility of quantum state transfer and quantum interface protocols based on phonon cooling in quantum applications.

physics.optics↗

arXiv2022

MINERVAS: Massive INterior EnviRonments VirtuAl Synthesis

With the rapid development of data-driven techniques, data has played an essential role in various computer vision tasks. Many realistic and synthetic datasets have been proposed to address different problems. However, there are lots of unresolved challenges: (1) the creation of dataset is usually a tedious process with manual annotations, (2) most datasets are only designed for a single specific task, (3) the modification or randomization of the 3D scene is difficult, and (4) the release of commercial 3D data may encounter copyright issue. This paper presents MINERVAS, a Massive INterior EnviRonments VirtuAl Synthesis system, to facilitate the 3D scene modification and the 2D image synthesis for various vision tasks. In particular, we design a programmable pipeline with Domain-Specific Language, allowing users to (1) select scenes from the commercial indoor scene database, (2) synthesize scenes for different tasks with customized rules, and (3) render various imagery data, such as visual color, geometric structures, semantic label. Our system eases the difficulty of customizing massive numbers of scenes for different tasks and relieves users from manipulating fine-grained scene configurations by providing user-controllable randomness using multi-level samplers. Most importantly, it empowers users to access commercial scene databases with millions of indoor scenes and protects the copyright of core data assets, e.g., 3D CAD models. We demonstrate the validity and flexibility of our system by using our synthesized data to improve the performance on different kinds of computer vision tasks.

cs.CV↗

arXiv2022

DiscrimLoss: A Universal Loss for Hard Samples and Incorrect Samples Discrimination

Given data with label noise (i.e., incorrect data), deep neural networks would gradually memorize the label noise and impair model performance. To relieve this issue, curriculum learning is proposed to improve model performance and generalization by ordering training samples in a meaningful (e.g., easy to hard) sequence. Previous work takes incorrect samples as generic hard ones without discriminating between hard samples (i.e., hard samples in correct data) and incorrect samples. Indeed, a model should learn from hard samples to promote generalization rather than overfit to incorrect ones. In this paper, we address this problem by appending a novel loss function DiscrimLoss, on top of the existing task loss. Its main effect is to automatically and stably estimate the importance of easy samples and difficult samples (including hard and incorrect samples) at the early stages of training to improve the model performance. Then, during the following stages, DiscrimLoss is dedicated to discriminating between hard and incorrect samples to improve the model generalization. Such a training strategy can be formulated dynamically in a self-supervised manner, effectively mimicking the main principle of curriculum learning. Experiments on image classification, image regression, text sequence regression, and event relation reasoning demonstrate the versatility and effectiveness of our method, particularly in the presence of diversified noise levels.

cs.LG↗

arXiv2022

Probing electron-hole weights of an Andreev bound state by transient currents

Andreev bound states (ABSs) are localized quantum states that contain both electron and hole components. They ubiquitously reside in inhomogeneous superconducting systems. Following theoretical analysis, we propose to probe the electron-hole weights of an ABS via a local tunneling measurement that detects the transient current under a steplike pulse bias. With our protocol, the ABS energy level can also be obtained from peaks of the Fourier spectrum of the transient current. Our protocol can be applied to detect robust zero-energy Majorana bound states (MBSs), which have equal electron-hole weights, in candidate platforms where local tunneling spectroscopy measurement is possible. In the 1D Majorana nanowire model, we numerically calculate the electron-hole weights for different types of low-energy bound states, including ABSs, quasi-MBSs, and MBSs.

cond-mat.mes-hall↗