arXiv subjects

Zhe Huang

Publications and source records attributed to Zhe Huang.

At least 55 records · Page 3Linked to original sources

arXiv2024

MGTR: Multi-Granular Transformer for Motion Prediction with LiDAR

Motion prediction has been an essential component of autonomous driving systems since it handles highly uncertain and complex scenarios involving moving agents of different types. In this paper, we propose a Multi-Granular TRansformer (MGTR) framework, an encoder-decoder network that exploits context features in different granularities for different kinds of traffic agents. To further enhance MGTR's capabilities, we leverage LiDAR point cloud data by incorporating LiDAR semantic features from an off-the-shelf LiDAR feature extractor. We evaluate MGTR on Waymo Open Dataset motion prediction benchmark and show that the proposed method achieved state-of-the-art performance, ranking 1st on its leaderboard (https://waymo.com/open/challenges/2023/motion-prediction/).

cs.CV↗

arXiv2023

Learning Task Skills and Goals Simultaneously from Physical Interaction

In real-world human-robot systems, it is essential for a robot to comprehend human objectives and respond accordingly while performing an extended series of motor actions. Although human objective alignment has recently emerged as a promising paradigm in the realm of physical human-robot interaction, its application is typically confined to generating simple motions due to inherent theoretical limitations. In this work, our goal is to develop a general formulation to learn manipulation functional modules and long-term task goals simultaneously from physical human-robot interaction. We show the feasibility of our framework in enabling robots to align their behaviors with the long-term task objectives inferred from human interactions.

cs.RO↗

arXiv2023

Hierarchical Intention Tracking for Robust Human-Robot Collaboration in Industrial Assembly Tasks

Collaborative robots require effective human intention estimation to safely and smoothly work with humans in less structured tasks such as industrial assembly, where human intention continuously changes. We propose the concept of intention tracking and introduce a collaborative robot system that concurrently tracks intentions at hierarchical levels. The high-level intention is tracked to estimate human's interaction pattern and enable robot to (1) avoid collision with human to minimize interruption and (2) assist human to correct failure. The low-level intention estimate provides robot with task-related information. We implement the system on a UR5e robot and demonstrate robust, seamless and ergonomic human-robot collaboration in an ablative pilot study of an assembly use case. Our robot demonstrations and videos are available at \url{https://sites.google.com/view/hierarchicalintentiontracking}.

cs.RO↗

arXiv2023

Towards Safe Multi-Level Human-Robot Interaction in Industrial Tasks

Multiple levels of safety measures are required by multiple interaction modes which collaborative robots need to perform industrial tasks with human co-workers. We develop three independent modules to account for safety in different types of human-robot interaction: vision-based safety monitoring pauses robot when human is present in a shared space; contact-based safety monitoring pauses robot when unexpected contact happens between human and robot; hierarchical intention tracking keeps robot in a safe distance from human when human and robot work independently, and switches robot to compliant mode when human intends to guide robot. We discuss the prospect of future research in development and integration of multi-level safety modules. We focus on how to provide safety guarantees for collaborative robot solutions with human behavior modeling.

cs.RO↗

arXiv2023

Workshop on Document Intelligence Understanding

Document understanding and information extraction include different tasks to understand a document and extract valuable information automatically. Recently, there has been a rising demand for developing document understanding among different domains, including business, law, and medicine, to boost the efficiency of work that is associated with a large number of documents. This workshop aims to bring together researchers and industry developers in the field of document intelligence and understanding diverse document types to boost automatic document processing and understanding techniques. We also released a data challenge on the recently introduced document-level VQA dataset, PDFVQA. The PDFVQA challenge examines the structural and contextual understandings of proposed models on the natural full document level of multiple consecutive document pages by including questions with a sequence of answers extracted from multi-pages of the full document. This task helps to boost the document understanding step from the single-page level to the full document level understanding.

cs.IR↗

arXiv2023

FSD: Fully-Specialized Detector via Neural Architecture Search

Most generic object detectors are mainly built for standard object detection tasks such as COCO and PASCAL VOC. They might not work well and/or efficiently on tasks of other domains consisting of images that are visually different from standard datasets. To this end, many advances have been focused on adapting a general-purposed object detector with limited domain-specific designs. However, designing a successful task-specific detector requires extraneous manual experiments and parameter tuning through trial and error. In this paper, we first propose and examine a fully-automatic pipeline to design a fully-specialized detector (FSD) which mainly incorporates a neural-architectural-searched model by exploring ideal network structures over the backbone and task-specific head. On the DeepLesion dataset, extensive results show that FSD can achieve 3.1 mAP gain while using approximately 40% fewer parameters on binary lesion detection task and improved the mAP by around 10% on multi-type lesion detection task via our region-aware graph modeling compared with existing general-purposed medical lesion detection networks.

cs.CV↗

arXiv2023

User-Friendly Safety Monitoring System for Manufacturing Cobots

Collaborative robots are being increasingly utilized in industrial production lines due to their efficiency and accuracy. However, the close proximity between humans and robots can pose safety risks due to the robot's high-speed movements and powerful forces. To address this, we developed a vision-based safety monitoring system that creates a 3D reconstruction of the collaborative scene. Our system records the human-robot interaction data in real-time and reproduce their virtual replicas in a simulator for offline analysis. The objective is to provide workers with a user-friendly visualization tool for reviewing performance and diagnosing failures, thereby enhancing safety in manufacturing settings.

cs.RO↗

arXiv2023

Kagome surface states and weak electronic correlation in vanadium-kagome metals

RV6Sn6 (R = Y and lanthanides) with two-dimensional vanadium-kagome surface states is an ideal platform to investigate kagome physics and manipulate the kagome features to realize novel phenomena. Utilizing the micron-scale spatially resolved angle-resolved photoemission spectroscopy and first-principles calculations, we report a systematical study of the electronic structures of RV6Sn6 (R = Gd, Tb, and Lu) on the two cleaved surfaces, i.e., the V- and RSn1-terminated (001) surfaces. The calculated bands without any renormalization match well with the main ARPES dispersive features, indicating the weak electronic correlation in this system. We observe 'W'-like kagome surface states around the Brillouin zone corners showing R-element-dependent intensities, which is probably due to various coupling strengths between V and RSn1 layers. Our finding suggests an avenue for tuning electronic states by interlayer coupling based on two-dimensional kagome lattices.

cond-mat.mtrl-sci↗

arXiv2023

Observation of electronic nematicity driven by three-dimensional charge density wave in kagome lattice KV$_3$Sb$_5$

Kagome superconductors AV$_3$Sb$_5$ (A = K, Rb, Cs) provide a fertile playground for studying intriguing phenomena, including non-trivial band topology, superconductivity, giant anomalous Hall effect and charge density wave (CDW). Recently, a $C_2$ symmetric nematic phase prior to the superconducting state in AV$_3$Sb$_5$ drew enormous attention due to its potential inheritance of the symmetry of the unusual superconductivity. However, direct evidence on the rotation symmetry breaking of the electronic structure in the CDW state from the reciprocal space is still rare, and the underlying mechanism remains ambiguous. The observation shows unconventional unidirectionality, indicative of rotation symmetry breaking from six-fold to two-fold. The interlayer coupling between adjacent planes with $π$-phase offset in the 2$\times$2$\times$2 CDW phase leads to the preferred two-fold symmetric electronic structure. These rarely observed unidirectional back-folded bands in KV$_3$Sb$_5$ may provide important insights into its peculiar charge order and superconductivity.

cond-mat.supr-con↗

arXiv2023

Fix-A-Step: Semi-supervised Learning from Uncurated Unlabeled Data

Semi-supervised learning (SSL) promises improved accuracy compared to training classifiers on small labeled datasets by also training on many unlabeled images. In real applications like medical imaging, unlabeled data will be collected for expediency and thus uncurated: possibly different from the labeled set in classes or features. Unfortunately, modern deep SSL often makes accuracy worse when given uncurated unlabeled data. Recent complex remedies try to detect out-of-distribution unlabeled images and then discard or downweight them. Instead, we introduce Fix-A-Step, a simpler procedure that views all uncurated unlabeled images as potentially helpful. Our first insight is that even uncurated images can yield useful augmentations of labeled data. Second, we modify gradient descent updates to prevent optimizing a multi-task SSL loss from hurting labeled-set accuracy. Fix-A-Step can repair many common deep SSL methods, improving accuracy on CIFAR benchmarks across all tested methods and levels of artificial class mismatch. On a new medical SSL benchmark called Heart2Heart, Fix-A-Step can learn from 353,500 truly uncurated ultrasound images to deliver gains that generalize across hospitals.

cs.LG↗

arXiv2023

Intention Aware Robot Crowd Navigation with Attention-Based Interaction Graph

We study the problem of safe and intention-aware robot navigation in dense and interactive crowds. Most previous reinforcement learning (RL) based methods fail to consider different types of interactions among all agents or ignore the intentions of people, which results in performance degradation. To learn a safe and efficient robot policy, we propose a novel recurrent graph neural network with attention mechanisms to capture heterogeneous interactions among agents through space and time. To encourage longsighted robot behaviors, we infer the intentions of dynamic agents by predicting their future trajectories for several timesteps. The predictions are incorporated into a model-free RL framework to prevent the robot from intruding into the intended paths of other agents. We demonstrate that our method enables the robot to achieve good navigation performance and non-invasiveness in challenging crowd navigation scenarios. We successfully transfer the policy learned in simulation to a real-world TurtleBot 2i. Our code and videos are available at https://sites.google.com/view/intention-aware-crowdnav/home.

cs.RO↗

arXiv2023

AirBirds: A Large-scale Challenging Dataset for Bird Strike Prevention in Real-world Airports

One fundamental limitation to the research of bird strike prevention is the lack of a large-scale dataset taken directly from real-world airports. Existing relevant datasets are either small in size or not dedicated for this purpose. To advance the research and practical solutions for bird strike prevention, in this paper, we present a large-scale challenging dataset AirBirds that consists of 118,312 time-series images, where a total of 409,967 bounding boxes of flying birds are manually, carefully annotated. The average size of all annotated instances is smaller than 10 pixels in 1920x1080 images. Images in the dataset are captured over 4 seasons of a whole year by a network of cameras deployed at a real-world airport, covering diverse bird species, lighting conditions and 13 meteorological scenarios. To the best of our knowledge, it is the first large-scale image dataset that directly collects flying birds in real-world airports for bird strike prevention. This dataset is publicly available at https://airbirdsdata.github.io/.

cs.CV↗

arXiv2023

Capacity-CRB Tradeoff in OFDM Integrated Sensing and Communication Systems

Integrated sensing and communication (ISAC) has emerged as a key technology for future communication systems. In this paper, we provide a general framework to reveal the fundamental tradeoff between sensing and communication in OFDM systems, where a unified ISAC waveform is exploited to perform both tasks. In particular, we define the Capacity-Bayesian Cramer Rao Bound (BCRB) region in the asymptotically case when the number of subcarriers is large. Specifically, we show that the asymptotically optimal input distribution that achieves the Pareto boundary point of the Capacity-BCRB region is Gaussian and the entire Pareto boundary can be obtained by solving a convex power allocation problem. Moreover, we characterize the structure of the sensing-optimal power allocation in the asymptotically case. Finally, numerical simulations are conducted to verify the theoretical analysis and provide useful insights.

cs.IT↗

arXiv2022

Flat bands, non-trivial band topology and electronic nematicity in layered kagome-lattice RbTi$_3$Bi$_5$

Layered kagome-lattice materials with 3$d$ transition metals provide a fertile playground for studies on geometry frustration, band topology and other novel ordered states. A representative class of materials AV$_3$Sb$_5$ (A=K, Rb, Cs) have been proved to possess various unconventional phases such as superconductivity, non-trivial $\mathbb{Z}_2$ band topology, and electronic nematicity, which are intertwined with multiple interlaced charge density waves (CDW). However, the interplay among these novel states and their mechanisms are still elusive. Recently, the discovery of isostructural titanium-based single-crystals ATi$_3$Bi$_5$ (A=K, Rb, Cs), which demonstrate similar multiple exotic states but in the absence of the concomitant intertwined CDW, has been offering an ideal opportunity to disentangle these complex novel states in kagome-lattice. Here, we combine the high-resolution angle-resolved photoemission spectroscopy and first-principles calculations to systematically investigate the low-lying electronic structure of RbTi$_3$Bi$_5$. For the first time, we experimentally demonstrate the coexistence of flat bands and multiple non-trivial topological states, including type-II Dirac nodal lines and non-trivial $\mathbb{Z}_2$ topological surface states therein. Furthermore, our findings as well provide the hint of rotation symmetry breaking in RbTi$_3$Bi$_5$, suggesting the directionality of the electronic structure and possibility of emerging pure electronic nematicity in this new family of kagome compounds, which may provide important insights into the electronic nematic phase in correlated kagome metals.

cond-mat.str-el↗

arXiv2022

Metal to Mott Insulator Transition in Two-dimensional 1T-TaSe$_2$

When electron-electron interaction dominates over other electronic energy scales, exotic, collective phenomena often emerge out of seemingly ordinary matter. The strongly correlated phenomena, such as quantum spin liquid and unconventional superconductivity, represent a major research frontier and a constant source of inspiration. Central to strongly correlated physics is the concept of Mott insulator, from which various other correlated phases derive. The advent of two-dimensional (2D) materials brings unprecedented opportunities to the study of strongly correlated physics in the 2D limit. In particular, the enhanced correlation and extreme tunability of 2D materials enables exploring strongly correlated systems across uncharted parameter space. Here, we discover an intriguing metal to Mott insulator transition in 1T-TaSe$_2$ as the material is thinned down to atomic thicknesses. Specifically, we discover, for the first time, that the bulk metallicity of 1T-TaSe$_2$ arises from a band crossing Fermi level. Reducing the dimensionality effectively quenches the kinetic energy of the initially itinerant electrons and drives the material into a Mott insulating state. The dimensionality-driven Metal to Mott insulator transition resolves the long-standing dichotomy between metallic bulk and insulating surface of 1T-TaSe$_2$. Our results additionally establish 1T-TaSe$_2$ as an ideal variable system for exploring various strongly correlated phenomena.

cond-mat.str-el↗

arXiv2022

Data Encryption based on 7D Complex Chaotic System with Cubic Memristor for Smart Grid

The information security has an irreplaceable position in the smart grid (SG). In order to avoid the malicious attack and ensure the information security, the cryptographic techniques are essential. This paper focuses on the encryption techniques to ensure the information security of SG. Firstly, an unusual 7-dimensional complex chaotic system (7D-CCS) combined with the cubic memristor is introduced. Besides its phase portraits, Lyapunov exponent, 0-1 test, complexity, and bifurcation diagram are investigated. Then, with the proposed 7D-CCS, we design a data encryption algorithm to ensure the encryption security. Finally, the data and monitoring images in SG are encrypted by the designed encryption scheme. Besides, the encryption performance is given in detailed. The experimental results show that the proposed encryption scheme has quite good encryption performance. Therefore, it can ensure the information security of SG.

cs.CR↗

arXiv2022

An Input-Aware Mimic Defense Theory and its Practice

The current security problems in cyberspace are characterized by strong and complex threats. Defenders face numerous problems such as lack of prior knowledge, various threats, and unknown vulnerabilities, which urgently need new fundamental theories to support. To address these issues, this article proposes a generic theoretical model for cyberspace defense and a new mimic defense framework, that is, Spatiotemporally heterogeneous, Input aware, and Dynamically updated Mimic Defense (SIDMD). We make the following contributions: (1) We first redefine vulnerabilities from the input space perspective to normalize the diverse cyberspace security problem. (2) We propose a novel unknown vulnerability discovery method and a dynamic scheduling strategy considering temporal and spatial dimensions without prior knowledge. Theoretical analysis and experimental results show that SIDMD has the best security performance in complex attack scenarios, and the probability of successful attacks is greatly reduced compared to the state-of-the-art.

cs.CR↗

arXiv2022

Addressing Optimism Bias in Sequence Modeling for Reinforcement Learning

Impressive results in natural language processing (NLP) based on the Transformer neural network architecture have inspired researchers to explore viewing offline reinforcement learning (RL) as a generic sequence modeling problem. Recent works based on this paradigm have achieved state-of-the-art results in several of the mostly deterministic offline Atari and D4RL benchmarks. However, because these methods jointly model the states and actions as a single sequencing problem, they struggle to disentangle the effects of the policy and world dynamics on the return. Thus, in adversarial or stochastic environments, these methods lead to overly optimistic behavior that can be dangerous in safety-critical systems like autonomous driving. In this work, we propose a method that addresses this optimism bias by explicitly disentangling the policy and world models, which allows us at test time to search for policies that are robust to multiple possible futures in the environment. We demonstrate our method's superior performance on a variety of autonomous driving tasks in simulation.

cs.LG↗