arXiv subjects

Liang Li

Publications and source records attributed to Liang Li.

At least 217 records · Page 12Linked to original sources

arXiv2023

LESS-Map: Lightweight and Evolving Semantic Map in Parking Lots for Long-term Self-Localization

Precise and long-term stable localization is essential in parking lots for tasks like autonomous driving or autonomous valet parking, \textit{etc}. Existing methods rely on a fixed and memory-inefficient map, which lacks robust data association approaches. And it is not suitable for precise localization or long-term map maintenance. In this paper, we propose a novel mapping, localization, and map update system based on ground semantic features, utilizing low-cost cameras. We present a precise and lightweight parameterization method to establish improved data association and achieve accurate localization at centimeter-level. Furthermore, we propose a novel map update approach by implementing high-quality data association for parameterized semantic features, allowing continuous map update and refinement during re-localization, while maintaining centimeter-level accuracy. We validate the performance of the proposed method in real-world experiments and compare it against state-of-the-art algorithms. The proposed method achieves an average accuracy improvement of 5cm during the registration process. The generated maps consume only a compact size of 450 KB/km and remain adaptable to evolving environments through continuous update.

cs.RO↗

arXiv2023

Self-supervised Cross-view Representation Reconstruction for Change Captioning

Change captioning aims to describe the difference between a pair of similar images. Its key challenge is how to learn a stable difference representation under pseudo changes caused by viewpoint change. In this paper, we address this by proposing a self-supervised cross-view representation reconstruction (SCORER) network. Concretely, we first design a multi-head token-wise matching to model relationships between cross-view features from similar/dissimilar images. Then, by maximizing cross-view contrastive alignment of two similar images, SCORER learns two view-invariant image representations in a self-supervised way. Based on these, we reconstruct the representations of unchanged objects by cross-attention, thus learning a stable difference representation for caption generation. Further, we devise a cross-modal backward reasoning to improve the quality of caption. This module reversely models a ``hallucination'' representation with the caption and ``before'' representation. By pushing it closer to the ``after'' representation, we enforce the caption to be informative about the difference in a self-supervised manner. Extensive experiments show our method achieves the state-of-the-art results on four datasets. The code is available at https://github.com/tuyunbin/SCORER.

cs.CV↗

arXiv2023

KDD-LOAM: Jointly Learned Keypoint Detector and Descriptors Assisted LiDAR Odometry and Mapping

Sparse keypoint matching based on distinct 3D feature representations can improve the efficiency and robustness of point cloud registration. Existing learning-based 3D descriptors and keypoint detectors are either independent or loosely coupled, so they cannot fully adapt to each other. In this work, we propose a tightly coupled keypoint detector and descriptor (TCKDD) based on a multi-task fully convolutional network with a probabilistic detection loss. In particular, this self-supervised detection loss fully adapts the keypoint detector to any jointly learned descriptors and benefits the self-supervised learning of descriptors. Extensive experiments on both indoor and outdoor datasets show that our TCKDD achieves state-of-the-art performance in point cloud registration. Furthermore, we design a keypoint detector and descriptors-assisted LiDAR odometry and mapping framework (KDD-LOAM), whose real-time odometry relies on keypoint descriptor matching-based RANSAC. The sparse keypoints are further used for efficient scan-to-map registration and mapping. Experiments on KITTI dataset demonstrate that KDD-LOAM significantly surpasses LOAM and shows competitive performance in odometry.

cs.CV↗

arXiv2023

Colossal linear magnetoelectricity in polar magnet Fe2Mo3O8

Linear magnetoelectric effect is an attractive phenomenon in condensed matters and provides indispensable technological functionalities. Here a colossal linear magnetoelectric effect with diagonal component alfa_33 reaching up to ~480 ps/m is reported in a polar magnet Fe2Mo3O8, and this effect can persist in a broad range of magnetic field (~20 T) and is orders of magnitude larger than reported values in literature. Such an exceptional experimental observation can be well reproduced by a theoretical model affirmatively unveiling the vital contributions from the exchange striction, while the sign difference of magnetocrystalline anisotropy can also be reasonably figured out.

cond-mat.mtrl-sci↗

arXiv2023

Response to "On the giant deformation and ferroelectricity of guanidinium nitrate" by Marek Szafrański and Andrzej Katrusiak

Following a well-established practice of publishing commentaries to articles of other authors who work on materials that were earlier studied by them (n.b. six published comments[1-6]), Marek Szafrański(MS) and Andrzej Katrusiak (AK) have filed on the preprint server arXiv a manuscript entitled "On the giant deformation and ferroelectricity of guanidinium nitrate"[7] with comments on our article "Exceptionally high work density of a ferroelectric dynamic organic crystal around room temperature" published in Nature Communications (2022, 13, 2823).[8] Both in the submitted comment as well as in the required (by the journal) direct communication with us preceding its posting, MS and AK have expressed dissatisfaction with the choice of literature references in our article, for which they felt that their previous work on this material has not been cited to a sufficient extent. In their comment, they summarize their other remarks on our article as "the structural determinations of GN [guanidinium nitrate] crystals, their phase transitions and associated giant deformation, as well as its detailed structural mechanism, the molecular dynamics and dielectric properties were reported before, while the semiconductivity, ferroelectricity, and fatigue resistance of the GN [guanidinium nitrate] crystals cannot be confirmed."[7] Apart from the sentiments of MS and AK on our choice of cited literature, we find their comments on the scientific content of our article to be strongly biased towards their own results and unfounded. Below, we provide a detailed response to their comments.

cond-mat.mtrl-sci↗

arXiv2023

Separate and Locate: Rethink the Text in Text-based Visual Question Answering

Text-based Visual Question Answering (TextVQA) aims at answering questions about the text in images. Most works in this field focus on designing network structures or pre-training tasks. All these methods list the OCR texts in reading order (from left to right and top to bottom) to form a sequence, which is treated as a natural language ``sentence''. However, they ignore the fact that most OCR words in the TextVQA task do not have a semantical contextual relationship. In addition, these approaches use 1-D position embedding to construct the spatial relation between OCR tokens sequentially, which is not reasonable. The 1-D position embedding can only represent the left-right sequence relationship between words in a sentence, but not the complex spatial position relationship. To tackle these problems, we propose a novel method named Separate and Locate (SaL) that explores text contextual cues and designs spatial position embedding to construct spatial relations between OCR texts. Specifically, we propose a Text Semantic Separate (TSS) module that helps the model recognize whether words have semantic contextual relations. Then, we introduce a Spatial Circle Position (SCP) module that helps the model better construct and reason the spatial position relationships between OCR texts. Our SaL model outperforms the baseline model by 4.44% and 3.96% accuracy on TextVQA and ST-VQA datasets. Compared with the pre-training state-of-the-art method pre-trained on 64 million pre-training samples, our method, without any pre-training tasks, still achieves 2.68% and 2.52% accuracy improvement on TextVQA and ST-VQA. Our code and models will be released at https://github.com/fangbufang/SaL.

cs.CV↗

arXiv2023

FPTQ: Fine-grained Post-Training Quantization for Large Language Models

In the era of large-scale language models, the substantial parameter size poses significant challenges for deployment. Being a prevalent compression technique, quantization has emerged as the mainstream practice to tackle this issue, which is mainly centered on two recipes W8A8 and W4A16 (i.e. weights and activations in such bit widths). In this study, we propose a novel W4A8 post-training quantization method for the available open-sourced LLMs, which combines the advantages of both two recipes. Therefore, we can leverage the benefit in the I/O utilization of 4-bit weight quantization and the acceleration due to 8-bit matrix computation. Nevertheless, the W4A8 faces notorious performance degradation. As a remedy, we involve layerwise activation quantization strategies which feature a novel logarithmic equalization for most intractable layers, and we combine them with fine-grained weight quantization. Without whistles and bells, we eliminate the necessity for further fine-tuning and obtain the state-of-the-art W4A8 quantized performance on BLOOM, LLaMA, and LLaMA-2 on standard benchmarks. We confirm that the W4A8 quantization is achievable for the deployment of large language models, fostering their wide-spreading real-world applications.

cs.CL↗

arXiv2023

Joint Intrinsic and Extrinsic LiDAR-Camera Calibration in Targetless Environments Using Plane-Constrained Bundle Adjustment

This paper introduces a novel targetless method for joint intrinsic and extrinsic calibration of LiDAR-camera systems using plane-constrained bundle adjustment (BA). Our method leverages LiDAR point cloud measurements from planes in the scene, alongside visual points derived from those planes. The core novelty of our method lies in the integration of visual BA with the registration between visual points and LiDAR point cloud planes, which is formulated as a unified optimization problem. This formulation achieves concurrent intrinsic and extrinsic calibration, while also imparting depth constraints to the visual points to enhance the accuracy of intrinsic calibration. Experiments are conducted on both public data sequences and self-collected dataset. The results showcase that our approach not only surpasses other state-of-the-art (SOTA) methods but also maintains remarkable calibration accuracy even within challenging environments. For the benefits of the robotics community, we have open sourced our codes.

cs.RO↗

arXiv2023

Predator-prey survival pressure is sufficient to evolve swarming behaviors

The comprehension of how local interactions arise in global collective behavior is of utmost importance in both biological and physical research. Traditional agent-based models often rely on static rules that fail to capture the dynamic strategies of the biological world. Reinforcement learning has been proposed as a solution, but most previous methods adopt handcrafted reward functions that implicitly or explicitly encourage the emergence of swarming behaviors. In this study, we propose a minimal predator-prey coevolution framework based on mixed cooperative-competitive multiagent reinforcement learning, and adopt a reward function that is solely based on the fundamental survival pressure, that is, prey receive a reward of $-1$ if caught by predators while predators receive a reward of $+1$. Surprisingly, our analysis of this approach reveals an unexpectedly rich diversity of emergent behaviors for both prey and predators, including flocking and swirling behaviors for prey, as well as dispersion tactics, confusion, and marginal predation phenomena for predators. Overall, our study provides novel insights into the collective behavior of organisms and highlights the potential applications in swarm robotics.

q-bio.PE↗

arXiv2023

A fast tunable driver of light source for the TRIDENT Pathfinder experiment

TRIDENT (The tRopIcal DEep-sea Neutrino Telescope) is a proposed next-generation neutrino telescope to be constructed in the South China Sea. In September 2021, the TRIDENT Pathfinder experiment (TRIDENT EXplorer, T-REX for short) was conducted to evaluate the in-situ optical properties of seawater. The T-REX experiment deployed three digital optical modules at a depth of 3420 meters, including a light emitter module (LEM) and two light receiver modules (LRMs) equipped with photomultiplier tubes (PMTs) and cameras to detect light signals. The LEM emits light in pulsing and steady modes. It features a fast tunable driver to activate light-emitting diodes (LEDs) that emit nanosecond-width light pulses with tunable intensity. The PMTs in the LRM receive single photo-electron (SPE) signals with an average photon number of approximately 0.3 per 1-microsecond time window, which is used to measure the arrival time distribution of the SPE signals. The fast tunable driver can be remotely controlled in real-time by the data acquisition system onboard the research vessel, allowing for convenient adjustments to the driver's parameters and facilitating the acquisition of high-quality experimental data. This paper describes the requirements, design scheme, and test results of the fast tunable driver, highlighting its successful implementation in the T-REX experiment and its potential for future deep-sea experiments.

physics.ins-det↗

arXiv2023

GRB-SN Association within the Binary-Driven Hypernova Model

The observations of supernovae (SNe) Ic occurring after the prompt emission of long gamma-ray bursts (GRBs) are addressed within the binary-driven hypernova (BdHN) model where GRBs originate from a binary composed of a $\sim10M_\odot$ carbon-oxygen (CO) star and a neutron star (NS). The CO core collapse gives the trigger, leading to a hypernova with a fast-spinning newborn NS ($ν$NS) at its center. The evolution depends strongly on the binary period, $P_{\rm bin}$. For $P_{\rm bin}\sim5$min, BdHNe I occur with energies $10^{52}$--$10^{54}$erg. The accretion of SN ejecta onto the NS leads to its collapse, forming a black hole (BH) originating the MeV/GeV radiation. For $P_{\rm bin}\sim 10$min, BdHNe II occur with energies $10^{50}$--$10^{52}$erg and for $P_{\rm bin}\sim$hours, BdHN III occurs with energies below $10^{50}$erg. {In BdHNe II and III,} no BH is formed. The $1$--$1000$ms $ν$NS originates, in all BdHNe, the X-ray-optical-radio afterglows by synchrotron emission. The hypernova follows an independent evolution, becoming an SN Ic, powered by nickel decay, observable after the GRB prompt emission. We report $24$ SNe Ic associated with BdHNe. Their optical peak luminosity and time of occurrence are similar and independent of the associated GRBs. {From previously identified $380$ BdHN I comprising redshifts up to $z=8.2$, we analyze} four examples with their associated hypernovae. By multiwavelength extragalactic observations, we identify seven new Episodes, theoretically explained, fortunately not yet detected in galactic sources, opening new research areas. Refinement of population synthesis simulations is needed to map the progenitors of such short-lived binary systems inside our galaxy.

astro-ph.HE↗

arXiv2023

CATS: A Pragmatic Chinese Answer-to-Sequence Dataset with Large Scale and High Quality

There are three problems existing in the popular data-to-text datasets. First, the large-scale datasets either contain noise or lack real application scenarios. Second, the datasets close to real applications are relatively small in size. Last, current datasets bias in the English language while leaving other languages underexplored. To alleviate these limitations, in this paper, we present CATS, a pragmatic Chinese answer-to-sequence dataset with large scale and high quality. The dataset aims to generate textual descriptions for the answer in the practical TableQA system. Further, to bridge the structural gap between the input SQL and table and establish better semantic alignments, we propose a Unified Graph Transformation approach to establish a joint encoding space for the two hybrid knowledge resources and convert this task to a graph-to-text problem. The experiment results demonstrate the effectiveness of our proposed method. Further analysis on CATS attests to both the high quality and challenges of the dataset.

cs.CL↗

arXiv2023

An absolutely convergent fixed-point fast sweeping WENO method on triangular meshes for steady state of hyperbolic conservation laws

High order fast sweeping methods for efficiently solving steady state solutions of hyperbolic PDEs were not available yet on unstructured meshes. In this paper, we extend high order fast sweeping methods to unstructured triangular meshes by applying fixed-point iterative sweeping techniques to a fifth-order finite volume unstructured WENO scheme, for solving steady state solutions of hyperbolic conservation laws. An advantage of fixed-point fast sweeping methods which distinguishes them from other fast sweeping methods is that they are explicit and do not require inverse operation of nonlinear local systems. As in the first order fast sweeping methods on triangular meshes, we introduce multiple reference points to determine alternating sweeping directions on unstructured meshes. All the cells on the mesh are ordered according to their centroids' distances to those reference points, and the resulted orderings provide sweeping directions for iterations. To make the residue of the fast sweeping iterations converge to machine zero / round off errors, we follow the approach in our early work of developing the absolutely convergent fixed-point fast sweeping WENO methods on rectangular meshes, and adopt high order WENO scheme with unequal-sized sub-stencils for spatial discretization. Extensive numerical experiments are performed to show the accuracy, computational efficiency, and absolute convergence of the presented fifth-order fast sweeping scheme on triangular meshes. Furthermore, the proposed method is compared with the forward Euler method and the popular third order total variation diminishing Rung-Kutta method for steady state computations. Numerical examples show that the developed fixed-point fast sweeping WENO method is the most efficient scheme among them, and especially it can save up to 70% CPU time costs than TVD-RK3 to converge to steady state solutions.

math.NA↗

arXiv2023

High Harmonic Generation in Solids: Particle and Wave Perspectives

High harmonic generation (HHG) from gas phase atoms (or molecules) has opened up a new frontier in ultrafast optics, where attosecond time resolution and Angstrom spatial resolution are accessible. The fundamental physical pictures of HHG are always explained by the laser-induced recollision of particle-like electron motion, which lay the foundation of attosecond spectroscopy. In recent years, HHG has also been observed in solids. One can expect the extension of attosecond spectroscopy to the condensed matter if a description capable of resolving ultrafast dynamics is provided. Thus, a large number of theoretical studies have been proposed to understand the underlying physics of HHG. Here, we revisit the recollision picture in solid HHG and show some challenges of current methods with particle perspective, and present the recently developed wave perspective Huygens-Fresnel picture in understanding dynamical systems within the ambit of strong-field physics.

physics.optics↗

arXiv2023

Improving Semiconductor Device Modeling for Electronic Design Automation by Machine Learning Techniques

The semiconductors industry benefits greatly from the integration of Machine Learning (ML)-based techniques in Technology Computer-Aided Design (TCAD) methods. The performance of ML models however relies heavily on the quality and quantity of training datasets. They can be particularly difficult to obtain in the semiconductor industry due to the complexity and expense of the device fabrication. In this paper, we propose a self-augmentation strategy for improving ML-based device modeling using variational autoencoder-based techniques. These techniques require a small number of experimental data points and does not rely on TCAD tools. To demonstrate the effectiveness of our approach, we apply it to a deep neural network-based prediction task for the Ohmic resistance value in Gallium Nitride devices. A 70% reduction in mean absolute error when predicting experimental results is achieved. The inherent flexibility of our approach allows easy adaptation to various tasks, thus making it highly relevant to many applications of the semiconductor industry.

cs.LG↗

arXiv2023

Learning to Dub Movies via Hierarchical Prosody Models

Given a piece of text, a video clip and a reference audio, the movie dubbing (also known as visual voice clone V2C) task aims to generate speeches that match the speaker's emotion presented in the video using the desired speaker voice as reference. V2C is more challenging than conventional text-to-speech tasks as it additionally requires the generated speech to exactly match the varying emotions and speaking speed presented in the video. Unlike previous works, we propose a novel movie dubbing architecture to tackle these problems via hierarchical prosody modelling, which bridges the visual information to corresponding speech prosody from three aspects: lip, face, and scene. Specifically, we align lip movement to the speech duration, and convey facial expression to speech energy and pitch via attention mechanism based on valence and arousal representations inspired by recent psychology findings. Moreover, we design an emotion booster to capture the atmosphere from global video scenes. All these embeddings together are used to generate mel-spectrogram and then convert to speech waves via existing vocoder. Extensive experimental results on the Chem and V2C benchmark datasets demonstrate the favorable performance of the proposed method. The source code and trained models will be released to the public.

cs.CL↗

arXiv2023

Recent Progress and Next Steps for the MATHUSLA LLP Detector

We report on recent progress and next steps in the design of the proposed MATHUSLA Long Lived Particle (LLP) detector for the HL-LHC as part of the Snowmass 2021 process. Our understanding of backgrounds has greatly improved, aided by detailed simulation studies, and significant R&D has been performed on designing the scintillator detectors and understanding their performance. The collaboration is on track to complete a Technical Design Report, and there are many opportunities for interested new members to contribute towards the goal of designing and constructing MATHUSLA in time for HL-LHC collisions, which would increase the sensitivity to a large variety of highly motivated LLP signals by orders of magnitude.

hep-ex↗

arXiv2023

Protective Self-Adaptive Pruning to Better Compress DNNs

Adaptive network pruning approach has recently drawn significant attention due to its excellent capability to identify the importance and redundancy of layers and filters and customize a suitable pruning solution. However, it remains unsatisfactory since current adaptive pruning methods rely mostly on an additional monitor to score layer and filter importance, and thus faces high complexity and weak interpretability. To tackle these issues, we have deeply researched the weight reconstruction process in iterative prune-train process and propose a Protective Self-Adaptive Pruning (PSAP) method. First of all, PSAP can utilize its own information, weight sparsity ratio, to adaptively adjust pruning ratio of layers before each pruning step. Moreover, we propose a protective reconstruction mechanism to prevent important filters from being pruned through supervising gradients and to avoid unrecoverable information loss as well. Our PSAP is handy and explicit because it merely depends on weights and gradients of model itself, instead of requiring an additional monitor as in early works. Experiments on ImageNet and CIFAR-10 also demonstrate its superiority to current works in both accuracy and compression ratio, especially for compressing with a high ratio or pruning from scratch.

cs.CV↗