Searcharxiv⌕ Search

arXiv subjects

Li Chen

Publications and source records attributed to Li Chen.

At least 433 records · Page 24Linked to original sources
arXiv2022

Level 2 Autonomous Driving on a Single Device: Diving into the Devils of Openpilot

Equipped with a wide span of sensors, predominant autonomous driving solutions are becoming more modular-oriented for safe system design. Though these sensors have laid a solid foundation, most massive-production solutions up to date still fall into L2 phase. Among these, Comma.ai comes to our sight, claiming one $999 aftermarket device mounted with a single camera and board inside owns the ability to handle L2 scenarios. Together with open-sourced software of the entire system released by Comma.ai, the project is named Openpilot. Is it possible? If so, how is it made possible? With curiosity in mind, we deep-dive into Openpilot and conclude that its key to success is the end-to-end system design instead of a conventional modular framework. The model is briefed as Supercombo, and it can predict the ego vehicle's future trajectory and other road semantics on the fly from monocular input. Unfortunately, the training process and massive amount of data to make all these work are not publicly available. To achieve an intensive investigation, we try to reimplement the training details and test the pipeline on public benchmarks. The refactored network proposed in this work is referred to as OP-Deepdive. For a fair comparison of our version to the original Supercombo, we introduce a dual-model deployment scheme to test the driving performance in the real world. Experimental results on nuScenes, Comma2k19, CARLA, and in-house realistic scenarios verify that a low-cost device can indeed achieve most L2 functionalities and be on par with the original Supercombo model. In this report, we would like to share our latest findings, shed some light on the new perspective of end-to-end autonomous driving from an industrial product-level side, and potentially inspire the community to continue improving the performance. Our code, benchmarks are at https://github.com/OpenPerceptionX/Openpilot-Deepdive.

cs.CV↗
arXiv2022

New insights into the structure of open clusters in the Gaia era

With the help of Gaia data, it is noted that in addition to the core components, there are low-density outer halo components in the extended region of open clusters. To study the extended structure beyond the core radius of the cluster ($\sim$ 10 pc), based on Gaia EDR3 data, taking up to 50 pc as the searching radius, we use the pyUPMASK algorithm to re-determine the member stars of the open cluster within 1-2 kpc. We obtain the member stars of 256 open clusters, especially those located in the outer halo region of open clusters. Furthermore, we find that most open clusters' radial density profile in the outer region deviates from the King's profile. To better describe the internal and external structural characteristics of open clusters, we propose a double components model for description: core components with King model distribution and outer halo components with logarithmic Gaussian distribution, and then suggest using four radii ( $r_c$, $r_t$, $r_o$, $r_e$) for describing the structure and distribution profile of star clusters, where $r_t$ and $r_e$ represent the boundaries of core components and outer halo components respectively. Finally, we provide a catalog of 256 clusters with structural parameters. In addition, our study shows the sizes of these radii are statistically linear related, which indicates that the inner and outer regions of the cluster are interrelated and follow similar evolutionary processes. Further, we show that the structure of two components can be used to better trace the cluster evolution properties in different stages.

astro-ph.GA↗
arXiv2022

An Efficient Cervical Whole Slide Image Analysis Framework Based on Multi-scale Semantic and Location Deep Features

Digital gigapixel whole slide image (WSI) is widely used in clinical diagnosis, and automated WSI analysis is key for computer-aided diagnosis. Currently, analyzing the integrated descriptor of probabilities or feature maps from massive local patches encoded by ResNet classifier is the main manner for WSI-level prediction. Feature representations of the sparse and tiny lesion cells in cervical slides, however, are still challenging, while the unused location representations are available to supply the semantics classification. This study designs a novel and efficient framework with a new module InCNet constructed lightweight model YOLCO (You Only Look Cytology Once). It directly extracts feature inside the single cell (cluster) instead of the traditional way that from image tile with a fixed size. The InCNet (Inline Connection Network) enriches the multi-scale connectivity without efficiency loss. The proposal allows the input size enlarged to megapixel that can stitch the WSI by the average repeats decreased from $10^3\sim10^4$ to $10^1\sim10^2$ for collecting features and predictions at two scales. Based on Transformer for classifying the integrated multi-scale multi-task WSI features, the experimental results appear $0.872$ AUC score better than the best conventional model on our dataset ($n$=2,019) from four scanners. The code is available at https://github.com/Chrisa142857/You-Only-Look-Cytopathology-Once , where the deployment version has the speed $\sim$70 s/WSI.

cs.CV↗
arXiv2022

Deep Learning with Label Noise: A Hierarchical Approach

Deep neural networks are susceptible to label noise. Existing methods to improve robustness, such as meta-learning and regularization, usually require significant change to the network architecture or careful tuning of the optimization procedure. In this work, we propose a simple hierarchical approach that incorporates a label hierarchy when training the deep learning models. Our approach requires no change of the network architecture or the optimization procedure. We investigate our hierarchical network through a wide range of simulated and real datasets and various label noise types. Our hierarchical approach improves upon regular deep neural networks in learning with label noise. Combining our hierarchical approach with pre-trained models achieves state-of-the-art performance in real-world noisy datasets.

cs.LG↗
arXiv2022

Combined mean-field and semiclassical limits of large fermionic systems

We study the time dependent Schrödinger equation for large spinless fermions with the semiclassical scale $\hbar = N^{-1/3}$ in three dimensions. By using the Husimi measure defined by coherent states, we rewrite the Schrödinger equation into a BBGKY type of hierarchy for the k particle Husimi measure. Further estimates are derived to obtain the weak compactness of the Husimi measure, and in addition uniform estimates for the remainder terms in the hierarchy are derived in order to show that in the semiclassical regime the weak limit of the Husimi measure is exactly the solution of the Vlasov equation.

math-ph↗
arXiv2022

Discovery of extended structure around open cluster COIN-Gaia 13 based on Gaia EDR3

COIN-Gaia 13 is a newly discovered open cluster revealed by Gaia DR2 data. It is a nearby open cluster with a distance of about 513 pc. Combined with the five-dimensional astrometric data of Gaia EDR3 with higher accuracy, we use the membership assignment algorithm (pyUPMASK) to determine the membership of COIN-Gaia 13 in a large extended spatial region. The cluster has found 478 candidate members. After obtaining reliable cluster members, we further study its basic properties and spatial distribution. Our results show that there is an obvious extended structure of the cluster in the X-Y plane. This elongated structure is distributed along the spiral arm, and the whole length is about 270 pc. The cluster age is 250 Myr, the total mass is about 439 M$_\odot$, and the tidal radius of the cluster is about 11 pc. Since more than half of the member stars (352 stars) are located outside twice the tidal radius, it is suspected that this cluster is undergoing the dynamic dissolution process. Furthermore, the spatial distribution and kinematic analysis indicate that the extended structure in COIN-Gaia 13 is more likely to be caused by the differential rotation of the Galaxy.

astro-ph.GA↗
arXiv2022

Maximum Flow and Minimum-Cost Flow in Almost-Linear Time

We give an algorithm that computes exact maximum flows and minimum-cost flows on directed graphs with $m$ edges and polynomially bounded integral demands, costs, and capacities in $m^{1+o(1)}$ time. Our algorithm builds the flow through a sequence of $m^{1+o(1)}$ approximate undirected minimum-ratio cycles, each of which is computed and processed in amortized $m^{o(1)}$ time using a new dynamic graph data structure. Our framework extends to algorithms running in $m^{1+o(1)}$ time for computing flows that minimize general edge-separable convex functions to high accuracy. This gives almost-linear time algorithms for several problems including entropy-regularized optimal transport, matrix scaling, $p$-norm flows, and $p$-norm isotonic regression on arbitrary directed acyclic graphs.

cs.DS↗
arXiv2022

Quasi-periodic oscillations of the X-ray burst from the magnetar SGR J1935+2154 and associated with the fast radio burst FRB 200428

The origin(s) and mechanism(s) of fast radio bursts (FRBs), which are short radio pulses from cosmological distances, have remained a major puzzle since their discovery. We report a strong Quasi-Periodic Oscillation(QPO) of 40 Hz in the X-ray burst from the magnetar SGR J1935+2154 and associated with FRB 200428, significantly detected with the Hard X-ray Modulation Telescope (Insight-HXMT) and also hinted by the Konus-Wind data. QPOs from magnetar bursts have only been rarely detected; our 3.4 sigma (p-value is 2.9e-4) detection of the QPO reported here reveals the strongest QPO signal observed from magnetars (except in some very rare giant flares), making this X-ray burst unique among magnetar bursts. The two X-ray spikes coinciding with the two FRB pulses are also among the peaks of the QPO. Our results suggest that at least some FRBs are related to strong oscillation processes of neutron stars. We also show that we may overestimate the significance of the QPO signal and underestimate the errors of QPO parameters if QPO exists only in a fraction of the time series of a X-ray burst which we use to calculate the Leahy-normalized periodogram.

astro-ph.HE↗
arXiv2022

Impacts of Personal Characteristics on User Trust in Conversational Recommender Systems

Conversational recommender systems (CRSs) imitate human advisors to assist users in finding items through conversations and have recently gained increasing attention in domains such as media and e-commerce. Like in human communication, building trust in human-agent communication is essential given its significant influence on user behavior. However, inspiring user trust in CRSs with a "one-size-fits-all" design is difficult, as individual users may have their own expectations for conversational interactions (e.g., who, user or system, takes the initiative), which are potentially related to their personal characteristics. In this study, we investigated the impacts of three personal characteristics, namely personality traits, trust propensity, and domain knowledge, on user trust in two types of text-based CRSs, i.e., user-initiative and mixed-initiative. Our between-subjects user study (N=148) revealed that users' trust propensity and domain knowledge positively influenced their trust in CRSs, and that users with high conscientiousness tended to trust the mixed-initiative system.

cs.HC↗
arXiv2022

ERNIE-SPARSE: Learning Hierarchical Efficient Transformer Through Regularized Self-Attention

Sparse Transformer has recently attracted a lot of attention since the ability for reducing the quadratic dependency on the sequence length. We argue that two factors, information bottleneck sensitivity and inconsistency between different attention topologies, could affect the performance of the Sparse Transformer. This paper proposes a well-designed model named ERNIE-Sparse. It consists of two distinctive parts: (i) Hierarchical Sparse Transformer (HST) to sequentially unify local and global information. (ii) Self-Attention Regularization (SAR) method, a novel regularization designed to minimize the distance for transformers with different attention topologies. To evaluate the effectiveness of ERNIE-Sparse, we perform extensive evaluations. Firstly, we perform experiments on a multi-modal long sequence modeling task benchmark, Long Range Arena (LRA). Experimental results demonstrate that ERNIE-Sparse significantly outperforms a variety of strong baseline methods including the dense attention and other efficient sparse attention methods and achieves improvements by 2.77% (57.78% vs. 55.01%). Secondly, to further show the effectiveness of our method, we pretrain ERNIE-Sparse and verified it on 3 text classification and 2 QA downstream tasks, achieve improvements on classification benchmark by 0.83% (92.46% vs. 91.63%), on QA benchmark by 3.24% (74.67% vs. 71.43%). Experimental results continue to demonstrate its superior performance.

cs.CL↗
arXiv2022

Conversational Recommendation: A Grand AI Challenge

Animated avatars, which look and talk like humans, are iconic visions of the future of AI-powered systems. Through many sci-fi movies we are acquainted with the idea of speaking to such virtual personalities as if they were humans. Today, we talk more and more to machines like Apple's Siri, e.g., to ask them for the weather forecast. However, when asked for recommendations, e.g., for a restaurant to go to, the limitations of such devices quickly become obvious. They do not engage in a conversation to find out what we might prefer, they often do not provide explanations for what they recommend, and they may have difficulties remembering what was said one minute earlier. Conversational recommender systems promise to address these limitations. In this paper, we review existing approaches to build such systems, which developments we observe today, which challenges are still open and why the development of conversational recommenders represents one of the next grand challenges of AI.

cs.AI↗
arXiv2022

A Survey of Surface Defect Detection of Industrial Products Based on A Small Number of Labeled Data

The surface defect detection method based on visual perception has been widely used in industrial quality inspection. Because defect data are not easy to obtain and the annotation of a large number of defect data will waste a lot of manpower and material resources. Therefore, this paper reviews the methods of surface defect detection of industrial products based on a small number of labeled data, and this method is divided into traditional image processing-based industrial product surface defect detection methods and deep learning-based industrial product surface defect detection methods suitable for a small number of labeled data. The traditional image processing-based industrial product surface defect detection methods are divided into statistical methods, spectral methods and model methods. Deep learning-based industrial product surface defect detection methods suitable for a small number of labeled data are divided into based on data augmentation, based on transfer learning, model-based fine-tuning, semi-supervised, weak supervised and unsupervised.

cs.CV↗
arXiv2022

New Open Cluster candidates Found in Galactic Disk Using Gaia DR2/EDR3 Data

We report 541 new open cluster candidates in Gaia EDR3 through revisiting the cluster results from an earlier analysis of the Gaia DR2, which revealed nearly a thousand open cluster candidates in the solar neighborhood (mostly d < 3 kpc) resideing at Galactic latitudes |b| < 20 degrees. A subsequent comparison with lists of known clusters shows a large increases of the cluster samples within 2 kpc from the Sun. We assign membership probabilities to the stars through the open source pyUPMASK algorithm, and also estimate the physical parameters through isochrone fitting for each candidate. Most of the new candidates show small total proper motion dispersions and clear features in the color-magnitude diagrams. Besides, the metallicity gradient of the new candidates is consistent with those found in the literature. The cluster parameters and member stars are available at CDS via anonymous ftp to cdsarc.u-strasbg.fr(130.79.128.5) or via https://cdsarc.unistra.fr/viz-bin/cat/J/ApJS. The discovery of these new objects shows that the open cluster samples in Gaia data is still not complete, and more discoveries are expected in the future researches.

astro-ph.GA↗
arXiv2022

Finding Optimal Tangent Points for Reducing Distortions of Hard-label Attacks

One major problem in black-box adversarial attacks is the high query complexity in the hard-label attack setting, where only the top-1 predicted label is available. In this paper, we propose a novel geometric-based approach called Tangent Attack (TA), which identifies an optimal tangent point of a virtual hemisphere located on the decision boundary to reduce the distortion of the attack. Assuming the decision boundary is locally flat, we theoretically prove that the minimum $\ell_2$ distortion can be obtained by reaching the decision boundary along the tangent line passing through such tangent point in each iteration. To improve the robustness of our method, we further propose a generalized method which replaces the hemisphere with a semi-ellipsoid to adapt to curved decision boundaries. Our approach is free of pre-training. Extensive experiments conducted on the ImageNet and CIFAR-10 datasets demonstrate that our approach can consume only a small number of queries to achieve the low-magnitude distortion. The implementation source code is released online at https://github.com/machanic/TangentAttack.

cs.CV↗
arXiv2022

Meta-learning an Intermediate Representation for Few-shot Block-wise Prediction of Landslide Susceptibility

Predicting a landslide susceptibility map (LSM) is essential for risk recognition and disaster prevention. Despite the successful application of data-driven approaches for LSM prediction, most methods generally apply a single global model to predict the LSM for an entire target region. However, in large-scale areas with significant environmental change, various parts of the region hold different landslide-inducing environments, and therefore, should be predicted with respective models. This study first segmented target scenarios into blocks for individual analysis. Then, the critical problem is that in each block with limited samples, conducting training and testing a model is impossible for a satisfactory LSM prediction, especially in dangerous mountainous areas where landslide surveying is expensive. To solve the problem, we trained an intermediate representation by the meta-learning paradigm, which is superior for capturing information valuable for few-shot adaption from LSM tasks. We hypothesized that there are more general and vital concepts concerning landslide causes and are sensitive to variations in input features. Thus, we can quickly few-shot adapt the models from the intermediate representation for different blocks or even unseen tasks using very few exemplar samples. Experimental results on the two study areas demonstrated the validity of our block-wise analysis in large scenarios and revealed the top few-shot adaption performances of the proposed methods.

cs.LG↗
arXiv2022

Probabilistic fair behaviors spark its boost in the Ultimatum Game: the strength of good Samaritans

Behavioral experiments on the Ultimatum Game have shown that we human beings have remarkable preference in fair play, contradicting the predictions by the game theory. Most of the existing models seeking for explanations, however, strictly follow the assumption of \emph{Homo economicus} in orthodox Economics that people are self-interested and fully rational to maximize their earnings. Here we relax this assumption by allowing that people probabilistically choose to be "good Samaritans", acting as fair players from time to time. For well-mixed and homogeneously structured populations, we numerically show that as this probability increases the level of fairness undergoes from the low scenario abruptly to the full fairness state, where occasional fair behaviors ($\sim5\%$) are sufficient to drive the whole population to behave in the half-half split manner. We also develop a mean-field theory, which correctly reproduces the first-order phase transition and points out that the bistability is an intrinsic property of this game and small fair acts lead to dramatical change due to its bifurcation structure. Heterogeneously structured populations, however, display continuous fairness transition; surprisingly, very few hub nodes acting as fair players are able to entrain the whole population to the full fairness state. Our results thus reveal the unexpected strength of "good Samaritans", which may constitute a new explanation for the emergence of fairness in our society.

q-bio.PE↗
arXiv2022

Decentralized Federated Learning: Balancing Communication and Computing Costs

Decentralized stochastic gradient descent (SGD) is a driving engine for decentralized federated learning (DFL). The performance of decentralized SGD is jointly influenced by inter-node communications and local updates. In this paper, we propose a general DFL framework, which implements both multiple local updates and multiple inter-node communications periodically, to strike a balance between communication efficiency and model consensus. It can provide a general decentralized SGD analytical framework. We establish strong convergence guarantees for the proposed DFL algorithm without the assumption of convex objectives. The convergence rate of DFL can be optimized to achieve the balance of communication and computing costs under constrained resources. For improving communication efficiency of DFL, compressed communication is further introduced to the proposed DFL as a new scheme, named DFL with compressed communication (C-DFL). The proposed C-DFL exhibits linear convergence for strongly convex objectives. Experiment results based on MNIST and CIFAR-10 datasets illustrate the superiority of DFL over traditional decentralized SGD methods and show that C-DFL further enhances communication efficiency.

cs.LG↗