arXiv subjects

En Li

Publications and source records attributed to En Li.

13 recordsLinked to original sources

arXiv2026

FAR-Dex: Few-shot Data Augmentation and Adaptive Residual Policy Refinement for Dexterous Manipulation

Achieving human-like dexterous manipulation through the collaboration of multi-fingered hands with robotic arms remains a longstanding challenge in robotics, primarily due to the scarcity of high-quality demonstrations and the complexity of high-dimensional action spaces. To address these challenges, we propose FAR-Dex, a hierarchical framework that integrates few-shot data augmentation with adaptive residual refinement to enable robust and precise arm-hand coordination in dexterous tasks. First, FAR-DexGen leverages the IsaacLab simulator to generate diverse and physically constrained trajectories from a few demonstrations, providing a data foundation for policy training. Second, FAR-DexRes introduces an adaptive residual module that refines policies by combining multi-step trajectory segments with observation features, thereby enhancing accuracy and robustness in manipulation scenarios. Experiments in both simulation and real-world demonstrate that FAR-Dex improves data quality by 13.4% and task success rates by 7% over state-of-the-art methods. It further achieves over 80% success in real-world tasks, enabling fine-grained dexterous manipulation with strong positional generalization.

cs.RO↗

arXiv2025

dots.llm1 Technical Report

Mixture of Experts (MoE) models have emerged as a promising paradigm for scaling language models efficiently by activating only a subset of parameters for each input token. In this report, we present dots.llm1, a large-scale MoE model that activates 14B parameters out of a total of 142B parameters, delivering performance on par with state-of-the-art models while reducing training and inference costs. Leveraging our meticulously crafted and efficient data processing pipeline, dots.llm1 achieves performance comparable to Qwen2.5-72B after pretraining on 11.2T high-quality tokens and post-training to fully unlock its capabilities. Notably, no synthetic data is used during pretraining. To foster further research, we open-source intermediate training checkpoints at every one trillion tokens, providing valuable insights into the learning dynamics of large language models.

cs.CL↗

arXiv2025

Designer polyradical nanographenes with strong spin entanglement and perturbation resilience via Clar's goblet extension

Polyradical nanographenes featuring strong spin entanglement and robust many-body spin states against external magnetic perturbations not only enable the exploration of correlated quantum magnetism at the molecular scale, but also constitute promising candidates for developing molecular qubits with chemical tunability and building scalable quantum networks. Here, we employed a predictive design strategy to achieve the on-surface synthesis of two homologues of Clar goblet, C62H22 and C76H26, via lateral and vertical extensions of the parent structure, respectively. Vertical extension increases the number of topologically frustrated zero-energy modes, which scale linearly with the total number of benzene ring rows. In contrast, the lateral extension enhances electron-electron interactions, leading to the emergence of additional radical states beyond those predicted by the topological zero-energy modes. Consequently, both structures exhibit correlated tetraradical character and a many-body singlet ground state as confirmed by multireference theoretical calculations. These magnetic states arise from unique magnetic origins and also display distinct resilience to external perturbations, which can be experimentally validated using nickelocene-functionalized scanning probe techniques. Our work presents a general strategy for rational design of highly entangled polyradical nanographenes with tunable spin numbers and resilience of their many-body spin states to perturbations, opening exciting possibilities for exploring novel correlated spin phases in molecular systems and advancing quantum information technologies.

cond-mat.mes-hall↗

arXiv2025

Teacher Motion Priors: Enhancing Robot Locomotion over Challenging Terrain

Achieving robust locomotion on complex terrains remains a challenge due to high dimensional control and environmental uncertainties. This paper introduces a teacher prior framework based on the teacher student paradigm, integrating imitation and auxiliary task learning to improve learning efficiency and generalization. Unlike traditional paradigms that strongly rely on encoder-based state embeddings, our framework decouples the network design, simplifying the policy network and deployment. A high performance teacher policy is first trained using privileged information to acquire generalizable motion skills. The teacher's motion distribution is transferred to the student policy, which relies only on noisy proprioceptive data, via a generative adversarial mechanism to mitigate performance degradation caused by distributional shifts. Additionally, auxiliary task learning enhances the student policy's feature representation, speeding up convergence and improving adaptability to varying terrains. The framework is validated on a humanoid robot, showing a great improvement in locomotion stability on dynamic terrains and significant reductions in development costs. This work provides a practical solution for deploying robust locomotion strategies in humanoid robots.

cs.RO↗

arXiv2024

Artificially built Kondo chains with organic radicals on metallic surfaces: new model system of heavy fermion quantum criticality

Heavy fermion quantum criticality is an extremely rich domain of research which represents a framework to understand strange metals as a consequence of a Kondo breakdown transition. Here we provide an experimental realization of such systems in terms of organic radicals on a metallic surface. The ground state of organic radicals is a Kramer's doublet that can be modeled by a spin 1/2 degree of freedom. Using on-surface synthesis and scanning tunneling microscopy (STM) tip manipulation, one can controllably engineer and characterize chains of organic radicals on a Au(111) surface. The spatial-resolved differential conductance reveals site-dependent low-energy excitations, which support the picture of emergent many-body Kondo physics. Using quantum Monte Carlo simulations, we show that a Kondo lattice model of spin chains on a metallic surface reproduces accurately the experimental results. This allows us to interpret the experimental results in terms of a heavy fermion metal, below the coherence temperature. We foresee that the tunability of these systems will pave the way to realize quantum simulators of heavy fermion criticality.

cond-mat.mes-hall↗

arXiv2022

Robust and accurate depth estimation by fusing LiDAR and Stereo

Depth estimation is one of the key technologies in some fields such as autonomous driving and robot navigation. However, the traditional method of using a single sensor is inevitably limited by the performance of the sensor. Therefore, a precision and robust method for fusing the LiDAR and stereo cameras is proposed. This method fully combines the advantages of the LiDAR and stereo camera, which can retain the advantages of the high precision of the LiDAR and the high resolution of images respectively. Compared with the traditional stereo matching method, the texture of the object and lighting conditions have less influence on the algorithm. Firstly, the depth of the LiDAR data is converted to the disparity of the stereo camera. Because the density of the LiDAR data is relatively sparse on the y-axis, the converted disparity map is up-sampled using the interpolation method. Secondly, in order to make full use of the precise disparity map, the disparity map and stereo matching are fused to propagate the accurate disparity. Finally, the disparity map is converted to the depth map. Moreover, the converted disparity map can also increase the speed of the algorithm. We evaluate the proposed pipeline on the KITTI benchmark. The experiment demonstrates that our algorithm has higher accuracy than several classic methods.

cs.CV↗

arXiv2022

Magnet-free nonreciprocal metasurface for on-demand bi-directional phase modulation

Unconstrained by Lorentz reciprocity, nonreciprocal metasurfaces are uniquely capable of encoding distinctive optical functions on forward- and backward-propagating waves. The nonreciprocal metasurfaces reported to date require external electric or magnetic field biasing or rely on nonlinear effects, both of which are challenging to practically implement. Here, we propose and experimentally realize a magnet-free, linear, and passive nonreciprocal metasurface based on self-biased magnetic meta-atoms. Record transmittance up to 77% and operation angle reaching 64 degree are experimentally demonstrated. Moreover, on-demand bidirectional phase modulation in a "LEGO-like" manner is theoretically proposed and experimentally demonstrated, enabling a cohort of nonreciprocal functionalities such as microwave isolation, nonreciprocal beam steering, nonreciprocal focusing, and nonreciprocal holography. The design can also be extended to MHz and optical frequencies, taking advantage of the wide variety of self-biased gyrotropic materials available. We foresee that the nonreciprocal metasurfaces demonstrated in this work will have a significant practical impact for applications ranging from nonreciprocal antennas and radomes to full-duplex wireless communication and radar systems.

physics.optics↗

arXiv2021

Lattice reconstruction induced multiple ultra-flat bands in twisted bilayer WSe2

Moir\'e superlattices in van der Waals heterostructures provide a tunable platform to study emergent properties that are absent in the natural crystal form. Twisted bilayer transition metal dichalcogenides (TB-TMDs) can host moir\'e flat bands over a wide range of twist angles. For twist angle close to 60{\deg}, it was predicted that TB-TMDs undergo a lattice reconstruction which causes the formation of ultra-flat bands. Here, by using scanning tunneling microscopy and spectroscopy, we show the emergence of multiple ultra-flat bands in twisted bilayer WSe2 when the twist angle is larger than 57{\deg}. The bandwidth, manifested as narrow tunneling conductance peaks, is estimated less than 10meV, which is only a fraction of the estimated on-site Coulomb repulsion energy. The number of these ultra-flat bands and spatial distribution of the wavefunctions match the theoretical predictions incredibly well, strongly evidencing that the observed ultra-flat bands are induced by lattice reconstruction. Our work provides a foundation for further study of the exotic correlated phases in TB-TMDs.

cond-mat.mes-hall↗

arXiv2020

Auxiliary Diagnosing Coronary Stenosis Using Machine Learning

How to accurately classify and diagnose whether an individual has Coronary Stenosis (CS) without invasive physical examination? This problem has not been solved satisfactorily. To this end, the four machine learning (ML) algorithms, i.e., Boosted Tree (BT), Decision Tree (DT), Logistic Regression (LR) and Random Forest (RF) are employed in this paper. First, eleven features including basic information of an individual, symptoms and results of routine physical examination are selected, as well as one label is specified, indicating whether an individual suffers from different severity of coronary artery stenosis or not. On the basis of it, a sample set is constructed. Second, each of these four ML algorithms learns from the sample set to obtain the corresponding optimal classified results, respectively. The experimental results show that: RF performs better than other three algorithms, and the former algorithm classifies whether an individual has CS with an accuracy of 95.7% (=90/94).

q-bio.TO↗

arXiv2020

Giant nonlinear Hall effect in twisted WSe$_2$

The recently discovered nonlinear Hall effect (NHE) in a few non-interacting systems provides a novel mechanism to generate second harmonic electrical Hall signals under time-reversal-symmetric conditions. Here, we introduce a new approach to engineering NHE by using twisted moir\'e structures. We find that the twisted WSe$_2$ bilayer exhibits a NHE when tuning the Fermi level to the moir\'e flat bands. Near half-filling of the first moir\'e band, the nonlinear Hall signal shows a sharp peak with the generation efficiency at least two orders of magnitude larger than those in previous experiments. We propose that the giant NHE and diverging generation efficiency originate from a mass-diverging type continuous Mott transition, which is evidenced by resistivity measurements. This work demonstrates not only how interaction effects can couple to Berry curvature dipoles to produce novel quantum phenomena, but also what NHE measurements can provide for developing a new tool to study the quantum criticality.

cond-mat.mes-hall↗

arXiv2019

Edge AI: On-Demand Accelerating Deep Neural Network Inference via Edge Computing

As a key technology of enabling Artificial Intelligence (AI) applications in 5G era, Deep Neural Networks (DNNs) have quickly attracted widespread attention. However, it is challenging to run computation-intensive DNN-based tasks on mobile devices due to the limited computation resources. What's worse, traditional cloud-assisted DNN inference is heavily hindered by the significant wide-area network latency, leading to poor real-time performance as well as low quality of user experience. To address these challenges, in this paper, we propose Edgent, a framework that leverages edge computing for DNN collaborative inference through device-edge synergy. Edgent exploits two design knobs: (1) DNN partitioning that adaptively partitions computation between device and edge for purpose of coordinating the powerful cloud resource and the proximal edge resource for real-time DNN inference; (2) DNN right-sizing that further reduces computing latency via early exiting inference at an appropriate intermediate DNN layer. In addition, considering the potential network fluctuation in real-world deployment, Edgentis properly design to specialize for both static and dynamic network environment. Specifically, in a static environment where the bandwidth changes slowly, Edgent derives the best configurations with the assist of regression-based prediction models, while in a dynamic environment where the bandwidth varies dramatically, Edgent generates the best execution plan through the online change point detection algorithm that maps the current bandwidth state to the optimal configuration. We implement Edgent prototype based on the Raspberry Pi and the desktop PC and the extensive experimental evaluations demonstrate Edgent's effectiveness in enabling on-demand low-latency edge intelligence.

cs.NI↗

arXiv2019

Edge Intelligence: Paving the Last Mile of Artificial Intelligence with Edge Computing

With the breakthroughs in deep learning, the recent years have witnessed a booming of artificial intelligence (AI) applications and services, spanning from personal assistant to recommendation systems to video/audio surveillance. More recently, with the proliferation of mobile computing and Internet-of-Things (IoT), billions of mobile and IoT devices are connected to the Internet, generating zillions Bytes of data at the network edge. Driving by this trend, there is an urgent need to push the AI frontiers to the network edge so as to fully unleash the potential of the edge big data. To meet this demand, edge computing, an emerging paradigm that pushes computing tasks and services from the network core to the network edge, has been widely recognized as a promising solution. The resulted new inter-discipline, edge AI or edge intelligence, is beginning to receive a tremendous amount of interest. However, research on edge intelligence is still in its infancy stage, and a dedicated venue for exchanging the recent advances of edge intelligence is highly desired by both the computer system and artificial intelligence communities. To this end, we conduct a comprehensive survey of the recent research efforts on edge intelligence. Specifically, we first review the background and motivation for artificial intelligence running at the network edge. We then provide an overview of the overarching architectures, frameworks and emerging key technologies for deep learning model towards training/inference at the network edge. Finally, we discuss future research opportunities on edge intelligence. We believe that this survey will elicit escalating attentions, stimulate fruitful discussions and inspire further research ideas on edge intelligence.

cs.DC↗

arXiv2018

Edge Intelligence: On-Demand Deep Learning Model Co-Inference with Device-Edge Synergy

As the backbone technology of machine learning, deep neural networks (DNNs) have have quickly ascended to the spotlight. Running DNNs on resource-constrained mobile devices is, however, by no means trivial, since it incurs high performance and energy overhead. While offloading DNNs to the cloud for execution suffers unpredictable performance, due to the uncontrolled long wide-area network latency. To address these challenges, in this paper, we propose Edgent, a collaborative and on-demand DNN co-inference framework with device-edge synergy. Edgent pursues two design knobs: (1) DNN partitioning that adaptively partitions DNN computation between device and edge, in order to leverage hybrid computation resources in proximity for real-time DNN inference. (2) DNN right-sizing that accelerates DNN inference through early-exit at a proper intermediate DNN layer to further reduce the computation latency. The prototype implementation and extensive evaluations based on Raspberry Pi demonstrate Edgent's effectiveness in enabling on-demand low-latency edge intelligence.

cs.DC↗