arXiv subjects

Lei Han

Publications and source records attributed to Lei Han.

At least 73 records · Page 4Linked to original sources

arXiv2024

VLN-Game: Vision-Language Equilibrium Search for Zero-Shot Semantic Navigation

Following human instructions to explore and search for a specified target in an unfamiliar environment is a crucial skill for mobile service robots. Most of the previous works on object goal navigation have typically focused on a single input modality as the target, which may lead to limited consideration of language descriptions containing detailed attributes and spatial relationships. To address this limitation, we propose VLN-Game, a novel zero-shot framework for visual target navigation that can process object names and descriptive language targets effectively. To be more precise, our approach constructs a 3D object-centric spatial map by integrating pre-trained visual-language features with a 3D reconstruction of the physical environment. Then, the framework identifies the most promising areas to explore in search of potential target candidates. A game-theoretic vision language model is employed to determine which target best matches the given language description. Experiments conducted on the Habitat-Matterport 3D (HM3D) dataset demonstrate that the proposed framework achieves state-of-the-art performance in both object goal navigation and language-based navigation tasks. Moreover, we show that VLN-Game can be easily deployed on real-world robots. The success of VLN-Game highlights the promising potential of using game-theoretic methods with compact vision-language models to advance decision-making capabilities in robotic systems. The supplementary video and code can be accessed via the following link: https://sites.google.com/view/vln-game.

cs.RO↗

arXiv2024

HumanVLA: Towards Vision-Language Directed Object Rearrangement by Physical Humanoid

Physical Human-Scene Interaction (HSI) plays a crucial role in numerous applications. However, existing HSI techniques are limited to specific object dynamics and privileged information, which prevents the development of more comprehensive applications. To address this limitation, we introduce HumanVLA for general object rearrangement directed by practical vision and language. A teacher-student framework is utilized to develop HumanVLA. A state-based teacher policy is trained first using goal-conditioned reinforcement learning and adversarial motion prior. Then, it is distilled into a vision-language-action model via behavior cloning. We propose several key insights to facilitate the large-scale learning process. To support general object rearrangement by physical humanoid, we introduce a novel Human-in-the-Room dataset encompassing various rearrangement tasks. Through extensive experiments and analysis, we demonstrate the effectiveness of the proposed approach.

cs.RO↗

arXiv2024

Memory Remedy: An AI-Enhanced Interactive Story Exploring Human-Robot Interaction and Companionship

We present our approach to using AI-generated content (AIGC) and multiple media to develop an immersive, game-based, interactive story experience. The narrative of the story, "Memory Remedy", unfolds through flashbacks, allowing the audience to gradually uncover the story and the complex relationship between the robot protagonist and the older adults. This exploration explores important themes such as the journey of life, the profound influence of memories, and the concept of post-human emotional care. By engaging with this AIGC-based interactive story, audiences are encouraged to reflect on the potential role of robotic companionship in the lives of older adults in the future; and to encourage deeper reflection on the complex relationship between artificial intelligence and humanity.

cs.HC↗

arXiv2024

Lead-free Hybrid Perovskite: An Efficient Room Temperature Spin Generator via Large Interfacial Rashba effect

Two-dimensional (2D) hybrid organic-inorganic perovskite (HOIP) demonstates great potential for developing flexible and wearable spintronic devices, by serving as spin sources via the bulk Rashba effect (BRE). However, the practical application of BRE in 2D HOIP faces huge challenges, particularly due to the toxicity of lead, which is crucial for achieving large spin-orbit coupling, and the restrictions in 2D HOIP candidates to meet specific symmetry-breaking requirements. To overcome these obstacles, we design a strategy to exploit the interfacial Rashba effect (IRE) of lead-free 2D HOIP (C6H5CH2CH2NH3)2CuCl4 (PEA-CuCl), manifesting as an efficient spin generator at room temperature. IRE of PEA-CuCl originates from the large orbital hybridization at the interface between PEA-CuCl and adjacent ferromagnetic layers. Spin-torque ferromagnetic resonance measurements further quantify a large Rashba effective field of 14.04 Oe per 10^11 A m-2, surpassing those of lead-based HOIP and traditional all-inorganic heterojunctions with noble metals. Our lead-free 2D HOIP PEA-CuCl, which harnesses large IRE for spin generation, is efficient, nontoxic, and economic, offering huge promise for future flexible and wearable spintronic devices.

cond-mat.mtrl-sci↗

arXiv2024

Tunable Quantum Anomalous Hall Effect via Crystal Order in Spin-Splitting Antiferromagnets

Quantum anomalous Hall (QAH) effect provides dissipationless chiral channels for spin transport, expected as an outstanding candidate in future low-power quantum computation. The spin-splitting band structure is vital for obtaining QAH effect in topological systems, with ferromagnetism indispensable to manipulate the Chern number. Herein, we challenge this wisdom by proposing tunable QAH effect in spin-splitting antiferromagnets with zero magnetization. Since the spin splitting of these unique magnets originates from the alternate crystal environment, the Chern number can be modulated not only by the conventional magnetic order, but also by the crystal order, opening an additional dimension for tuning QAH effect. Our concept is illustrated based on two-dimensional (2D) MnBi2Te4 (MBT) with even septuple layers (SLs), a typical axion insulator with fully magnetic compensation. By interlayer rotation and translation operations, sublattices of MBT with opposite magnetizations are no longer connected by inversion or mirror symmetries, leading to the transition to the QAH insulator. The flexible stacking of 2D materials enables the reversible Chern number by crystal design. Our work fundamentally reveals the crystal-order-dependent QAH effect in spin-splitting antiferromagnets, which would advance QAH effect-based devices towards high controllability, integration density and operation speed.

cond-mat.mes-hall↗

arXiv2024

VidEgoThink: Assessing Egocentric Video Understanding Capabilities for Embodied AI

Recent advancements in Multi-modal Large Language Models (MLLMs) have opened new avenues for applications in Embodied AI. Building on previous work, EgoThink, we introduce VidEgoThink, a comprehensive benchmark for evaluating egocentric video understanding capabilities. To bridge the gap between MLLMs and low-level control in Embodied AI, we design four key interrelated tasks: video question-answering, hierarchy planning, visual grounding and reward modeling. To minimize manual annotation costs, we develop an automatic data generation pipeline based on the Ego4D dataset, leveraging the prior knowledge and multimodal capabilities of GPT-4o. Three human annotators then filter the generated data to ensure diversity and quality, resulting in the VidEgoThink benchmark. We conduct extensive experiments with three types of models: API-based MLLMs, open-source image-based MLLMs, and open-source video-based MLLMs. Experimental results indicate that all MLLMs, including GPT-4o, perform poorly across all tasks related to egocentric video understanding. These findings suggest that foundation models still require significant advancements to be effectively applied to first-person scenarios in Embodied AI. In conclusion, VidEgoThink reflects a research trend towards employing MLLMs for egocentric vision, akin to human capabilities, enabling active observation and interaction in the complex real-world environments.

cs.CV↗

arXiv2024

Fairness without Sensitive Attributes via Knowledge Sharing

While model fairness improvement has been explored previously, existing methods invariably rely on adjusting explicit sensitive attribute values in order to improve model fairness in downstream tasks. However, we observe a trend in which sensitive demographic information becomes inaccessible as public concerns around data privacy grow. In this paper, we propose a confidence-based hierarchical classifier structure called "Reckoner" for reliable fair model learning under the assumption of missing sensitive attributes. We first present results showing that if the dataset contains biased labels or other hidden biases, classifiers significantly increase the bias gap across different demographic groups in the subset with higher prediction confidence. Inspired by these findings, we devised a dual-model system in which a version of the model initialised with a high-confidence data subset learns from a version of the model initialised with a low-confidence data subset, enabling it to avoid biased predictions. Our experimental results show that Reckoner consistently outperforms state-of-the-art baselines in COMPAS dataset and New Adult dataset, considering both accuracy and fairness metrics.

cs.LG↗

arXiv2024

Navigating Munk's Abyssal Recipes: Reconciling the Paradoxes and Suggesting an Upwelling Mechanism for Bottom Water in a Flat-Bottom Ocean

Walter Munk's classical work, known as the "abyssal recipes", introduced a foundational framework for comprehending the upwelling of abyssal waters. While it has spurred numerous investigations into the complexities of deep-ocean processes from theoretical, laboratory, and field perspectives, it has also faced challenges when compared to direct observations spanning decades. Two particularly intriguing paradoxes emerge: the dichotomy of diffusivities and the conundrum of interior downwelling. This study attempts to resolve these paradoxes by examining the isopycnal displacement velocity within the dynamic framework of Munk's abyssal recipes. A relieving inference is that it seems no longer an imperative to seek a globally averaged diapycnal diffusivity of O(1) cm2/s to close the mass budget associated with the bottom-water formation rate. Through box-model experiments, a novel erosion-intrusion model is proposed to elucidate the mechanism of abyssal upwelling in flat-bottom ocean, drawing an analogy to the growth of "tree rings". Based upon this model, the average rising velocity in the abyssal North Pacific is estimated using observational data of biogeochemical tracers. The proposed model does not seek to disprove existing theories but rather acts as a complement for situations that prevailing theories may not apply, such as cases with flat-bottom topography or non-bottom-intensified diffusivity.

physics.ao-ph↗

arXiv2024

SilverCycling: Exploring the Impact of Bike-Based Locomotion on Spatial Orientation for Older Adults in VR

Spatial orientation is essential for people to effectively navigate and interact with the environment in everyday life. With age-related cognitive decline, providing VR locomotion techniques with better spatial orientation performance for older adults becomes important. Such advancements not only make VR more accessible to older adults but also enable them to reap the potential health benefits of VR technology. Natural motion-based locomotion has been shown to be effective in enhancing younger users' performance in VR navigation tasks that require spatial orientation. However, there is a lack of understanding regarding the impact of natural motion-based locomotion on spatial orientation for older adults in VR. To address this gap, we selected the SilverCycling system, a VR bike-based locomotion technique that we developed, as a representative of natural motion-based locomotion, guided by findings from our pilot study. We conducted a user study with 16 older adults to compare SilverCycling with the joystick-based controller. The findings suggest SilverCycling's potential to significantly enhance spatial orientation in the open-road urban environment for older adults, offering a better user experience. Based on our findings, we identify key factors influencing spatial orientation and propose design recommendations to make VR locomotion more accessible and user-friendly for older adults.

cs.HC↗

arXiv2024

Lifelike Agility and Play in Quadrupedal Robots using Reinforcement Learning and Generative Pre-trained Models

Knowledge from animals and humans inspires robotic innovations. Numerous efforts have been made to achieve agile locomotion in quadrupedal robots through classical controllers or reinforcement learning approaches. These methods usually rely on physical models or handcrafted rewards to accurately describe the specific system, rather than on a generalized understanding like animals do. Here we propose a hierarchical framework to construct primitive-, environmental- and strategic-level knowledge that are all pre-trainable, reusable and enrichable for legged robots. The primitive module summarizes knowledge from animal motion data, where, inspired by large pre-trained models in language and image understanding, we introduce deep generative models to produce motor control signals stimulating legged robots to act like real animals. Then, we shape various traversing capabilities at a higher level to align with the environment by reusing the primitive module. Finally, a strategic module is trained focusing on complex downstream tasks by reusing the knowledge from previous levels. We apply the trained hierarchical controllers to the MAX robot, a quadrupedal robot developed in-house, to mimic animals, traverse complex obstacles and play in a designed challenging multi-agent chase tag game, where lifelike agility and strategy emerge in the robots.

cs.RO↗

arXiv2024

Computational Graph Representation of Equations System Constructors in Hierarchical Circuit Simulation

Equations system constructors of hierarchical circuits play a central role in device modeling, nonlinear equations solving, and circuit design automation. However, existing constructors present limitations in applications to different extents. For example, the costs of developing and reusing device models -- especially coarse-grained equivalent models of circuit modules -- remain high while parameter sensitivity analysis is complex and inefficient. Inspired by differentiable programming and leveraging the ecosystem benefits of open-source software, we propose an equations system constructor using the computational graph representation, along with its JSON format netlist, to address these limitations. This representation allows for runtime dependencies between signals and subcircuit/device parameters. The proposed method streamlines the model development process and facilitates end-to-end computation of gradients of equations remainders with respect to parameters. This paper discusses in detail the overarching concept of hierarchical subcircuit/device decomposition and nested invocation by drawing parallels to functions in programming languages, and introduces rules for parameters passing and gradient propagation across hierarchical circuit modules. The presented numerical examples, including (1) an uncoupled CMOS model representation using "equivalent circuit decomposition+dynamic parameters" and (2) operational amplifier (OpAmp) auto device sizing, have demonstrated that the proposed method supports circuit simulation and design and particularly subcircuit modeling with improved efficiency, simplicity, and decoupling compared to existing techniques.

math.NA↗

arXiv2024

Q-Star Meets Scalable Posterior Sampling: Bridging Theory and Practice via HyperAgent

We propose HyperAgent, a reinforcement learning (RL) algorithm based on the hypermodel framework for exploration in RL. HyperAgent allows for the efficient incremental approximation of posteriors associated with an optimal action-value function ($Q^\star$) without the need for conjugacy and follows the greedy policies w.r.t. these approximate posterior samples. We demonstrate that HyperAgent offers robust performance in large-scale deep RL benchmarks. It can solve Deep Sea hard exploration problems with episodes that optimally scale with problem size and exhibits significant efficiency gains in the Atari suite. Implementing HyperAgent requires minimal code addition to well-established deep RL frameworks like DQN. We theoretically prove that, under tabular assumptions, HyperAgent achieves logarithmic per-step computational complexity while attaining sublinear regret, matching the best known randomized tabular RL algorithm.

cs.LG↗

arXiv2024

Efficient Preference-based Reinforcement Learning via Aligned Experience Estimation

Preference-based reinforcement learning (PbRL) has shown impressive capabilities in training agents without reward engineering. However, a notable limitation of PbRL is its dependency on substantial human feedback. This dependency stems from the learning loop, which entails accurate reward learning compounded with value/policy learning, necessitating a considerable number of samples. To boost the learning loop, we propose SEER, an efficient PbRL method that integrates label smoothing and policy regularization techniques. Label smoothing reduces overfitting of the reward model by smoothing human preference labels. Additionally, we bootstrap a conservative estimate $\widehat{Q}$ using well-supported state-action pairs from the current replay memory to mitigate overestimation bias and utilize it for policy learning regularization. Our experimental results across a variety of complex tasks, both in online and offline settings, demonstrate that our approach improves feedback efficiency, outperforming state-of-the-art methods by a large margin. Ablation studies further reveal that SEER achieves a more accurate Q-function compared to prior work.

cs.LG↗

arXiv2024

Variational Dynamic for Self-Supervised Exploration in Deep Reinforcement Learning

Efficient exploration remains a challenging problem in reinforcement learning, especially for tasks where extrinsic rewards from environments are sparse or even totally disregarded. Significant advances based on intrinsic motivation show promising results in simple environments but often get stuck in environments with multimodal and stochastic dynamics. In this work, we propose a variational dynamic model based on the conditional variational inference to model the multimodality and stochasticity. We consider the environmental state-action transition as a conditional generative process by generating the next-state prediction under the condition of the current state, action, and latent variable, which provides a better understanding of the dynamics and leads a better performance in exploration. We derive an upper bound of the negative log-likelihood of the environmental transition and use such an upper bound as the intrinsic reward for exploration, which allows the agent to learn skills by self-supervised exploration without observing extrinsic rewards. We evaluate the proposed method on several image-based simulation tasks and a real robotic manipulating task. Our method outperforms several state-of-the-art environment model-based exploration approaches.

cs.LG↗

arXiv2024

Electrical-controllable antiferromagnet-based tunnel junction

Electrical-controllable antiferromagnet tunnel junction is a key goal in spintronics, holding immense promise for ultra-dense and ultra-stable antiferromagnetic memory with high processing speed for modern information technology. Here, we have advanced towards this goal by achieving an electrical-controllable antiferromagnet-based tunnel junction of Pt/Co/Pt/Co/IrMn/MgO/Pt. The exchange coupling between antiferromagnetic IrMn and Co/Pt perpendicular magnetic multilayers results in the formation of interfacial exchange bias and exchange spring in IrMn. Encoding information states 0 and 1 is realized through the exchange spring in IrMn, which can be electrically written by spin-orbit torque switching with high cyclability and electrically read by antiferromagnetic tunneling anisotropic magnetoresistance. Combining spin-orbit torque switching of both exchange spring andexchange bias, 16 Boolean logic operation is successfully demonstrated. With both memory and logic functionalities integrated into our electrical-controllable antiferromagnetic-based tunnel junction, we chart the course toward high-performance antiferromagnetic logic-in-memory.

physics.app-ph↗

arXiv2024

Observation of non-volatile anomalous Nernst effect in altermagnet with collinear Néel vector

Anomalous Nernst effect (ANE), a widely investigated transverse thermoelectric effect that converts waste heat into electrical energy with remarkable flexibility and integration capability, has been extended to antiferromagnets with non-collinear spin texture recently. ANE in compensated magnet with collinear Néel vector will bring more opportunities to construct magnetic-field-immune and ultrafast transverse thermoelectric converters, but remains unachieved for long. It is due to the degenerated band structure of traditional collinear compensated magnet excludes non-zero Berry curvature. Here, we realize non-volatile ANE in altermagnet Mn5Si3 thin film with collinear Neel vector, whose unique alternating spin-splitting band structure plays vital role in creating non-zero Berry curvature and hotpots of anomalous Nernst conductivity near band intersections. Interestingly, ANE is relatively weak in stoichiometric Mn5Si3, but undergoes a sixfold enhancement through strategically raising the Fermi level by additional Mn doping, indicating sensitive intrinsic influence from specific location of the Fermi level on ANE in altermagnet. Moreover, our investigation reveals a unique Neel-vector-dependent temperature-scaling relationship of anomalous Nernst conductivity in Mn5Si3. Our work not only fills a longstanding gap by confirming the presence of non-volatile ANE in collinear compensated magnet, but also enlightens thermoelectric physics related to exotic spin-splitting band structure in altermagnet.

cond-mat.mtrl-sci↗

arXiv2024

An Efficient Model-Based Approach on Learning Agile Motor Skills without Reinforcement

Learning-based methods have improved locomotion skills of quadruped robots through deep reinforcement learning. However, the sim-to-real gap and low sample efficiency still limit the skill transfer. To address this issue, we propose an efficient model-based learning framework that combines a world model with a policy network. We train a differentiable world model to predict future states and use it to directly supervise a Variational Autoencoder (VAE)-based policy network to imitate real animal behaviors. This significantly reduces the need for real interaction data and allows for rapid policy updates. We also develop a high-level network to track diverse commands and trajectories. Our simulated results show a tenfold sample efficiency increase compared to reinforcement learning methods such as PPO. In real-world testing, our policy achieves proficient command-following performance with only a two-minute data collection period and generalizes well to new speeds and paths.

cs.RO↗

arXiv2024

Towards Robust Offline Reinforcement Learning under Diverse Data Corruption

Offline reinforcement learning (RL) presents a promising approach for learning reinforced policies from offline datasets without the need for costly or unsafe interactions with the environment. However, datasets collected by humans in real-world environments are often noisy and may even be maliciously corrupted, which can significantly degrade the performance of offline RL. In this work, we first investigate the performance of current offline RL algorithms under comprehensive data corruption, including states, actions, rewards, and dynamics. Our extensive experiments reveal that implicit Q-learning (IQL) demonstrates remarkable resilience to data corruption among various offline RL algorithms. Furthermore, we conduct both empirical and theoretical analyses to understand IQL's robust performance, identifying its supervised policy learning scheme as the key factor. Despite its relative robustness, IQL still suffers from heavy-tail targets of Q functions under dynamics corruption. To tackle this challenge, we draw inspiration from robust statistics to employ the Huber loss to handle the heavy-tailedness and utilize quantile estimators to balance penalization for corrupted data and learning stability. By incorporating these simple yet effective modifications into IQL, we propose a more robust offline RL approach named Robust IQL (RIQL). Extensive experiments demonstrate that RIQL exhibits highly robust performance when subjected to diverse data corruption scenarios.

cs.LG↗