arXiv subjects

Tao Yu

Publications and source records attributed to Tao Yu.

At least 181 records · Page 10Linked to original sources

arXiv2024

Enhancement of Magnon Transport by Superconductor Meissner Screening

Recent experiments observe the spin-wave-Meissner-current modes in ferromagnetic insulator-superconductor heterostructures, in which the coherently excited spin waves seemingly do not decay as usual beneath the superconductor strip [Borst et al., Science 382, 430 (2023)]. We interpret this phenomenon by demonstrating that the stray magnetic field emitted by the magnetization dynamics is reflected, focused, and enhanced inside the ferromagnet by the supercurrent induced in the superconductor, such that the group velocity of spin waves is strongly enhanced. Analytical and numerical calculations based on this model predict that the coherent transport of magnons is enhanced by close to 500% for yttrium iron garnet capped by superconducting NbN with a decay length exceeding millimeters. Our finding may augment the performance of magnons in quantum information and quantum transport processing.

cond-mat.mes-hall↗

arXiv2024

The 3d-index of the 3d-skein module via the quantum trace map

We define a map from the skein module of a cusped hyperbolic 3-manifold to the ring of Laurent series in one variable with integer coefficients that satisfies two properties: its evaluation at peripheral curves coincides with the Dimofte--Gaiotto--Gukov 3d-index, and it factors through the 3d-quantum trace map associated to a suitable ideal triangulation of the manifold. The map fulfills a supersymmetry prediction of mathematical physics and is part of a conjectural 3+1 dimensional topological quantum field theory.

math.GT↗

arXiv2024

HHMR: Holistic Hand Mesh Recovery by Enhancing the Multimodal Controllability of Graph Diffusion Models

Recent years have witnessed a trend of the deep integration of the generation and reconstruction paradigms. In this paper, we extend the ability of controllable generative models for a more comprehensive hand mesh recovery task: direct hand mesh generation, inpainting, reconstruction, and fitting in a single framework, which we name as Holistic Hand Mesh Recovery (HHMR). Our key observation is that different kinds of hand mesh recovery tasks can be achieved by a single generative model with strong multimodal controllability, and in such a framework, realizing different tasks only requires giving different signals as conditions. To achieve this goal, we propose an all-in-one diffusion framework based on graph convolution and attention mechanisms for holistic hand mesh recovery. In order to achieve strong control generation capability while ensuring the decoupling of multimodal control signals, we map different modalities to a shared feature space and apply cross-scale random masking in both modality and feature levels. In this way, the correlation between different modalities can be fully exploited during the learning of hand priors. Furthermore, we propose Condition-aligned Gradient Guidance to enhance the alignment of the generated model with the control signals, which significantly improves the accuracy of the hand mesh reconstruction and fitting. Experiments show that our novel framework can realize multiple hand mesh recovery tasks simultaneously and outperform the existing methods in different tasks, which provides more possibilities for subsequent downstream applications including gesture recognition, pose generation, mesh editing, and so on.

cs.CV↗

arXiv2024

Interaction-Induced Dimensional Crossover through Full 3D to 1D

The exploration of dimensional crossover carries profound fundamental significance, serving as a crucial bridge in comprehending the remarkable disparities observed in transitional phenomena across the two distinct dimensions of a physical system. The prevalent strategy for manipulating the dimensionality involves meticulously controlling the external trapping geometry, thereby restricting the degrees of freedom of the kinetic energy from three-dimensional (3D) to lower-dimensional spaces, while maintaining the 3D nature of the interaction energy degrees of freedom. The aim of this work is to introduce an innovative scenario to achieve dimensional crossover, characterized by lower-D nature of both the kinetic and the interaction energy degrees of freedom. To accomplish this objective, we delve deeply into the realm of a 2D optically trapped Bose gas, focusing specifically on its finite-range interaction. Our emphasis lies in exploring the lattice-induced dimensional crossover from full 3D to 1D in both kinetic and interaction terms. Utilizing the functional path integral method, we derive the equation of states of the model system, encompassing crucial quantities such as the ground state energy and quantum depletion. These equations enable us to analyze the combined effects of finite range interaction and an optical lattice on quantum fluctuations of the BEC system. Notably, our analytical findings reconcile the Lee-Huang-Yang (LHY) correction to the ground state energy in 3D and Lieb-Liniger (LL) ones in 1D limit, thereby providing fresh insights into the intriguing disparities between LHY and LL corrections.

cond-mat.quant-gas↗

arXiv2024

OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments

Autonomous agents that accomplish complex computer tasks with minimal human interventions have the potential to transform human-computer interaction, significantly enhancing accessibility and productivity. However, existing benchmarks either lack an interactive environment or are limited to environments specific to certain applications or domains, failing to reflect the diverse and complex nature of real-world computer use, thereby limiting the scope of tasks and agent scalability. To address this issue, we introduce OSWorld, the first-of-its-kind scalable, real computer environment for multimodal agents, supporting task setup, execution-based evaluation, and interactive learning across various operating systems such as Ubuntu, Windows, and macOS. OSWorld can serve as a unified, integrated computer environment for assessing open-ended computer tasks that involve arbitrary applications. Building upon OSWorld, we create a benchmark of 369 computer tasks involving real web and desktop apps in open domains, OS file I/O, and workflows spanning multiple applications. Each task example is derived from real-world computer use cases and includes a detailed initial state setup configuration and a custom execution-based evaluation script for reliable, reproducible evaluation. Extensive evaluation of state-of-the-art LLM/VLM-based agents on OSWorld reveals significant deficiencies in their ability to serve as computer assistants. While humans can accomplish over 72.36% of the tasks, the best model achieves only 12.24% success, primarily struggling with GUI grounding and operational knowledge. Comprehensive analysis using OSWorld provides valuable insights for developing multimodal generalist agents that were not possible with previous benchmarks. Our code, environment, baseline models, and data are publicly available at https://os-world.github.io.

cs.AI↗

arXiv2024

Text2Reward: Reward Shaping with Language Models for Reinforcement Learning

Designing reward functions is a longstanding challenge in reinforcement learning (RL); it requires specialized knowledge or domain data, leading to high costs for development. To address this, we introduce Text2Reward, a data-free framework that automates the generation and shaping of dense reward functions based on large language models (LLMs). Given a goal described in natural language, Text2Reward generates shaped dense reward functions as an executable program grounded in a compact representation of the environment. Unlike inverse RL and recent work that uses LLMs to write sparse reward codes or unshaped dense rewards with a constant function across timesteps, Text2Reward produces interpretable, free-form dense reward codes that cover a wide range of tasks, utilize existing packages, and allow iterative refinement with human feedback. We evaluate Text2Reward on two robotic manipulation benchmarks (ManiSkill2, MetaWorld) and two locomotion environments of MuJoCo. On 13 of the 17 manipulation tasks, policies trained with generated reward codes achieve similar or better task success rates and convergence speed than expert-written reward codes. For locomotion tasks, our method learns six novel locomotion behaviors with a success rate exceeding 94%. Furthermore, we show that the policies trained in the simulator with our method can be deployed in the real world. Finally, Text2Reward further improves the policies by refining their reward functions with human feedback. Video results are available at https://text-to-reward.github.io/ .

cs.LG↗

arXiv2024

IREE Oriented Green 6G Networks: A Radial Basis Function Based Approach

In order to provide design guidelines for energy efficient 6G networks, we propose a novel radial basis function (RBF) based optimization framework to maximize the integrated relative energy efficiency (IREE) metric. Different from the conventional energy efficient optimization schemes, we maximize the transformed utility for any given IREE using spectrum efficiency oriented RBF network and gradually update the IREE metric using proposed Dinkelbach's algorithm. The existence and uniqueness properties of RBF networks are provided, and the convergence conditions of the entire framework are discussed as well. Through some numerical experiments, we show that the proposed IREE outperforms many existing SE or EE oriented designs and find a new Jensen-Shannon (JS) divergence constrained region, which behaves differently from the conventional EE-SE region. Meanwhile, by studying IREE-SE trade-offs under different traffic requirements, we suggest that network operators shall spend more efforts to balance the distributions of traffic demands and network capacities in order to improve the IREE performance, especially when the spatial variations of the traffic distribution are significant.

cs.NI↗

arXiv2024

Audio Matters Too! Enhancing Markerless Motion Capture with Audio Signals for String Performance Capture

In this paper, we touch on the problem of markerless multi-modal human motion capture especially for string performance capture which involves inherently subtle hand-string contacts and intricate movements. To fulfill this goal, we first collect a dataset, named String Performance Dataset (SPD), featuring cello and violin performances. The dataset includes videos captured from up to 23 different views, audio signals, and detailed 3D motion annotations of the body, hands, instrument, and bow. Moreover, to acquire the detailed motion annotations, we propose an audio-guided multi-modal motion capture framework that explicitly incorporates hand-string contacts detected from the audio signals for solving detailed hand poses. This framework serves as a baseline for string performance capture in a completely markerless manner without imposing any external devices on performers, eliminating the potential of introducing distortion in such delicate movements. We argue that the movements of performers, particularly the sound-producing gestures, contain subtle information often elusive to visual methods but can be inferred and retrieved from audio cues. Consequently, we refine the vision-based motion capture results through our innovative audio-guided approach, simultaneously clarifying the contact relationship between the performer and the instrument, as deduced from the audio. We validate the proposed framework and conduct ablation studies to demonstrate its efficacy. Our results outperform current state-of-the-art vision-based algorithms, underscoring the feasibility of augmenting visual motion capture with audio modality. To the best of our knowledge, SPD is the first dataset for musical instrument performance, covering fine-grained hand motion details in a multi-modal, large-scale collection.

cs.MM↗

arXiv2024

Collage: Light-Weight Low-Precision Strategy for LLM Training

Large models training is plagued by the intense compute cost and limited hardware memory. A practical solution is low-precision representation but is troubled by loss in numerical accuracy and unstable training rendering the model less useful. We argue that low-precision floating points can perform well provided the error is properly compensated at the critical locations in the training process. We propose Collage which utilizes multi-component float representation in low-precision to accurately perform operations with numerical errors accounted. To understand the impact of imprecision to training, we propose a simple and novel metric which tracks the lost information during training as well as differentiates various precision strategies. Our method works with commonly used low-precision such as half-precision ($16$-bit floating points) and can be naturally extended to work with even lower precision such as $8$-bit. Experimental results show that pre-training using Collage removes the requirement of using $32$-bit floating-point copies of the model and attains similar/better training performance compared to $(16, 32)$-bit mixed-precision strategy, with up to $3.7\times$ speedup and $\sim 15\%$ to $23\%$ less memory usage in practice.

cs.LG↗

arXiv2024

Chiral-Damping-Enhanced Magnon Transmission

The inevitable Gilbert damping in magnetization dynamics is usually regarded as detrimental to spin transport. Here we apply a general feature of chiral non-Hermitian dynamics to a ferromagnetic-insulator--normal-metal heterostructure to show that the strong momentum dependence and chirality of the eddy-current-induced damping also causes beneficial scattering properties: A potential barrier that reflects magnon wave packets becomes unidirectionally transparent in the presence of a metallic cap layer. Passive magnon gates that turn presumably harmful dissipation into useful functionalities should be useful for future quantum magnonic devices.

cond-mat.mes-hall↗

arXiv2024

Shadow Cones: A Generalized Framework for Partial Order Embeddings

Hyperbolic space has proven to be well-suited for capturing hierarchical relations in data, such as trees and directed acyclic graphs. Prior work introduced the concept of entailment cones, which uses partial orders defined by nested cones in the Poincaré ball to model hierarchies. Here, we introduce the ``shadow cones" framework, a physics-inspired entailment cone construction. Specifically, we model partial orders as subset relations between shadows formed by a light source and opaque objects in hyperbolic space. The shadow cones framework generalizes entailment cones to a broad class of formulations and hyperbolic space models beyond the Poincaré ball. This results in clear advantages over existing constructions: for example, shadow cones possess better optimization properties over constructions limited to the Poincaré ball. Our experiments on datasets of various sizes and hierarchical structures show that shadow cones consistently and significantly outperform existing entailment cone constructions. These results indicate that shadow cones are an effective way to model partial orders in hyperbolic space, offering physically intuitive and novel insights about the nature of such structures.

cs.LG↗

arXiv2024

MMVP: A Multimodal MoCap Dataset with Vision and Pressure Sensors

Foot contact is an important cue for human motion capture, understanding, and generation. Existing datasets tend to annotate dense foot contact using visual matching with thresholding or incorporating pressure signals. However, these approaches either suffer from low accuracy or are only designed for small-range and slow motion. There is still a lack of a vision-pressure multimodal dataset with large-range and fast human motion, as well as accurate and dense foot-contact annotation. To fill this gap, we propose a Multimodal MoCap Dataset with Vision and Pressure sensors, named MMVP. MMVP provides accurate and dense plantar pressure signals synchronized with RGBD observations, which is especially useful for both plausible shape estimation, robust pose fitting without foot drifting, and accurate global translation tracking. To validate the dataset, we propose an RGBD-P SMPL fitting method and also a monocular-video-based baseline framework, VP-MoCap, for human motion capture. Experiments demonstrate that our RGBD-P SMPL Fitting results significantly outperform pure visual motion capture. Moreover, VP-MoCap outperforms SOTA methods in foot-contact and global translation estimation accuracy. We believe the configuration of the dataset and the baseline frameworks will stimulate the research in this direction and also provide a good reference for MoCap applications in various domains. Project page: https://metaverse-ai-lab-thu.github.io/MMVP-Dataset/.

cs.CV↗

arXiv2024

A Quantum trace map for 3-manifolds

We define a quantum trace map from the skein module of a 3-manifold with torus boundary components to a module (left and right quotient of a quantum torus) constructed from an ideal triangulation. Our map is a 3-dimensional version of the well-known quantum trace map on surfaces introduced by Bonahon and Wong and further developed by Lê.

math.GT↗

arXiv2024

Den-SOFT: Dense Space-Oriented Light Field DataseT for 6-DOF Immersive Experience

We have built a custom mobile multi-camera large-space dense light field capture system, which provides a series of high-quality and sufficiently dense light field images for various scenarios. Our aim is to contribute to the development of popular 3D scene reconstruction algorithms such as IBRnet, NeRF, and 3D Gaussian splitting. More importantly, the collected dataset, which is much denser than existing datasets, may also inspire space-oriented light field reconstruction, which is potentially different from object-centric 3D reconstruction, for immersive VR/AR experiences. We utilized a total of 40 GoPro 10 cameras, capturing images of 5k resolution. The number of photos captured for each scene is no less than 1000, and the average density (view number within a unit sphere) is 134.68. It is also worth noting that our system is capable of efficiently capturing large outdoor scenes. Addressing the current lack of large-space and dense light field datasets, we made efforts to include elements such as sky, reflections, lights and shadows that are of interest to researchers in the field of 3D reconstruction during the data capture process. Finally, we validated the effectiveness of our provided dataset on three popular algorithms and also integrated the reconstructed 3DGS results into the Unity engine, demonstrating the potential of utilizing our datasets to enhance the realism of virtual reality (VR) and create feasible interactive spaces. The dataset is available at our project website.

cs.CV↗

arXiv2024

AutoDroid: LLM-powered Task Automation in Android

Mobile task automation is an attractive technique that aims to enable voice-based hands-free user interaction with smartphones. However, existing approaches suffer from poor scalability due to the limited language understanding ability and the non-trivial manual efforts required from developers or end-users. The recent advance of large language models (LLMs) in language understanding and reasoning inspires us to rethink the problem from a model-centric perspective, where task preparation, comprehension, and execution are handled by a unified language model. In this work, we introduce AutoDroid, a mobile task automation system capable of handling arbitrary tasks on any Android application without manual efforts. The key insight is to combine the commonsense knowledge of LLMs and domain-specific knowledge of apps through automated dynamic analysis. The main components include a functionality-aware UI representation method that bridges the UI with the LLM, exploration-based memory injection techniques that augment the app-specific domain knowledge of LLM, and a multi-granularity query optimization module that reduces the cost of model inference. We integrate AutoDroid with off-the-shelf LLMs including online GPT-4/GPT-3.5 and on-device Vicuna, and evaluate its performance on a new benchmark for memory-augmented Android task automation with 158 common tasks. The results demonstrated that AutoDroid is able to precisely generate actions with an accuracy of 90.9%, and complete tasks with a success rate of 71.3%, outperforming the GPT-4-powered baselines by 36.4% and 39.7%. The demo, benchmark suites, and source code of AutoDroid will be released at url{https://autodroid-sys.github.io/}.

cs.AI↗

arXiv2024

Does Collaborative Human-LM Dialogue Generation Help Information Extraction from Human Dialogues?

The capabilities of pretrained language models have opened opportunities to explore new application areas, but applications involving human-human interaction are limited by the fact that most data is protected from public release for privacy reasons. Problem-solving human dialogues in real applications can be much more complex than existing Wizard-of-Oz collections, preventing successful domain transfer. To support information extraction (IE) for a private call center dataset, we introduce a human-in-the-loop dialogue generation framework capable of synthesizing realistic dialogues. In IE experiments with auto insurance call center dialogues, we observe 25\% relative improvement in $F_1$ after augmenting a small set of real human conversations with synthetic data. We release code and our synthetic dataset to illustrate the complexity of real-world call center conversations and encourage development of complex dialogue datasets that are more representative of natural data.

cs.CL↗

arXiv2024

Smart Mobility Digital Twin Based Automated Vehicle Navigation System: A Proof of Concept

Digital twins (DTs) have driven major advancements across various industrial domains over the past two decades. With the rapid advancements in autonomous driving and vehicle-to-everything (V2X) technologies, integrating DTs into vehicular platforms is anticipated to further revolutionize smart mobility systems. In this paper, a new smart mobility DT (SMDT) platform is proposed for the control of connected and automated vehicles (CAVs) over next-generation wireless networks. In particular, the proposed platform enables cloud services to leverage the abilities of DTs to promote the autonomous driving experience. To enhance traffic efficiency and road safety measures, a novel navigation system that exploits available DT information is designed. The SMDT platform and navigation system are implemented with state-of-the-art products, e.g., CAVs and roadside units (RSUs), and emerging technologies, e.g., cloud and cellular V2X (C-V2X). In addition, proof-of-concept (PoC) experiments are conducted to validate system performance. The performance of SMDT is evaluated from two standpoints: (i) the rewards of the proposed navigation system on traffic efficiency and safety and, (ii) the latency and reliability of the SMDT platform. Our experimental results using SUMO-based large-scale traffic simulations show that the proposed SMDT can reduce the average travel time and the blocking probability due to unexpected traffic incidents. Furthermore, the results record a peak overall latency for DT modeling and route planning services to be 155.15 ms and 810.59 ms, respectively, which validates that our proposed design aligns with the 3GPP requirements for emerging V2X use cases and fulfills the targets of the proposed design. Our demonstration video can be found at https://youtu.be/3waQwlaHQkk.

eess.SY↗

arXiv2024

OS-Copilot: Towards Generalist Computer Agents with Self-Improvement

Autonomous interaction with the computer has been a longstanding challenge with great potential, and the recent proliferation of large language models (LLMs) has markedly accelerated progress in building digital agents. However, most of these agents are designed to interact with a narrow domain, such as a specific software or website. This narrow focus constrains their applicability for general computer tasks. To this end, we introduce OS-Copilot, a framework to build generalist agents capable of interfacing with comprehensive elements in an operating system (OS), including the web, code terminals, files, multimedia, and various third-party applications. We use OS-Copilot to create FRIDAY, a self-improving embodied agent for automating general computer tasks. On GAIA, a general AI assistants benchmark, FRIDAY outperforms previous methods by 35%, showcasing strong generalization to unseen applications via accumulated skills from previous tasks. We also present numerical and quantitative evidence that FRIDAY learns to control and self-improve on Excel and Powerpoint with minimal supervision. Our OS-Copilot framework and empirical findings provide infrastructure and insights for future research toward more capable and general-purpose computer agents.

cs.AI↗