Searcharxiv⌕ Search

arXiv subjects

Zhuoran Li

Publications and source records attributed to Zhuoran Li.

71 records · Page 4Linked to original sources
arXiv2024

D$^3$Fields: Dynamic 3D Descriptor Fields for Zero-Shot Generalizable Rearrangement

Scene representation is a crucial design choice in robotic manipulation systems. An ideal representation is expected to be 3D, dynamic, and semantic to meet the demands of diverse manipulation tasks. However, previous works often lack all three properties simultaneously. In this work, we introduce D$^3$Fields -- dynamic 3D descriptor fields. These fields are implicit 3D representations that take in 3D points and output semantic features and instance masks. They can also capture the dynamics of the underlying 3D environments. Specifically, we project arbitrary 3D points in the workspace onto multi-view 2D visual observations and interpolate features derived from visual foundational models. The resulting fused descriptor fields allow for flexible goal specifications using 2D images with varied contexts, styles, and instances. To evaluate the effectiveness of these descriptor fields, we apply our representation to rearrangement tasks in a zero-shot manner. Through extensive evaluation in real worlds and simulations, we demonstrate that D$^3$Fields are effective for zero-shot generalizable rearrangement tasks. We also compare D$^3$Fields with state-of-the-art implicit 3D representations and show significant improvements in effectiveness and efficiency.

cs.RO↗
arXiv2024

RoboEXP: Action-Conditioned Scene Graph via Interactive Exploration for Robotic Manipulation

We introduce the novel task of interactive scene exploration, wherein robots autonomously explore environments and produce an action-conditioned scene graph (ACSG) that captures the structure of the underlying environment. The ACSG accounts for both low-level information (geometry and semantics) and high-level information (action-conditioned relationships between different entities) in the scene. To this end, we present the Robotic Exploration (RoboEXP) system, which incorporates the Large Multimodal Model (LMM) and an explicit memory design to enhance our system's capabilities. The robot reasons about what and how to explore an object, accumulating new information through the interaction process and incrementally constructing the ACSG. Leveraging the constructed ACSG, we illustrate the effectiveness and efficiency of our RoboEXP system in facilitating a wide range of real-world manipulation tasks involving rigid, articulated objects, nested objects, and deformable objects.

cs.RO↗
arXiv2024

Value-Based Deep Multi-Agent Reinforcement Learning with Dynamic Sparse Training

Deep Multi-agent Reinforcement Learning (MARL) relies on neural networks with numerous parameters in multi-agent scenarios, often incurring substantial computational overhead. Consequently, there is an urgent need to expedite training and enable model compression in MARL. This paper proposes the utilization of dynamic sparse training (DST), a technique proven effective in deep supervised learning tasks, to alleviate the computational burdens in MARL training. However, a direct adoption of DST fails to yield satisfactory MARL agents, leading to breakdowns in value learning within deep sparse value-based MARL models. Motivated by this challenge, we introduce an innovative Multi-Agent Sparse Training (MAST) framework aimed at simultaneously enhancing the reliability of learning targets and the rationality of sample distribution to improve value learning in sparse models. Specifically, MAST incorporates the Soft Mellowmax Operator with a hybrid TD-($λ$) schema to establish dependable learning targets. Additionally, it employs a dual replay buffer mechanism to enhance the distribution of training samples. Building upon these aspects, MAST utilizes gradient-based topology evolution to exclusively train multiple MARL agents using sparse networks. Our comprehensive experimental investigation across various value-based MARL algorithms on multiple benchmarks demonstrates, for the first time, significant reductions in redundancy of up to $20\times$ in Floating Point Operations (FLOPs) for both training and inference, with less than $3\%$ performance degradation.

cs.LG↗
arXiv2024

Improving Zero-Shot Cross-Lingual Transfer via Progressive Code-Switching

Code-switching is a data augmentation scheme mixing words from multiple languages into source lingual text. It has achieved considerable generalization performance of cross-lingual transfer tasks by aligning cross-lingual contextual word representations. However, uncontrolled and over-replaced code-switching would augment dirty samples to model training. In other words, the excessive code-switching text samples will negatively hurt the models' cross-lingual transferability. To this end, we propose a Progressive Code-Switching (PCS) method to gradually generate moderately difficult code-switching examples for the model to discriminate from easy to hard. The idea is to incorporate progressively the preceding learned multilingual knowledge using easier code-switching data to guide model optimization on succeeding harder code-switching data. Specifically, we first design a difficulty measurer to measure the impact of replacing each word in a sentence based on the word relevance score. Then a code-switcher generates the code-switching data of increasing difficulty via a controllable temperature variable. In addition, a training scheduler decides when to sample harder code-switching data for model training. Experiments show our model achieves state-of-the-art results on three different zero-shot cross-lingual transfer tasks across ten languages.

cs.CL↗
arXiv2024

Unify the effect of anharmonicity in double-wells and anharmonic oscillators

We study the effect of anharmonicity in quantum anharmonic oscillators, by computing the energy gap between the ground and the 1st excited state using the numerical bootstrap method. Based on perturbative formulae of limiting coupling regimes, we propose a qualitative formula of the energy gap across all coupling values. Except detailed numerical parameter values, the proposed formula has the same functional form as the formula of ground state level splitting of double-well potentials, proposed recently in arXiv:2308.11516. This unifies the effects of anharmonicity in both the anharmonic oscillators and the double-well potentials, although the underlying physical process of them are completely different. We give an explanation of this connection of their anharmonicity from the viewpoint of quantum phase transitions. This connection is justified up to the octic anharmonicities by the numerical bootstrap.

quant-ph↗
arXiv2023

Restriction estimates for one class of hypersurfaces with vanishing curvature in $\mathbb{R}^n$

In this paper, we study the restriction problem for one class of hypersurfaces with vanishing curvature in $\mathbb{R}^n$ with $n$ being odd. We obtain an $L^2-L^p$ restriction estimate, which is optimal except at the endpoint. Furthermore, we establish an $L^s-L^p$ restriction estimate for these hypersurfaces, which is achieved by improving the known $L^{\infty}$ restriction estimate for hypersurfaces with $\frac{n-1}{2}$ positive principal curvatures and $\frac{n-1}{2}$ negative principal curvatures.

math.AP↗
arXiv2023

Sensing User's Activity, Channel, and Location with Near-Field Extra-Large-Scale MIMO

This paper proposes a grant-free massive access scheme based on the millimeter wave (mmWave) extra-large-scale multiple-input multiple-output (XL-MIMO) to support massive Internet-of-Things (IoT) devices with low latency, high data rate, and high localization accuracy in the upcoming sixth-generation (6G) networks. The XL-MIMO consists of multiple antenna subarrays that are widely spaced over the service area to ensure line-of-sight (LoS) transmissions. First, we establish the XL-MIMO-based massive access model considering the near-field spatial non-stationary (SNS) property. Then, by exploiting the block sparsity of subarrays and the SNS property, we propose a structured block orthogonal matching pursuit algorithm for efficient active user detection (AUD) and channel estimation (CE). Furthermore, different sensing matrices are applied in different pilot subcarriers for exploiting the diversity gains. Additionally, a multi-subarray collaborative localization algorithm is designed for localization. In particular, the angle of arrival (AoA) and time difference of arrival (TDoA) of the LoS links between active users and related subarrays are extracted from the estimated XL-MIMO channels, and then the coordinates of active users are acquired by jointly utilizing the AoAs and TDoAs. Simulation results show that the proposed algorithms outperform existing algorithms in terms of AUD and CE performance and can achieve centimeter-level localization accuracy.

cs.IT↗
arXiv2023

Sensing User's Channel and Location with Terahertz Extra-Large Reconfigurable Intelligent Surface under Hybrid-Field Beam Squint Effect

This paper investigates the sensing of user's uplink channel and location in terahertz extra-large reconfigurable intelligent surface (XL-RIS) systems, where the unique hybrid far-near field effect and the beam squint effect caused by the XL array aperture as well as the XL bandwidth are overcome. Specifically, we first propose a joint channel and location sensing scheme, which consists of a location-assisted generalized multiple measurement vector orthogonal matching pursuit (LA-GMMV-OMP) algorithm for channel estimation (CE) and a complete dictionary based localization (CDL) scheme, where a frequency selective polar-domain redundant dictionary is proposed to overcome the hybrid field beam squint effect. The CE module outputs coarse on-grid angle estimation (respectively observed from the BS and RIS) to the localization module, which returns the fine off-grid angle estimation to improve CE. Particularly, with RIS, CDL can obtain user's location via line intersection, and a polar-domain gradient descent (PGD) algorithm at the base station is proposed to achieve the off-grid angle estimation with super-resolution accuracy. Additionally, to further reduce the sensing overhead, we propose a partial dictionary-based localization scheme, which is decoupled from CE, where RIS is served as an anchor to lock the user on the hyperbola according to time difference of arrival and the user's off-grid location can be obtained by using the proposed PGD algorithm. Simulation results demonstrate the superiority of the two proposed localization schemes and the proposed CE scheme over state-of-the-art baseline approaches.

eess.SP↗
arXiv2023

Indium-Tin-Oxide for High-performance Electro-optic Modulation

Advances in opto-electronics are often led by discovery and development of materials featuring unique properties. Recently the material class of transparent conductive oxides (TCO) has attracted attention for active photonic devices on-chip. In particular Indium Tin Oxide (ITO) is found to have refractive index changes on the order of unity. This property makes it possible to achieve electro-optic modulation of sub-wavelength device scales, when thin ITO films are interfaced with optical light confinement techniques such as found in plasmonics; optical modes are compressed to nanometer scale to create strong light-matter-interactions. Here we review efforts towards utilizing this novel material for high-performance and ultra-compact modulation. While high performance metrics are achieved experimentally, there are open questions pertaining the permittivity modulation mechanism of ITO. Furthermore, we show that a footprint-saving waveguide inline cavity can enhance obtainable extinction-ratio to insertion-loss ratios by about one order of magnitude over non-cavity based version. Moreover, we offer a speed analysis that shows that the device is resistance limited, but not capacitance or drift-carrier limited. Interestingly, two bias options exist for ITO and we find that a side-connection enables devices that should in principle enable several hundred of GHz fast devices, using our routinely achievable ITO film resistivities. Finally, we offer a brief discuss about footprint savings of compact ITO modulators showing a 3-orders of magnitude smaller footprint over Silicon photonic MZI-based modulators.

physics.optics↗
arXiv2022

Time Series Analysis in American Stock Market Recovering in Post COVID-19 Pandemic Period

Every financial crisis has caused a dual shock to the global economy. The shortage of market liquidity, such as default in debt and bonds, has led to the spread of bankruptcies, such as Lehman Brothers in 2008. Using the data for the ETFs of the S&P 500, Nasdaq 100, and Dow Jones Industrial Average collected from Yahoo Finance, this study implemented Deep Learning, Neuro Network, and Time-series to analyze the trend of the American Stock Market in the post-COVID-19 period. LSTM model in Neuro Network to predict the future trend, which suggests the US stock market keeps falling for the post-COVID-19 period. This study reveals a reasonable allocation method of Long Short-Term Memory for which there is strong evidence.

q-fin.ST↗
arXiv2022

$L^2$ Schrödinger maximal estimates associated with finite type phases in $\mathbb{R}^2$

In this paper, we establish Schrödinger maximal estimates associated with the finite type phases \begin{equation*} ϕ(ξ_1,ξ_2):=ξ^m_1+ξ^m_2,\;(ξ_1,ξ_2)\in [0,1]^2, \end{equation*} where $m \geq 4$ is an even number. Following [12], we prove an $L^2$ fractal restriction estimate associated with the surfaces \begin{equation*} F^2_m:=\{(ξ_1,ξ_2,ϕ(ξ_1,ξ_2)):\;(ξ_1,ξ_2)\in [0,1]^2\} \end{equation*} as the main result, which also gives results on the average Fourier decay of fractal measures associated with these surfaces. The key ingredients of the proof include the rescaling technique from [16], Bourgain-Demeter's $\ell^2$ decoupling inequality, the reduction of dimension arguments from [17] and induction on scales.

math.CA↗
arXiv2021

$\ell^2$ decoupling for certain surfaces of finite type in $\mathbb{R}^3$

In this article, we establish an $\ell^2$ decoupling inequality for the surface $$F_4^2:=\Big\{(ξ_1,ξ_2,ξ_1^4+ξ_2^4): (ξ_1,ξ_2) \in [0,1]^2\Big\}$$ associated with the decomposition adapted to finite type geometry from our previous work. The key ingredients of the proof include the so-called generalized rescaling technique, an $\ell^2$ decoupling inequality for the surfaces $$\Big\{(ξ_1,ξ_2,ϕ_1(ξ_1)+ξ_2^4): (ξ_1,ξ_2) \in [0,1]^2\Big\}$$ with $ϕ_1$ being non-degenerate, reduction of dimension arguments and induction on scales.

math.AP↗
arXiv2020

Assessing the Quality-of-Experience of Adaptive Bitrate Video Streaming

The diversity of video delivery pipeline poses a grand challenge to the evaluation of adaptive bitrate (ABR) streaming algorithms and objective quality-of-experience (QoE) models. Here we introduce so-far the largest subject-rated database of its kind, namely WaterlooSQoE-IV, consisting of 1350 adaptive streaming videos created from diverse source contents, video encoders, network traces, ABR algorithms, and viewing devices. We collect human opinions for each video with a series of carefully designed subjective experiments. Subsequent data analysis and testing/comparison of ABR algorithms and QoE models using the database lead to a series of novel observations and interesting findings, in terms of the effectiveness of subjective experiment methodologies, the interactions between user experience and source content, viewing device and encoder type, the heterogeneities in the bias and preference of user experiences, the behaviors of ABR algorithms, and the performance of objective QoE models. Most importantly, our results suggest that a better objective QoE model, or a better understanding of human perceptual experience and behaviour, is the most dominating factor in improving the performance of ABR algorithms, as opposed to advanced optimization frameworks, machine learning strategies or bandwidth predictors, where a majority of ABR research has been focused on in the past decade. On the other hand, our performance evaluation of 11 QoE models shows only a moderate correlation between state-of-the-art QoE models and subjective ratings, implying rooms for improvement in both QoE modeling and ABR algorithms. The database is made publicly available at: \url{https://ece.uwaterloo.ca/~zduanmu/waterloosqoe4/}.

eess.IV↗
arXiv2020

Restriction estimates for certain surfaces of finite type in $\mathbb{R}^3$

In this paper, we study the restriction estimate for a certain surface of finite type in $\mathbb{R}^3$, and partially improves the results of Buschenhenke-Müller-Vargas. The key ingredients of the proof include the so called generalized rescaling technique based on a decomposition adapted to finite type geometry, a decoupling inequality and reduction of dimension arguments.

math.AP↗
arXiv2020

Characterizing Generalized Rate-Distortion Performance of Video Coding: An Eigen Analysis Approach

Rate-distortion (RD) theory is at the heart of lossy data compression. Here we aim to model the generalized RD (GRD) trade-off between the visual quality of a compressed video and its encoding profiles (e.g., bitrate and spatial resolution). We first define the theoretical functional space $\mathcal{W}$ of the GRD function by analyzing its mathematical properties.We show that $\mathcal{W}$ is a convex set in a Hilbert space, inspiring a computational model of the GRD function, and a method of estimating model parameters from sparse measurements. To demonstrate the feasibility of our idea, we collect a large-scale database of real-world GRD functions, which turn out to live in a low-dimensional subspace of $\mathcal{W}$. Combining the GRD reconstruction framework and the learned low-dimensional space, we create a low-parameter eigen GRD method to accurately estimate the GRD function of a source video content from only a few queries. Experimental results on the database show that the learned GRD method significantly outperforms state-of-the-art empirical RD estimation methods both in accuracy and efficiency. Last, we demonstrate the promise of the proposed model in video codec comparison.

eess.IV↗
arXiv2019

A Knowledge-Driven Quality-of-Experience Model for Adaptive Streaming Videos

The fundamental conflict between the enormous space of adaptive streaming videos and the limited capacity for subjective experiment casts significant challenges to objective Quality-of-Experience (QoE) prediction. Existing objective QoE models exhibit complex functional form, failing to generalize well in diverse streaming environments. In this study, we propose an objective QoE model namely knowledge-driven streaming quality index (KSQI) to integrate prior knowledge on the human visual system and human annotated data in a principled way. By analyzing the subjective characteristics towards streaming videos from a corpus of subjective studies, we show that a family of QoE functions lies in a convex set. Using a variant of projected gradient descent, we optimize the objective QoE model over a database of training videos. The proposed KSQI demonstrates strong generalizability to diverse streaming environments, evident by state-of-the-art performance on four publicly available benchmark datasets.

cs.MM↗