Searcharxiv⌕ Search

arXiv subjects

Zhenyu Li

Publications and source records attributed to Zhenyu Li.

At least 109 records · Page 6Linked to original sources
arXiv2023

PatchFusion: An End-to-End Tile-Based Framework for High-Resolution Monocular Metric Depth Estimation

Single image depth estimation is a foundational task in computer vision and generative modeling. However, prevailing depth estimation models grapple with accommodating the increasing resolutions commonplace in today's consumer cameras and devices. Existing high-resolution strategies show promise, but they often face limitations, ranging from error propagation to the loss of high-frequency details. We present PatchFusion, a novel tile-based framework with three key components to improve the current state of the art: (1) A patch-wise fusion network that fuses a globally-consistent coarse prediction with finer, inconsistent tiled predictions via high-level feature guidance, (2) A Global-to-Local (G2L) module that adds vital context to the fusion network, discarding the need for patch selection heuristics, and (3) A Consistency-Aware Training (CAT) and Inference (CAI) approach, emphasizing patch overlap consistency and thereby eradicating the necessity for post-processing. Experiments on UnrealStereo4K, MVS-Synth, and Middleburry 2014 demonstrate that our framework can generate high-resolution depth maps with intricate details. PatchFusion is independent of the base model for depth estimation. Notably, our framework built on top of SOTA ZoeDepth brings improvements for a total of 17.3% and 29.4% in terms of the root mean squared error (RMSE) on UnrealStereo4K and MVS-Synth, respectively.

cs.CV↗
arXiv2023

Enhancing the Spatial Awareness Capability of Multi-Modal Large Language Model

The Multi-Modal Large Language Model (MLLM) refers to an extension of the Large Language Model (LLM) equipped with the capability to receive and infer multi-modal data. Spatial awareness stands as one of the crucial abilities of MLLM, encompassing diverse skills related to understanding spatial relationships among objects and between objects and the scene area. Industries such as autonomous driving, smart healthcare, robotics, virtual, and augmented reality heavily demand MLLM's spatial awareness capabilities. However, there exists a noticeable gap between the current spatial awareness capabilities of MLLM and the requirements set by human needs. To address this issue, this paper proposes using more precise spatial position information between objects to guide MLLM in providing more accurate responses to user-related inquiries. Specifically, for a particular multi-modal task, we utilize algorithms for acquiring geometric spatial information and scene graphs to obtain relevant geometric spatial information and scene details of objects involved in the query. Subsequently, based on this information, we direct MLLM to address spatial awareness-related queries posed by the user. Extensive experiments were conducted in benchmarks such as MME, MM-Vet, and other multi-modal large language models. The experimental results thoroughly confirm the efficacy of the proposed method in enhancing the spatial awareness tasks and associated tasks of MLLM.

cs.AI↗
arXiv2023

Enhancing Subtask Performance of Multi-modal Large Language Model

Multi-modal Large Language Model (MLLM) refers to a model expanded from a Large Language Model (LLM) that possesses the capability to handle and infer multi-modal data. Current MLLMs typically begin by using LLMs to decompose tasks into multiple subtasks, then employing individual pre-trained models to complete specific subtasks, and ultimately utilizing LLMs to integrate the results of each subtasks to obtain the results of the task. In real-world scenarios, when dealing with large projects, it is common practice to break down the project into smaller sub-projects, with different teams providing corresponding solutions or results. The project owner then decides which solution or result to use, ensuring the best possible outcome for each subtask and, consequently, for the entire project. Inspired by this, this study considers selecting multiple pre-trained models to complete the same subtask. By combining the results from multiple pre-trained models, the optimal subtask result is obtained, enhancing the performance of the MLLM. Specifically, this study first selects multiple pre-trained models focused on the same subtask based on distinct evaluation approaches, and then invokes these models in parallel to process input data and generate corresponding subtask results. Finally, the results from multiple pre-trained models for the same subtask are compared using the LLM, and the best result is chosen as the outcome for that subtask. Extensive experiments are conducted in this study using GPT-4 annotated datasets and human-annotated datasets. The results of various evaluation metrics adequately demonstrate the effectiveness of the proposed approach in this paper.

cs.CL↗
arXiv2023

SpOctA: A 3D Sparse Convolution Accelerator with Octree-Encoding-Based Map Search and Inherent Sparsity-Aware Processing

Point-cloud-based 3D perception has attracted great attention in various applications including robotics, autonomous driving and AR/VR. In particular, the 3D sparse convolution (SpConv) network has emerged as one of the most popular backbones due to its excellent performance. However, it poses severe challenges to real-time perception on general-purpose platforms, such as lengthy map search latency, high computation cost, and enormous memory footprint. In this paper, we propose SpOctA, a SpConv accelerator that enables high-speed and energy-efficient point cloud processing. SpOctA parallelizes the map search by utilizing algorithm-architecture co-optimization based on octree encoding, thereby achieving 8.8-21.2x search speedup. It also attenuates the heavy computational workload by exploiting inherent sparsity of each voxel, which eliminates computation redundancy and saves 44.4-79.1% processing latency. To optimize on-chip memory management, a SpConv-oriented non-uniform caching strategy is introduced to reduce external memory access energy by 57.6% on average. Implemented on a 40nm technology and extensively evaluated on representative benchmarks, SpOctA rivals the state-of-the-art SpConv accelerators by 1.1-6.9x speedup with 1.5-3.1x energy efficiency improvement.

cs.AR↗
arXiv2023

Secure Split Learning against Property Inference, Data Reconstruction, and Feature Space Hijacking Attacks

Split learning of deep neural networks (SplitNN) has provided a promising solution to learning jointly for the mutual interest of a guest and a host, which may come from different backgrounds, holding features partitioned vertically. However, SplitNN creates a new attack surface for the adversarial participant, holding back its practical use in the real world. By investigating the adversarial effects of highly threatening attacks, including property inference, data reconstruction, and feature hijacking attacks, we identify the underlying vulnerability of SplitNN and propose a countermeasure. To prevent potential threats and ensure the learning guarantees of SplitNN, we design a privacy-preserving tunnel for information exchange between the guest and the host. The intuition is to perturb the propagation of knowledge in each direction with a controllable unified solution. To this end, we propose a new activation function named R3eLU, transferring private smashed data and partial loss into randomized responses in forward and backward propagations, respectively. We give the first attempt to secure split learning against three threatening attacks and present a fine-grained privacy budget allocation scheme. The analysis proves that our privacy-preserving SplitNN solution provides a tight privacy budget, while the experimental results show that our solution performs better than existing solutions in most cases and achieves a good tradeoff between defense and model usability.

cs.LG↗
arXiv2023

Circuit-Depth Reduction of Unitary-Coupled-Cluster Ansatz by Energy Sorting

Quantum computation represents a revolutionary approach for solving problems in quantum chemistry. However, due to the limited quantum resources in the current noisy intermediate-scale quantum (NISQ) devices, quantum algorithms for large chemical systems remains a major task. In this work, we demonstrate that the circuit depth of the unitary coupled cluster (UCC) and UCC-based ansatzes in the algorithm of variational quantum eigensolver can be significantly reduced by an energy-sorting strategy. Specifically, subsets of excitation operators are first pre-screened from the operator pool according to its contribution to the total energy. The quantum circuit ansatz is then iteratively constructed until the convergence of the final energy to a typical accuracy. For demonstration, this method has been successfully applied to molecular and periodic systems. Particularly, a reduction of 50\%$\sim$98\% in the number of operators is observed while retaining the accuracy of the origin UCCSD operator pools. This method can be straightforwardly extended to general parametric variational ansatzes.

quant-ph↗
arXiv2023

Bridging the Language Gap: Knowledge Injected Multilingual Question Answering

Question Answering (QA) is the task of automatically answering questions posed by humans in natural languages. There are different settings to answer a question, such as abstractive, extractive, boolean, and multiple-choice QA. As a popular topic in natural language processing tasks, extractive question answering task (extractive QA) has gained extensive attention in the past few years. With the continuous evolvement of the world, generalized cross-lingual transfer (G-XLT), where question and answer context are in different languages, poses some unique challenges over cross-lingual transfer (XLT), where question and answer context are in the same language. With the boost of corresponding development of related benchmarks, many works have been done to improve the performance of various language QA tasks. However, only a few works are dedicated to the G-XLT task. In this work, we propose a generalized cross-lingual transfer framework to enhance the model's ability to understand different languages. Specifically, we first assemble triples from different languages to form multilingual knowledge. Since the lack of knowledge between different languages greatly limits models' reasoning ability, we further design a knowledge injection strategy via leveraging link prediction techniques to enrich the model storage of multilingual knowledge. In this way, we can profoundly exploit rich semantic knowledge. Experiment results on real-world datasets MLQA demonstrate that the proposed method can improve the performance by a large margin, outperforming the baseline method by 13.18%/12.00% F1/EM on average.

cs.CL↗
arXiv2023

Augment and Criticize: Exploring Informative Samples for Semi-Supervised Monocular 3D Object Detection

In this paper, we improve the challenging monocular 3D object detection problem with a general semi-supervised framework. Specifically, having observed that the bottleneck of this task lies in lacking reliable and informative samples to train the detector, we introduce a novel, simple, yet effective `Augment and Criticize' framework that explores abundant informative samples from unlabeled data for learning more robust detection models. In the `Augment' stage, we present the Augmentation-based Prediction aGgregation (APG), which aggregates detections from various automatically learned augmented views to improve the robustness of pseudo label generation. Since not all pseudo labels from APG are beneficially informative, the subsequent `Criticize' phase is presented. In particular, we introduce the Critical Retraining Strategy (CRS) that, unlike simply filtering pseudo labels using a fixed threshold (e.g., classification score) as in 2D semi-supervised tasks, leverages a learnable network to evaluate the contribution of unlabeled images at different training timestamps. This way, the noisy samples prohibitive to model evolution could be effectively suppressed. To validate our framework, we apply it to MonoDLE and MonoFlex. The two new detectors, dubbed 3DSeMo_DLE and 3DSeMo_FLEX, achieve state-of-the-art results with remarkable improvements for over 3.5% AP_3D/BEV (Easy) on KITTI, showing its effectiveness and generality. Code and models will be released.

cs.CV↗
arXiv2023

Towards practical and massively parallel quantum computing emulation for quantum chemistry

Quantum computing is moving beyond its early stage and seeking for commercial applications in chemical and biomedical sciences. In the current noisy intermediate-scale quantum computing era, quantum resource is too scarce to support these explorations. Therefore, it is valuable to emulate quantum computing on classical computers for developing quantum algorithms and validating quantum hardware. However, existing simulators mostly suffer from the memory bottleneck so developing the approaches for large-scale quantum chemistry calculations remains challenging. Here we demonstrate a high-performance and massively parallel variational quantum eigensolver (VQE) simulator based on matrix product states, combined with embedding theory for solving large-scale quantum computing emulation for quantum chemistry on HPC platforms. We apply this method to study the torsional barrier of ethane and the quantification of the protein-ligand interactions. Our largest simulation reaches $1000$ qubits, and a performance of $216.9$ PFLOPS is achieved on a new Sunway supercomputer, which sets the state-of-the-art for quantum computing emulation for quantum chemistry

quant-ph↗
arXiv2023

mmWave Coverage Extension Using Reconfigurable Intelligent Surfaces in Indoor Dense Spaces

In this work, we consider the deployment of reconfigurable intelligent surfaces (RISs) to extend the coverage of a millimeter-wave (mmWave) network in indoor dense spaces. We first integrate RIS into ray-tracing simulations to realistically capture the propagation characteristics, then formulate a non-convex optimization problem that minimizes the number of RISs under rate constraints. We propose a feasible point pursuit and successive convex approximation-based algorithm, which solves the problem by jointly selecting the RIS locations, optimizing the RIS phase-shifts, and allocating time resources to user equipments (UEs). The numerical results demonstrate substantial coverage extension by using at least four RISs, and a data rate of 130 Mbit/s is guaranteed for UEs in the considered area of an airplane cabin.

cs.IT↗
arXiv2023

Not Just Plain Text! Fuel Document-Level Relation Extraction with Explicit Syntax Refinement and Subsentence Modeling

Document-level relation extraction (DocRE) aims to identify semantic labels among entities within a single document. One major challenge of DocRE is to dig decisive details regarding a specific entity pair from long text. However, in many cases, only a fraction of text carries required information, even in the manually labeled supporting evidence. To better capture and exploit instructive information, we propose a novel expLicit syntAx Refinement and Subsentence mOdeliNg based framework (LARSON). By introducing extra syntactic information, LARSON can model subsentences of arbitrary granularity and efficiently screen instructive ones. Moreover, we incorporate refined syntax into text representations which further improves the performance of LARSON. Experimental results on three benchmark datasets (DocRED, CDR, and GDA) demonstrate that LARSON significantly outperforms existing methods.

cs.CL↗
arXiv2023

Quantum Neural Network Inspired Hardware Adaptable Ansatz for Efficient Quantum Simulation of Chemical Systems

The variational quantum eigensolver is a promising way to solve the Schrödinger equation on a noisy intermediate-scale quantum (NISQ) computer, while its success relies on a well-designed wavefunction ansatz. Compared to physically motivated ansatzes, hardware heuristic ansatzes usually lead to a shallower circuit, but it may still be too deep for an NISQ device. Inspired by the quantum neural network, we propose a new hardware heuristic ansatz where the circuit depth can be significantly reduced by introducing ancilla qubits, which makes a practical simulation of a chemical reaction with more than 20 atoms feasible on a currently available quantum computer. More importantly, the expressibility of this new ansatz can be improved by increasing either the depth or the width of the circuit, which makes it adaptable to different hardware environments. These results open a new avenue to develop practical applications of quantum computation in the NISQ era.

quant-ph↗
arXiv2023

Quantum circuit matrix product state ansatz for large-scale simulations of molecules

As in the density matrix renormalization group (DMRG) method, approximating many-body wave function of electrons using a matrix product state (MPS) is a promising way to solve electronic structure problems. The expressibility of an MPS is determined by the size of the matrices or in other words the bond dimension, which unfortunately should be very large in many cases. In this study, we propose to calculate the ground state energies of molecular systems by variationally optimizing quantum circuit MPS (QCMPS) with a relatively small number of qubits. It is demonstrated that with carefully chosen circuit structure and orbital localization scheme, QCMPS can reach a similar accuracy as that achieved in DMRG with an exponentially large bond dimension. QCMPS simulation of a linear molecule with 50 orbitals can reach the chemical accuracy using only 6 qubits at a moderate circuit depth. These results suggest that QCMPS is a promising wave function ansatz in the variational quantum eigensolver algorithm for molecular systems.

quant-ph↗
arXiv2022

BEVDistill: Cross-Modal BEV Distillation for Multi-View 3D Object Detection

3D object detection from multiple image views is a fundamental and challenging task for visual scene understanding. Owing to its low cost and high efficiency, multi-view 3D object detection has demonstrated promising application prospects. However, accurately detecting objects through perspective views is extremely difficult due to the lack of depth information. Current approaches tend to adopt heavy backbones for image encoders, making them inapplicable for real-world deployment. Different from the images, LiDAR points are superior in providing spatial cues, resulting in highly precise localization. In this paper, we explore the incorporation of LiDAR-based detectors for multi-view 3D object detection. Instead of directly training a depth prediction network, we unify the image and LiDAR features in the Bird-Eye-View (BEV) space and adaptively transfer knowledge across non-homogenous representations in a teacher-student paradigm. To this end, we propose \textbf{BEVDistill}, a cross-modal BEV knowledge distillation (KD) framework for multi-view 3D object detection. Extensive experiments demonstrate that the proposed method outperforms current KD approaches on a highly-competitive baseline, BEVFormer, without introducing any extra cost in the inference phase. Notably, our best model achieves 59.4 NDS on the nuScenes test leaderboard, achieving new state-of-the-art in comparison with various image-based detectors. Code will be available at https://github.com/zehuichen123/BEVDistill.

cs.CV↗
arXiv2022

Efficient Single-Image Depth Estimation on Mobile Devices, Mobile AI & AIM 2022 Challenge: Report

Various depth estimation models are now widely used on many mobile and IoT devices for image segmentation, bokeh effect rendering, object tracking and many other mobile tasks. Thus, it is very crucial to have efficient and accurate depth estimation models that can run fast on low-power mobile chipsets. In this Mobile AI challenge, the target was to develop deep learning-based single image depth estimation solutions that can show a real-time performance on IoT platforms and smartphones. For this, the participants used a large-scale RGB-to-depth dataset that was collected with the ZED stereo camera capable to generated depth maps for objects located at up to 50 meters. The runtime of all models was evaluated on the Raspberry Pi 4 platform, where the developed solutions were able to generate VGA resolution depth maps at up to 27 FPS while achieving high fidelity results. All models developed in the challenge are also compatible with any Android or Linux-based mobile devices, their detailed description is provided in this paper.

cs.CV↗
arXiv2022

LiteDepth: Digging into Fast and Accurate Depth Estimation on Mobile Devices

Monocular depth estimation is an essential task in the computer vision community. While tremendous successful methods have obtained excellent results, most of them are computationally expensive and not applicable for real-time on-device inference. In this paper, we aim to address more practical applications of monocular depth estimation, where the solution should consider not only the precision but also the inference time on mobile devices. To this end, we first develop an end-to-end learning-based model with a tiny weight size (1.4MB) and a short inference time (27FPS on Raspberry Pi 4). Then, we propose a simple yet effective data augmentation strategy, called R2 crop, to boost the model performance. Moreover, we observe that the simple lightweight model trained with only one single loss term will suffer from performance bottleneck. To alleviate this issue, we adopt multiple loss terms to provide sufficient constraints during the training stage. Furthermore, with a simple dynamic re-weight strategy, we can avoid the time-consuming hyper-parameter choice of loss terms. Finally, we adopt the structure-aware distillation to further improve the model performance. Notably, our solution named LiteDepth ranks 2nd in the MAI&AIM2022 Monocular Depth Estimation Challenge}, with a si-RMSE of 0.311, an RMSE of 3.79, and the inference time is 37$ms$ tested on the Raspberry Pi 4. Notably, we provide the fastest solution to the challenge. Codes and models will be released at \url{https://github.com/zhyever/LiteDepth}.

cs.CV↗
arXiv2022

Divide-and-conquer variational quantum algorithms for large-scale electronic structure simulations

Exploring the potential application of quantum computers in material design and drug discovery has attracted a lot of interest in the age of quantum computing. However, the quantum resource requirement for solving practical electronic structure problems are far beyond the capacity of near-term quantum devices. In this work, we integrate the divide-and-conquer (DC) approaches into the variational quantum eigensolver (VQE) for large-scale quantum computational chemistry simulations. Two popular divide-and-conquer schemes, including many-body expansion~(MBE) fragmentation theory and density matrix embedding theory~(DMET), are employed to divide complicated problems into many small parts that are easy to implement on near-term quantum computers. Pilot applications of these methods to systems consisting of tens of atoms are performed with adaptive VQE algorithms. This work should encourage further studies of using the philosophy of DC to solve electronic structure problems on quantum computers.

quant-ph↗
arXiv2022

Q$^2$Chemistry: A quantum computation platform for quantum chemistry

Quantum computer provides new opportunities for quantum chemistry. In this article, we present a versatile, extensible, and efficient software package, named Q$^2$Chemistry, for developing quantum algorithms and quantum inspired classical algorithms in the field of quantum chemistry. In Q$^2$Chemistry, wave function and Hamiltonian can be conveniently mapped into the qubit space, then quantum circuits can be generated according to a specific quantum algorithm already implemented in the package or newly developed by the users. The generated circuits can be dispatched to either a physical quantum computer, if available, or to the internal virtual quantum computer realized by simulating quantum circuit on classical supercomputers. As demonstrated by our benchmark simulations with up to 72 qubit, Q$^2$Chemistry achieves excellent performance in simulating medium scale quantum circuits. Application of Q$^2$Chemistry to simulate molecules and periodic systems are given with performance analysis.

quant-ph↗