arXiv subjects

Hao Zhang

Publications and source records attributed to Hao Zhang.

At least 649 records · Page 36Linked to original sources

arXiv2024

Preference Alignment Improves Language Model-Based TTS

Recent advancements in text-to-speech (TTS) have shown that language model (LM)-based systems offer competitive performance to their counterparts. Further optimization can be achieved through preference alignment algorithms, which adjust LMs to align with the preferences of reward models, enhancing the desirability of the generated content. This study presents a thorough empirical evaluation of how preference alignment algorithms, particularly Direct Preference Optimization (DPO), enhance LM-based TTS. With a 1.15B parameter LM-based TTS model, we demonstrate that preference alignment consistently improves intelligibility, speaker similarity, and proxy subjective evaluation scores, with the latter two metrics surpassing even human speech in certain evaluations. We also show preference alignment is applicable to low-resource scenarios and effectively generalized to out-of-domain applications.

cs.CL↗

arXiv2024

Explaining Generalization Power of a DNN Using Interactive Concepts

This paper explains the generalization power of a deep neural network (DNN) from the perspective of interactions. Although there is no universally accepted definition of the concepts encoded by a DNN, the sparsity of interactions in a DNN has been proved, i.e., the output score of a DNN can be well explained by a small number of interactions between input variables. In this way, to some extent, we can consider such interactions as interactive concepts encoded by the DNN. Therefore, in this paper, we derive an analytic explanation of inconsistency of concepts of different complexities. This may shed new lights on using the generalization power of concepts to explain the generalization power of the entire DNN. Besides, we discover that the DNN with stronger generalization power usually learns simple concepts more quickly and encodes fewer complex concepts. We also discover the detouring dynamics of learning complex concepts, which explains both the high learning difficulty and the low generalization power of complex concepts. The code will be released when the paper is accepted.

cs.LG↗

arXiv2024

Approximately counting maximal independent set is equivalent to #SAT

A maximal independent set is an independent set that is not a subset of any other independent set. It is also the key problem of mathematics, computer science, and other fields. A counting problem is a type of computational problem that associated with the number of solutions. Besides, counting problems help us better understand several fields such as algorithm analysis, complexity theory, artificial intelligence, etc. The problem of counting maximal independent sets is #P-complete. So it is natural to think about approximate counting for maximal independent sets problem. In this article, we study the complexity of approximately counting maximal independent sets. Specifically, we are the first to prove that the #MIS problem is AP-interreducible with the #SAT of a given general graph.

cs.CC↗

arXiv2024

DECOLLAGE: 3D Detailization by Controllable, Localized, and Learned Geometry Enhancement

We present a 3D modeling method which enables end-users to refine or detailize 3D shapes using machine learning, expanding the capabilities of AI-assisted 3D content creation. Given a coarse voxel shape (e.g., one produced with a simple box extrusion tool or via generative modeling), a user can directly "paint" desired target styles representing compelling geometric details, from input exemplar shapes, over different regions of the coarse shape. These regions are then up-sampled into high-resolution geometries which adhere with the painted styles. To achieve such controllable and localized 3D detailization, we build on top of a Pyramid GAN by making it masking-aware. We devise novel structural losses and priors to ensure that our method preserves both desired coarse structures and fine-grained features even if the painted styles are borrowed from diverse sources, e.g., different semantic parts and even different shape categories. Through extensive experiments, we show that our ability to localize details enables novel interactive creative workflows and applications. Our experiments further demonstrate that in comparison to prior techniques built on global detailization, our method generates structure-preserving, high-resolution stylized geometries with more coherent shape details and style transitions.

cs.CV↗

arXiv2024

VQCrystal: Leveraging Vector Quantization for Discovery of Stable Crystal Structures

Discovering functional crystalline materials through computational methods remains a formidable challenge in materials science. Here, we introduce VQCrystal, an innovative deep learning framework that leverages discrete latent representations to overcome key limitations in current approaches to crystal generation and inverse design. VQCrystal employs a hierarchical VQ-VAE architecture to encode global and atom-level crystal features, coupled with a machine learning-based inter-atomic potential(IAP) model and a genetic algorithm to realize property-targeted inverse design. Benchmark evaluations on diverse datasets demonstrate VQCrystal's advanced capabilities in representation learning and novel crystal discovery. Notably, VQCrystal achieves state-of-the-art performance with 91.93\% force validity and a Fréchet Distance of 0.152 on MP-20, indicating both strong validity and high diversity in the sampling process. To demonstrate real-world applicability, we apply VQCrystal for both 3D and 2D material design. For 3D materials, the density-functional theory validation confirmed that 63.04\% of bandgaps and 99\% of formation energies of the 56 filtered materials matched the target range. Moreover, 437 generated materials were validated as existing entries in the full database outside the training set. For the discovery of 2D materials, 73.91\% of 23 filtered structures exhibited high stability with formation energies below -1 eV/atom. Our results highlight VQCrystal's potential to accelerate the discovery of novel materials with tailored properties.

cond-mat.mtrl-sci↗

arXiv2024

Combining physics-informed graph neural network and finite difference for solving forward and inverse spatiotemporal PDEs

The great success of Physics-Informed Neural Networks (PINN) in solving partial differential equations (PDEs) has significantly advanced our simulation and understanding of complex physical systems in science and engineering. However, many PINN-like methods are poorly scalable and are limited to in-sample scenarios. To address these challenges, this work proposes a novel discrete approach termed Physics-Informed Graph Neural Network (PIGNN) to solve forward and inverse nonlinear PDEs. In particular, our approach seamlessly integrates the strength of graph neural networks (GNN), physical equations and finite difference to approximate solutions of physical systems. Our approach is compared with the PINN baseline on three well-known nonlinear PDEs (heat, Burgers and FitzHugh-Nagumo). We demonstrate the excellent performance of the proposed method to work with irregular meshes, longer time steps, arbitrary spatial resolutions, varying initial conditions (ICs) and boundary conditions (BCs) by conducting extensive numerical experiments. Numerical results also illustrate the superiority of our approach in terms of accuracy, time extrapolability, generalizability and scalability. The main advantage of our approach is that models trained in small domains with simple settings have excellent fitting capabilities and can be directly applied to more complex situations in large domains.

math.NA↗

arXiv2024

Scalable tensor network algorithm for thermal quantum many-body systems in two dimension

Simulating strongly-correlated quantum many-body systems at finite temperatures is a significant challenge in computational physics. In this work, we present a scalable finite-temperature tensor network algorithm for two-dimensional quantum many-body systems. We employ the (fermionic) projected entangled pair state (PEPS) to represent the vectorization of the quantum thermal state and utilize a stochastic reconfiguration method to cool down the quantum states from infinite temperature. We validate our method by benchmarking it against the 2D antiferromagnetic Heisenberg model, the $J_1$-$J_2$ model, and the Fermi-Hubbard model, comparing physical properties such as internal energy, specific heat, and magnetic susceptibility with results obtained from stochastic series expansion (SSE), exact diagonalization, and determinant quantum Monte Carlo (DQMC).

cond-mat.str-el↗

arXiv2024

Broad-line Region of the Quasar PG 2130+099. II. Doubling the Size Over Four Years?

Over the past three decades, multiple reverberation mapping (RM) campaigns conducted for the quasar PG 2130+099 have exhibited inconsistent findings with time delays ranging from $\sim$10 to $\sim$200 days. To achieve a comprehensive understanding of the geometry and dynamics of the broad-line region (BLR) in PG 2130+099, we continued an ongoing high-cadence RM monitoring campaign using the Calar Alto Observatory 2.2m optical telescope for an extra four years from 2019 to 2022. We measured the time lags of several broad emission lines (including He II, He I, H$β$, and Fe II) with respect to the 5100 Å continuum, and their time lags continuously vary through the years. Especially, the H$β$ time lags exhibited approximately a factor of two increase in the last two years. Additionally, the velocity-resolved time delays of the broad H$β$ emission line reveal a back-and-forth change between signs of virial motion and inflow in the BLR. The combination of negligible ($\sim$10%) continuum change and substantial time-lag variation (over two times) results in significant scatter in the intrinsic $R_{\rm Hβ}-L_{\rm 5100}$ relationship for PG 2130+099. Taking into account the consistent changes in the continuum variability time scale and the size of the BLR, we tentatively propose that the changes in the measurement of the BLR size may be affected by 'geometric dilution'.

astro-ph.GA↗

arXiv2024

Efficient LLM Scheduling by Learning to Rank

In Large Language Model (LLM) inference, the output length of an LLM request is typically regarded as not known a priori. Consequently, most LLM serving systems employ a simple First-come-first-serve (FCFS) scheduling strategy, leading to Head-Of-Line (HOL) blocking and reduced throughput and service quality. In this paper, we reexamine this assumption -- we show that, although predicting the exact generation length of each request is infeasible, it is possible to predict the relative ranks of output lengths in a batch of requests, using learning to rank. The ranking information offers valuable guidance for scheduling requests. Building on this insight, we develop a novel scheduler for LLM inference and serving that can approximate the shortest-job-first (SJF) schedule better than existing approaches. We integrate this scheduler with the state-of-the-art LLM serving system and show significant performance improvement in several important applications: 2.8x lower latency in chatbot serving and 6.5x higher throughput in synthetic data generation. Our code is available at https://github.com/hao-ai-lab/vllm-ltr.git

cs.LG↗

arXiv2024

SiCP: Simultaneous Individual and Cooperative Perception for 3D Object Detection in Connected and Automated Vehicles

Cooperative perception for connected and automated vehicles is traditionally achieved through the fusion of feature maps from two or more vehicles. However, the absence of feature maps shared from other vehicles can lead to a significant decline in 3D object detection performance for cooperative perception models compared to standalone 3D detection models. This drawback impedes the adoption of cooperative perception as vehicle resources are often insufficient to concurrently employ two perception models. To tackle this issue, we present Simultaneous Individual and Cooperative Perception (SiCP), a generic framework that supports a wide range of the state-of-the-art standalone perception backbones and enhances them with a novel Dual-Perception Network (DP-Net) designed to facilitate both individual and cooperative perception. In addition to its lightweight nature with only 0.13M parameters, DP-Net is robust and retains crucial gradient information during feature map fusion. As demonstrated in a comprehensive evaluation on the V2V4Real and OPV2V datasets, thanks to DP-Net, SiCP surpasses state-of-the-art cooperative perception solutions while preserving the performance of standalone perception solutions.

cs.CV↗

arXiv2024

Through-the-Wall Radar Human Activity Micro-Doppler Signature Representation Method Based on Joint Boulic-Sinusoidal Pendulum Model

With the help of micro-Doppler signature, ultra-wideband (UWB) through-the-wall radar (TWR) enables the reconstruction of range and velocity information of limb nodes to accurately identify indoor human activities. However, existing methods are usually trained and validated directly using range-time maps (RTM) and Doppler-time maps (DTM), which have high feature redundancy and poor generalization ability. In order to solve this problem, this paper proposes a human activity micro-Doppler signature representation method based on joint Boulic-sinusoidal pendulum motion model. In detail, this paper presents a simplified joint Boulic-sinusoidal pendulum human motion model by taking head, torso, both hands and feet into consideration improved from Boulic-Thalmann kinematic model. The paper also calculates the minimum number of key points needed to describe the Doppler and micro-Doppler information sufficiently. Both numerical simulations and experiments are conducted to verify the effectiveness. The results demonstrate that the proposed number of key points of micro-Doppler signature can precisely represent the indoor human limb node motion characteristics, and substantially improve the generalization capability of the existing methods for different testers.

eess.SP↗

arXiv2024

High-areal-capacity Na-ion battery electrode with uncompromised energy and power densities by simultaneous electrospinning-spraying fabrication

Sodium-ion batteries (SIBs) are cost-effective alternatives to lithium-ion batteries (LIBs), but their low energy density remains a challenge. Current electrode designs fail to simultaneously achieve high areal loading, high active content, and superior performance. In response, this work introduces an ideal electrode structure, featuring a continuous conductive network with active particles securely trapped in the absence of binder, fabricated using a universal technique that combines electrospinning and electrospraying (co-ESP). We found that the particle size must be larger than the network's pores for optimised performance, an aspect overlooked in previous research. The free-standing co-ESP Na2V3(PO4)3 (NVP) cathodes demonstrated state-of-the-art 296 mg cm-2 areal loading with 97.5 wt.% active content, as well as remarkable rate-performance and cycling stability. Co-ESP full cells showed uncompromised energy and power densities (231.6 Wh kg-1 and 7152.6 W kg-1), leading among reported SIBs with industry-relevant areal loadings. The structural merit is analysed using multi-scale X-ray computed tomography, providing valuable design insights.Finally, the superior performance is validated in the pouch cells, highlighting the electrode's scalability and potential for commercial application.

cond-mat.mtrl-sci↗

arXiv2024

On Optimizing the Communication of Model Parallelism

We study a novel and important communication pattern in large-scale model-parallel deep learning (DL), which we call cross-mesh resharding. This pattern emerges when the two paradigms of model parallelism - intra-operator and inter-operator parallelism - are combined to support large models on large clusters. In cross-mesh resharding, a sharded tensor needs to be sent from a source device mesh to a destination device mesh, on which the tensor may be distributed with the same or different layouts. We formalize this as a many-to-many multicast communication problem, and show that existing approaches either are sub-optimal or do not generalize to different network topologies or tensor layouts, which result from different model architectures and parallelism strategies. We then propose two contributions to address cross-mesh resharding: an efficient broadcast-based communication system, and an "overlapping-friendly" pipeline schedule. On microbenchmarks, our overall system outperforms existing ones by up to 10x across various tensor and mesh layouts. On end-to-end training of two large models, GPT-3 and U-Transformer, we improve throughput by 10% and 50%, respectively.

cs.LG↗

arXiv2024

Collaborative Cross-modal Fusion with Large Language Model for Recommendation

Despite the success of conventional collaborative filtering (CF) approaches for recommendation systems, they exhibit limitations in leveraging semantic knowledge within the textual attributes of users and items. Recent focus on the application of large language models for recommendation (LLM4Rec) has highlighted their capability for effective semantic knowledge capture. However, these methods often overlook the collaborative signals in user behaviors. Some simply instruct-tune a language model, while others directly inject the embeddings of a CF-based model, lacking a synergistic fusion of different modalities. To address these issues, we propose a framework of Collaborative Cross-modal Fusion with Large Language Models, termed CCF-LLM, for recommendation. In this framework, we translate the user-item interactions into a hybrid prompt to encode both semantic knowledge and collaborative signals, and then employ an attentive cross-modal fusion strategy to effectively fuse latent embeddings of both modalities. Extensive experiments demonstrate that CCF-LLM outperforms existing methods by effectively utilizing semantic and collaborative signals in the LLM4Rec context.

cs.IR↗

arXiv2024

Hard superconducting gap in InSb nanowires

Topological superconductivity is a state of matter that can host Majorana modes, the building blocks of a topological quantum computer. Many experimental platforms predicted to show such a topological state rely on proximity-induced superconductivity. However, accessing the topological properties requires an induced hard superconducting gap, which is challenging to achieve for most material systems. We have systematically studied how the interface between an InSb semiconductor nanowire and a NbTiN superconductor affects the induced superconducting properties. Step by step, we improve the homogeneity of the interface while ensuring a barrier-free electrical contact to the superconductor, and obtain a hard gap in the InSb nanowire. The magnetic field stability of NbTiN allows the InSb nanowire to maintain a hard gap and a supercurrent in the presence of magnetic fields (~ 0.5 Tesla), a requirement for topological superconductivity in one-dimensional systems. Our study provides a guideline to induce superconductivity in various experimental platforms such as semiconductor nanowires, two dimensional electron gases and topological insulators, and holds relevance for topological superconductivity and quantum computation.

cond-mat.mes-hall↗

arXiv2024

HETHUB: A Distributed Training System with Heterogeneous Cluster for Large-Scale Models

Training large-scale models relies on a vast number of computing resources. For example, training the GPT-4 model (1.8 trillion parameters) requires 25000 A100 GPUs . It is a challenge to build a large-scale cluster with one type of GPU-accelerator. Using multiple types of GPU-accelerators to construct a large-scale cluster is an effective way to solve the problem of insufficient homogeneous GPU-accelerators. However, the existing distributed training systems for large-scale models only support homogeneous GPU-accelerators, not support heterogeneous GPU-accelerators. To address the problem, this paper proposes a distributed training system with hybrid parallelism, HETHUB, for large-scale models, which supports heterogeneous cluster, including AMD, Nvidia GPU and other types of GPU-accelerators . It introduces a distributed unified communicator to realize the communication between heterogeneous GPU-accelerators, a distributed performance predictor, and an automatic parallel planner to develop and train models efficiently with heterogeneous GPU-accelerators. Compared to the distributed training system with homogeneous GPU-accelerators, our system can support six combinations of heterogeneous GPU-accelerators. We train the Llama-140B model on a heterogeneous cluster with 768 GPU-accelerators(128 AMD and 640 GPU-accelerator A). The experiment results show that the optimal performance of our system in the heterogeneous cluster has achieved up to 97.49% of the theoretical upper bound performance.

cs.DC↗

arXiv2024

DPA-Net: Structured 3D Abstraction from Sparse Views via Differentiable Primitive Assembly

We present a differentiable rendering framework to learn structured 3D abstractions in the form of primitive assemblies from sparse RGB images capturing a 3D object. By leveraging differentiable volume rendering, our method does not require 3D supervision. Architecturally, our network follows the general pipeline of an image-conditioned neural radiance field (NeRF) exemplified by pixelNeRF for color prediction. As our core contribution, we introduce differential primitive assembly (DPA) into NeRF to output a 3D occupancy field in place of density prediction, where the predicted occupancies serve as opacity values for volume rendering. Our network, coined DPA-Net, produces a union of convexes, each as an intersection of convex quadric primitives, to approximate the target 3D object, subject to an abstraction loss and a masking loss, both defined in the image space upon volume rendering. With test-time adaptation and additional sampling and loss designs aimed at improving the accuracy and compactness of the obtained assemblies, our method demonstrates superior performance over state-of-the-art alternatives for 3D primitive abstraction from sparse views.

cs.CV↗

arXiv2024

MPC-Minimized Secure LLM Inference

Many inference services based on large language models (LLMs) pose a privacy concern, either revealing user prompts to the service or the proprietary weights to the user. Secure inference offers a solution to this problem through secure multi-party computation (MPC), however, it is still impractical for modern LLM workload due to the large overhead imposed by MPC. To address this overhead, we propose Marill, a framework that adapts LLM fine-tuning to minimize MPC usage during secure inference. Marill introduces high-level architectural changes during fine-tuning that significantly reduce the number of expensive operations needed within MPC during inference, by removing some and relocating others outside MPC without compromising security. As a result, Marill-generated models are more efficient across all secure inference protocols and our approach complements MPC-friendly approximations for such operations. Compared to standard fine-tuning, Marill results in 3.6-11.3x better runtime and 2.4-6.9x better communication during secure inference across various MPC settings, while typically preserving over 90% performance across downstream tasks.

cs.CR↗