arXiv subjects

Linfeng Zhang

Publications and source records attributed to Linfeng Zhang.

At least 235 records · Page 13Linked to original sources

arXiv2024

ICT: Image-Object Cross-Level Trusted Intervention for Mitigating Object Hallucination in Large Vision-Language Models

Despite the recent breakthroughs achieved by Large Vision Language Models (LVLMs) in understanding and responding to complex visual-textual contexts, their inherent hallucination tendencies limit their practical application in real-world scenarios that demand high levels of precision. Existing methods typically either fine-tune the LVLMs using additional data, which incurs extra costs in manual annotation and computational resources or perform comparisons at the decoding stage, which may eliminate useful language priors for reasoning while introducing inference time overhead. Therefore, we propose ICT, a lightweight, training-free method that calculates an intervention direction to shift the model's focus towards different levels of visual information, enhancing its attention to high-level and fine-grained visual details. During the forward pass stage, the intervention is applied to the attention heads that encode the overall image information and the fine-grained object details, effectively mitigating the phenomenon of overly language priors, and thereby alleviating hallucinations. Extensive experiments demonstrate that ICT achieves strong performance with a small amount of data and generalizes well across different datasets and models. Our code will be public.

cs.CV↗

arXiv2024

Multi-Stage Vision Token Dropping: Towards Efficient Multimodal Large Language Model

The vision tokens in multimodal large language models usually exhibit significant spatial and temporal redundancy and take up most of the input tokens, which harms their inference efficiency. To solve this problem, some recent works were introduced to drop the unimportant tokens during inference where the importance of each token is decided only by the information in either the vision encoding stage or the prefilling stage. In this paper, we propose Multi-stage Token Dropping (MustDrop) to measure the importance of each token from the whole lifecycle, including the vision encoding stage, prefilling stage, and decoding stage. Concretely, in the visual encoding stage, MustDrop merges spatially adjacent tokens with high similarity, and establishes a key token set to retain the most vision-critical tokens, preventing them from being discarded in later stages. In the prefilling stage, MustDrop further compresses vision tokens by the guidance of text semantics, with a dual-attention filtering strategy. In the decoding stage, an output-aware cache policy is proposed to further reduce the size of the KV cache. By leveraging tailored strategies in the multi-stage process, MustDrop can more precisely recognize the important and redundant tokens, thus achieving an optimal balance between performance and efficiency. For instance, MustDrop reduces about 88.5\% FLOPs on LLaVA with a compression ratio of 92.2\% while maintaining comparable accuracy. Our codes are available at \url{https://github.com/liuting20/MustDrop}.

cs.CV↗

arXiv2024

Deep learning tight-binding approach for large-scale electronic simulations at finite temperatures with $ab$ $initio$ accuracy

Simulating electronic behavior in materials and devices with realistic large system sizes remains a formidable task within the $ab$ $initio$ framework due to its computational intensity. Here we show DeePTB, an efficient deep learning-based tight-binding approach with $ab$ $initio$ accuracy to address this issue. By training on structural data and corresponding $ab$ $initio$ eigenvalues, the DeePTB model can efficiently predict tight-binding Hamiltonians for unseen structures, enabling efficient simulations of large-size systems under external perturbations such as finite temperatures and strain. This capability is vital for semiconductor band gap engineering and materials design. When combined with molecular dynamics, DeePTB facilitates efficient and accurate finite-temperature simulations of both atomic and electronic behavior simultaneously. This is demonstrated by computing the temperature-dependent electronic properties of a gallium phosphide system with $10^6$ atoms. The availability of DeePTB bridges the gap between accuracy and scalability in electronic simulations, potentially advancing materials science and related fields by enabling large-scale electronic structure calculations.

cond-mat.mtrl-sci↗

arXiv2024

SciAssess: Benchmarking LLM Proficiency in Scientific Literature Analysis

Recent breakthroughs in Large Language Models (LLMs) have revolutionized scientific literature analysis. However, existing benchmarks fail to adequately evaluate the proficiency of LLMs in this domain, particularly in scenarios requiring higher-level abilities beyond mere memorization and the handling of multimodal data. In response to this gap, we introduce SciAssess, a benchmark specifically designed for the comprehensive evaluation of LLMs in scientific literature analysis. It aims to thoroughly assess the efficacy of LLMs by evaluating their capabilities in Memorization (L1), Comprehension (L2), and Analysis \& Reasoning (L3). It encompasses a variety of tasks drawn from diverse scientific fields, including biology, chemistry, material, and medicine. To ensure the reliability of SciAssess, rigorous quality control measures have been implemented, ensuring accuracy, anonymization, and compliance with copyright standards. SciAssess evaluates 11 LLMs, highlighting their strengths and areas for improvement. We hope this evaluation supports the ongoing development of LLM applications in scientific literature analysis. SciAssess and its resources are available at \url{https://github.com/sci-assess/SciAssess}.

cs.CL↗

arXiv2024

DiTFastAttn: Attention Compression for Diffusion Transformer Models

Diffusion Transformers (DiT) excel at image and video generation but face computational challenges due to the quadratic complexity of self-attention operators. We propose DiTFastAttn, a post-training compression method to alleviate the computational bottleneck of DiT. We identify three key redundancies in the attention computation during DiT inference: (1) spatial redundancy, where many attention heads focus on local information; (2) temporal redundancy, with high similarity between the attention outputs of neighboring steps; (3) conditional redundancy, where conditional and unconditional inferences exhibit significant similarity. We propose three techniques to reduce these redundancies: (1) Window Attention with Residual Sharing to reduce spatial redundancy; (2) Attention Sharing across Timesteps to exploit the similarity between steps; (3) Attention Sharing across CFG to skip redundant computations during conditional generation. We apply DiTFastAttn to DiT, PixArt-Sigma for image generation tasks, and OpenSora for video generation tasks. Our results show that for image generation, our method reduces up to 76% of the attention FLOPs and achieves up to 1.8x end-to-end speedup at high-resolution (2k x 2k) generation.

cs.CV↗

arXiv2024

REEF: Representation Encoding Fingerprints for Large Language Models

Protecting the intellectual property of open-source Large Language Models (LLMs) is very important, because training LLMs costs extensive computational resources and data. Therefore, model owners and third parties need to identify whether a suspect model is a subsequent development of the victim model. To this end, we propose a training-free REEF to identify the relationship between the suspect and victim models from the perspective of LLMs' feature representations. Specifically, REEF computes and compares the centered kernel alignment similarity between the representations of a suspect model and a victim model on the same samples. This training-free REEF does not impair the model's general capabilities and is robust to sequential fine-tuning, pruning, model merging, and permutations. In this way, REEF provides a simple and effective way for third parties and models' owners to protect LLMs' intellectual property together. The code is available at https://github.com/tmylla/REEF.

cs.CL↗

arXiv2024

Accelerating Diffusion Models with One-to-Many Knowledge Distillation

Significant advancements in image generation have been made with diffusion models. Nevertheless, when contrasted with previous generative models, diffusion models face substantial computational overhead, leading to failure in real-time generation. Recent approaches have aimed to accelerate diffusion models by reducing the number of sampling steps through improved sampling techniques or step distillation. However, the methods to diminish the computational cost for each timestep remain a relatively unexplored area. Observing the fact that diffusion models exhibit varying input distributions and feature distributions at different timesteps, we introduce one-to-many knowledge distillation (O2MKD), which distills a single teacher diffusion model into multiple student diffusion models, where each student diffusion model is trained to learn the teacher's knowledge for a subset of continuous timesteps. Experiments on CIFAR10, LSUN Church, CelebA-HQ with DDPM and COCO30K with Stable Diffusion show that O2MKD can be applied to previous knowledge distillation and fast sampling methods to achieve significant acceleration. Codes will be released in Github.

cs.CV↗

arXiv2024

A Comprehensive Framework for Estimating Aircraft Fuel Consumption Based on Flight Trajectories

Accurate calculation of aircraft fuel consumption plays an irreplaceable role in flight operations, optimization, and pollutant accounting. Calculating aircraft fuel consumption accurately is tricky because it changes based on different flying conditions and physical factors. Utilizing flight surveillance data, this study developed a comprehensive mathematical framework and established a link between flight dynamics and fuel consumption, providing a set of high-precision, high-resolution fuel calculation methods. It also allows other practitioners to select data sources according to specific needs through this framework. The methodology begins by addressing the functional aspects of interval fuel consumption. We apply spectral transformation techniques to mine Automatic Dependent Surveillance-Broadcast (ADS-B) data, identifying key aspects of the flight profile and establishing their theoretical relationships with fuel consumption. Subsequently, a deep neural network with tunable parameters is used to fit this multivariate function, facilitating high-precision calculations of interval fuel consumption. Furthermore, a second-order smooth monotonic interpolation method was constructed along with a novel estimation method for instantaneous fuel consumption. Numerical results have validated the effectiveness of the model. Using ADS-B and Aircraft Communications Addressing and Reporting System (ACARS) data from 2023 for testing, the average error of interval fuel consumption can be reduced to as low as $3.31\%$, and the error in the integral sense of instantaneous fuel consumption is $8.86\%$. These results establish this model as the state of the art, achieving the lowest estimation errors in aircraft fuel consumption calculations to date.

stat.AP↗

arXiv2024

NLP-Powered Repository and Search Engine for Academic Papers: A Case Study on Cyber Risk Literature with CyLit

As the body of academic literature continues to grow, researchers face increasing difficulties in effectively searching for relevant resources. Existing databases and search engines often fall short of providing a comprehensive and contextually relevant collection of academic literature. To address this issue, we propose a novel framework that leverages Natural Language Processing (NLP) techniques. This framework automates the retrieval, summarization, and clustering of academic literature within a specific research domain. To demonstrate the effectiveness of our approach, we introduce CyLit, an NLP-powered repository specifically designed for the cyber risk literature. CyLit empowers researchers by providing access to context-specific resources and enabling the tracking of trends in the dynamic and rapidly evolving field of cyber risk. Through the automatic processing of large volumes of data, our NLP-powered solution significantly enhances the efficiency and specificity of academic literature searches. We compare the literature categorization results of CyLit to those presented in survey papers or generated by ChatGPT, highlighting the distinctive insights this tool provides into cyber risk research literature. Using NLP techniques, we aim to revolutionize the way researchers discover, analyze, and utilize academic resources, ultimately fostering advancements in various domains of knowledge.

cs.IR↗

arXiv2024

Revealing the molecular structures of a-Al2O3(0001)-water interface by machine learning based computational vibrational spectroscopy

Solid-water interfaces are crucial to many physical and chemical processes and are extensively studied using surface-specific sum-frequency generation (SFG) spectroscopy. To establish clear correlations between specific spectral signatures and distinct interfacial water structures, theoretical calculations using molecular dynamics (MD) simulations are required. These MD simulations typically need relatively long trajectories (a few nanoseconds) to achieve reliable SFG response function calculations via the dipole-polarizability time correlation function. However, the requirement for long trajectories limits the use of computationally expensive techniques such as ab initio MD (AIMD) simulations, particularly for complex solid-water interfaces. In this work, we present a pathway for calculating vibrational spectra (IR, Raman, SFG) of solid-water interfaces using machine learning (ML)-accelerated methods. We employ both the dipole moment-polarizability correlation function and the surface-specific velocity-velocity correlation function approaches to calculate SFG spectra. Our results demonstrate the successful acceleration of AIMD simulations and the calculation of SFG spectra using ML methods. This advancement provides an opportunity to calculate SFG spectra for the complicated solid-water systems more rapidly and at a lower computational cost with the aid of ML.

cond-mat.mtrl-sci↗

arXiv2024

Towards a Unified Benchmark and Framework for Deep Learning-Based Prediction of Nuclear Magnetic Resonance Chemical Shifts

The study of structure-spectrum relationships is essential for spectral interpretation, impacting structural elucidation and material design. Predicting spectra from molecular structures is challenging due to their complex relationships. Herein, we introduce NMRNet, a deep learning framework using the SE(3) Transformer for atomic environment modeling, following a pre-training and fine-tuning paradigm. To support the evaluation of NMR chemical shift prediction models, we have established a comprehensive benchmark based on previous research and databases, covering diverse chemical systems. Applying NMRNet to these benchmark datasets, we achieve state-of-the-art performance in both liquid-state and solid-state NMR datasets, demonstrating its robustness and practical utility in real-world scenarios. This marks the first integration of solid and liquid state NMR within a unified model architecture, highlighting the need for domainspecific handling of different atomic environments. Our work sets a new standard for NMR prediction, advancing deep learning applications in analytical and structural chemistry.

physics.comp-ph↗

arXiv2024

RiD-kit: Software package designed to do enhanced sampling using reinforced dynamics

Developing an efficient method to accelerate the speed of molecular dynamics is a central theme in the field of molecular simulation. One category among the methods are collective-variable-based methods, which rely on predefined collective variables (CVs). The difficulty of selecting a few important CVs hinders the methods to be applied to large systems easily. Here we present a CV-based enhanced sampling method RiD-kit, which could handle a large number of CVs and perform efficient sampling. The method could be applied to various kinds of systems, including biomolecules, chemical reactions and materials. In this protocol, we guide the users through all phases of the RiD-kit workflow, from preparing the input files, setting the simulation parameters and analyzing the results. The RiD-kit workflow provides an efficient and user-friendly command line tool which could submit jobs to various kinds of platforms including the high-performance computers (HPC), cloud server and local machines.

physics.chem-ph↗

arXiv2024

DPA-2: a large atomic model as a multi-task learner

The rapid advancements in artificial intelligence (AI) are catalyzing transformative changes in atomic modeling, simulation, and design. AI-driven potential energy models have demonstrated the capability to conduct large-scale, long-duration simulations with the accuracy of ab initio electronic structure methods. However, the model generation process remains a bottleneck for large-scale applications. We propose a shift towards a model-centric ecosystem, wherein a large atomic model (LAM), pre-trained across multiple disciplines, can be efficiently fine-tuned and distilled for various downstream tasks, thereby establishing a new framework for molecular modeling. In this study, we introduce the DPA-2 architecture as a prototype for LAMs. Pre-trained on a diverse array of chemical and materials systems using a multi-task approach, DPA-2 demonstrates superior generalization capabilities across multiple downstream tasks compared to the traditional single-task pre-training and fine-tuning methodologies. Our approach sets the stage for the development and broad application of LAMs in molecular and materials simulation research.

physics.chem-ph↗

arXiv2024

A Pre-trained Deep Potential Model for Sulfide Solid Electrolytes with Broad Coverage and High Accuracy

Solid electrolytes with fast ion transport are one of the key challenges for solid state lithium metal batteries. To improve ion conductivity, chemical doping has been the most effective strategy, and atomistic simulation with machine-learning potential helps find optimized doping by predicting ion conductivity for arbitrary composition. Yet most existing machine-learning models are trained on narrow chemistry, and new model has to be trained for each system, wasting transferable knowledge and incurring significant cost. Here, we propose a pre-trained deep potential model purpose-built for sulfide electrolytes with attention mechanism, known as DPA-SSE. The training set encompasses 15 elements and consists of both equilibrium and extensive out-of-equilibrium configurations. DPA-SSE achieves a high energy resolution of less than 2 meV/atom for dynamical trajectories up to 1150 K, and reproduces experimental ion conductivity of sulfide electrolytes with remarkable accuracy. DPA-SSE exhibits good transferability, covering a range of complex electrolytes with mixes of cation and anion atoms. Highly efficient dynamical simulation with DPA-SSE can be realized by model distillation which generates a faster model for given systems. DPA-SSE also serves as a platform for continuous learning, and the model fine-tune requires only a portion of downstream data. These results demonstrate the possibility of a new pathway for AI-driven development of solid electrolytes with exceptional performance.

physics.chem-ph↗

arXiv2024

Uni-ELF: A Multi-Level Representation Learning Framework for Electrolyte Formulation Design

Advancements in lithium battery technology heavily rely on the design and engineering of electrolytes. However, current schemes for molecular design and recipe optimization of electrolytes lack an effective computational-experimental closed loop and often fall short in accurately predicting diverse electrolyte formulation properties. In this work, we introduce Uni-ELF, a novel multi-level representation learning framework to advance electrolyte design. Our approach involves two-stage pretraining: reconstructing three-dimensional molecular structures at the molecular level using the Uni-Mol model, and predicting statistical structural properties (e.g., radial distribution functions) from molecular dynamics simulations at the mixture level. Through this comprehensive pretraining, Uni-ELF is able to capture intricate molecular and mixture-level information, which significantly enhances its predictive capability. As a result, Uni-ELF substantially outperforms state-of-the-art methods in predicting both molecular properties (e.g., melting point, boiling point, synthesizability) and formulation properties (e.g., conductivity, Coulombic efficiency). Moreover, Uni-ELF can be seamlessly integrated into an automatic experimental design workflow. We believe this innovative framework will pave the way for automated AI-based electrolyte design and engineering.

physics.chem-ph↗

arXiv2024

Uni-Mol2: Exploring Molecular Pretraining Model at Scale

In recent years, pretraining models have made significant advancements in the fields of natural language processing (NLP), computer vision (CV), and life sciences. The significant advancements in NLP and CV are predominantly driven by the expansion of model parameters and data size, a phenomenon now recognized as the scaling laws. However, research exploring scaling law in molecular pretraining models remains unexplored. In this work, we present Uni-Mol2 , an innovative molecular pretraining model that leverages a two-track transformer to effectively integrate features at the atomic level, graph level, and geometry structure level. Along with this, we systematically investigate the scaling law within molecular pretraining models, characterizing the power-law correlations between validation loss and model size, dataset size, and computational resources. Consequently, we successfully scale Uni-Mol2 to 1.1 billion parameters through pretraining on 800 million conformations, making it the largest molecular pretraining model to date. Extensive experiments show consistent improvement in the downstream tasks as the model size grows. The Uni-Mol2 with 1.1B parameters also outperforms existing methods, achieving an average 27% improvement on the QM9 and 14% on COMPAS-1D dataset.

cs.LG↗

arXiv2024

Uni-SMART: Universal Science Multimodal Analysis and Research Transformer

In scientific research and its application, scientific literature analysis is crucial as it allows researchers to build on the work of others. However, the fast growth of scientific knowledge has led to a massive increase in scholarly articles, making in-depth literature analysis increasingly challenging and time-consuming. The emergence of Large Language Models (LLMs) has offered a new way to address this challenge. Known for their strong abilities in summarizing texts, LLMs are seen as a potential tool to improve the analysis of scientific literature. However, existing LLMs have their own limits. Scientific literature often includes a wide range of multimodal elements, such as tables, charts, and molecule, which are hard for text-focused LLMs to understand and analyze. This issue points to the urgent need for new solutions that can fully understand and analyze multimodal content in scientific literature. To answer this demand, we present \textbf{Uni-SMART} (Universal Science Multimodal Analysis and Research Transformer), an innovative model designed for in-depth understanding of multimodal scientific literature. Through rigorous quantitative evaluation across several domains, Uni-SMART demonstrates superior performance over other text-focused LLMs. Furthermore, our exploration extends to practical applications, including patent infringement detection and nuanced analysis of charts. These applications not only highlight Uni-SMART's adaptability but also its potential to revolutionize how we interact with scientific literature.

cs.CL↗

arXiv2024

Efficient and Precise Force Field Optimization for Biomolecules Using DPA-2

Molecular simulations are essential tools in computational chemistry, enabling the prediction and understanding of molecular interactions and thermodynamic properties of biomolecules. However, traditional force fields face significant challenges in accurately representing novel molecules and complex chemical environments due to the labor-intensive process of manually setting optimization parameters and the high computational cost of quantum mechanical calculations. To overcome these difficulties, we fine-tuned a high-accuracy DPA-2 pre-trained model and applied it to optimize force field parameters on-the-fly, significantly reducing computational costs. Our method combines this fine-tuned DPA-2 model with a node-embedding-based similarity metric, allowing seamless augmentation to new chemical species without manual intervention. We applied this process to the TYK2 inhibitor and PTP1B systems and demonstrated its effectiveness through the improvement of free energy perturbation calculation results. This advancement contributes valuable insights and tools for the computational chemistry community.

physics.chem-ph↗