Search

A comprehensive comparison of quantum evolution between the quantum and classical mechanically motion of nuclei in a finite-dimensional quantum chemistry model is presented. A modified version of Tavis-Cummings-Hubbard model with two two-level artificial atoms in optical cavities is described for simulating the association and dissociation of neutral hydrogen molecule. The initial circumstances that led to the formation and decomposition of neutral hydrogen molecule are discussed. The dissipative process of Markovian open system is simulated through solving quantum master equation - Lindbladian. The motion of these two atoms (nuclei) both quantum and classical mechanically is compared. In quantum form, nuclei's mobility is portrayed by tunneling effect of nuclei. And we describe the effect of the classical motion of nuclei on the interaction within the system by fluctuation of strengths. Consideration is also given to the dark state, which is produced along with the dissociation process and has a non-negligible impact on the final result of the evolution., Comment: 12 pages, 8 figures

Entropic dynamics of a multiqubit cavity quantum electrodynamics system is simulated and various aspects of entropy are explored. In the modified version of the Tavis-Cummings-Hubbard model, atoms are held in optical cavities through optical tweezers and can jump between different cavities through the tunneling effect. The interaction of atom with the cavity results in different electronic transitions and the creation and annihilation of corresponding types of photon. Electron spin and the Pauli exclusion principle are considered. Formation and break of covalent bond and creation and annihilation of phonon are also introduced into the model. The system is bipartite. The effect of all kinds of interactions on entropy is studied. And the von Neumann entropy of different subsystems is compared. The results show that the entropic dynamics can be controlled by selectively choosing system parameters, and the entropy values of different subsystems satisfy certain inequality relationships., Comment: 16 pages, 14 figures, 2 tables

A general scheme is given for supercomputer simulation of quantum processes, which are described by various modifications of finite-dimensional cavity quantum electrodynamics models, including Jaynes-Cummings-Hubbard model and Tavis-Cummings-Hubbard model. Conclusions and recommendations are illustrated using two examples: approximate model of hydrogen bonding and model of photon motion on a two-dimensional plane., Comment: 8 pages, 8 figures, 3 tables

A distributed computing approach to solve the curse of dimensionality, caused by the complex quantum system modeling, is discussed. With the help of Cannon's algorithm, the distributed computing transformation of numerical method for simulating quantum unitary evolution is achieved. Based on the Tavis-Cummings model, a large number of atoms are added into the optical cavity to obtain a high-dimensional quantum closed system, implemented on the supercomputer platform. The comparison of time cost and speedup of different distributed computing strategies is discussed., Comment: 8 pages, 6 figures, 4 tables

The recent advance in vision-language models is largely attributed to the abundance of image-text data. We aim to replicate this success for video-language models, but there simply is not enough human-curated video-text data available. We thus resort to fine-tuning a video-language model from a strong image-language baseline with synthesized instructional data. The resulting video model by video-instruction-tuning (VIIT) is then used to auto-label millions of videos to generate high-quality captions. We show the adapted video-language model performs well on a wide range of video-language benchmarks. For instance, it surpasses the best prior result on open-ended NExT-QA by 2.8%. Besides, our model generates detailed descriptions for previously unseen videos, which provide better textual supervision than existing methods. Experiments show that a video-language dual-encoder model contrastively trained on these auto-generated captions is 3.8% better than the strongest baseline that also leverages vision-language models. Our best model outperforms state-of-the-art methods on MSR-VTT zero-shot text-to-video retrieval by 6%. As a side product, we generate the largest video caption dataset to date., Comment: CVPR 2024. Project page: https://zhaoyue-zephyrus.github.io/video-instruction-tuning

We present VisionFM, a foundation model pre-trained with 3.4 million ophthalmic images from 560,457 individuals, covering a broad range of ophthalmic diseases, modalities, imaging devices, and demography. After pre-training, VisionFM provides a foundation to foster multiple ophthalmic artificial intelligence (AI) applications, such as disease screening and diagnosis, disease prognosis, subclassification of disease phenotype, and systemic biomarker and disease prediction, with each application enhanced with expert-level intelligence and accuracy. The generalist intelligence of VisionFM outperformed ophthalmologists with basic and intermediate levels in jointly diagnosing 12 common ophthalmic diseases. Evaluated on a new large-scale ophthalmic disease diagnosis benchmark database, as well as a new large-scale segmentation and detection benchmark database, VisionFM outperformed strong baseline deep neural networks. The ophthalmic image representations learned by VisionFM exhibited noteworthy explainability, and demonstrated strong generalizability to new ophthalmic modalities, disease spectrum, and imaging devices. As a foundation model, VisionFM has a large capacity to learn from diverse ophthalmic imaging data and disparate datasets. To be commensurate with this capacity, in addition to the real data used for pre-training, we also generated and leveraged synthetic ophthalmic imaging data. Experimental results revealed that synthetic data that passed visual Turing tests, can also enhance the representation learning capability of VisionFM, leading to substantial performance gains on downstream ophthalmic AI tasks. Beyond the ophthalmic AI applications developed, validated, and demonstrated in this work, substantial further applications can be achieved in an efficient and cost-effective manner using VisionFM as the foundation.

With the rapid development of the image generation technologies, the malicious abuses of the GAN-generated fingerprint images poses a significant threat to the public safety in certain circumstances. Although the existing universal deep forgery detection approach can be applied to detect the fake fingerprint images, they are easily attacked and have poor robustness. Meanwhile, there is no specifically designed deep forgery detection method for fingerprint images. In this paper, we propose the first deep forgery detection approach for fingerprint images, which combines unique ridge features of fingerprint and generation artifacts of the GAN-generated images, to the best of our knowledge. Specifically, we firstly construct a ridge stream, which exploits the grayscale variations along the ridges to extract unique fingerprint-specific features. Then, we construct a generation artifact stream, in which the FFT-based spectrums of the input fingerprint images are exploited, to extract more robust generation artifact features. At last, the unique ridge features and generation artifact features are fused for binary classification (i.e., real or fake). Comprehensive experiments demonstrate that our proposed approach is effective and robust with low complexities., Comment: 10 pages, 8 figures

In this paper, we try to study the quantum discord dynamics in a complex correlated photon-matter model, which is modified from the Tavis-Cummings-Hubbard model - a common cavity quantum electrodynamics model. The target model consists of two hydrogen atoms. A neutral hydrogen molecule can be obtained through an association reaction and disintegrated through dissociation reaction. The formation and breaking of covalent bond is accompanied by the creation and annihilation of phonon. Compared with previous efforts, studying the quantum discord dynamics of this complicated system is more challenging than it was for the simple quantum system, which consisted of a single two-level atom. For convenience, we adopt a bipartite split of the multiqubit system and the two-qubit von Neumann projective measurement on the observed subsystem. We attempt to examine the dissipative dynamics in open quantum system in addition to the unitary evolution of closed quantum system. We are dedicated to identifying the regularity of quantum correlation as the basis for future research on more complex quantum systems, specifically including the impacts of nuclei tunneling effect, covalent bond formation strength, and dissipation intensities of photon (phonon) on quantum discord., Comment: 20 pages, 14 figures, 2 tables

Based on the two-qubit Jaynes-Cummings model - a common cavity quantum electrodynamics model, and extending to modification of the three-qubit Tavis-Cummings model, we investigate the quantum correlation between light and matter in bipartite quantum systems. By resolving the quantum master equation, we are able to derive the dissipative dynamics in open systems. To gauge the degree of quantum entanglement, some entanglement measurements are introduced: von Neumann entropy, concurrence and quantum discord. In addition, consideration is given to the impacts of initial entanglement and dissipation strength on quantum discord. Finally we discussed two different cases of nuclei motion: quantum and classical., Comment: 14 pages, 9 figures, 1 table

We introduce the association-dissociation model of neutral hydrogen molecule, which is a finite-dimensional cavity quantum electrodynamics model of chemistry with two two-level artificial atoms on quantum dots placed in optical cavities, based on the Tavis-Cummings-Hubbard model. The motion of the nuclei can be represented in quantum form. Electron spin transition and spin-spin interaction between electron and nucleus are both considered. Consideration is also given to the effects of nuclear and electron spins on the formation of neutral hydrogen molecule., Comment: 9 pages, 5 figures. arXiv admin note: substantial text overlap with arXiv:2209.09607

Mobile-centric AI applications have high requirements for resource-efficiency of model inference. Input filtering is a promising approach to eliminate the redundancy so as to reduce the cost of inference. Previous efforts have tailored effective solutions for many applications, but left two essential questions unanswered: (1) theoretical filterability of an inference workload to guide the application of input filtering techniques, thereby avoiding the trial-and-error cost for resource-constrained mobile applications; (2) robust discriminability of feature embedding to allow input filtering to be widely effective for diverse inference tasks and input content. To answer them, we first formalize the input filtering problem and theoretically compare the hypothesis complexity of inference models and input filters to understand the optimization potential. Then we propose the first end-to-end learnable input filtering framework that covers most state-of-the-art methods and surpasses them in feature embedding with robust discriminability. We design and implement InFi that supports six input modalities and multiple mobile-centric deployments. Comprehensive evaluations confirm our theoretical results and show that InFi outperforms strong baselines in applicability, accuracy, and efficiency. InFi achieve 8.5x throughput and save 95% bandwidth, while keeping over 90% accuracy, for a video analytics application on mobile platforms., Comment: IEEE Transactions on Mobile Computing (TMC) 2023

A finite-dimensional chemistry model with two two-level artificial atoms on quantum dots positioned in optical cavities, called the association-dissociation model of neutral hydrogen molecule, is described. The initial circumstances that led to the formation of the synthetic neutral hydrogen molecule are explained. In quantum form, nuclei's mobility is portrayed. The association of atoms in the molecule is simulated through a quantum master equation, incorporating hybridization of atomic orbitals into molecular - depending on the position of the nuclei. Consideration is also given to electron spin transitions. Investigated are the effects of temperature variation of various photonic modes on quantum evolution and neutral hydrogen molecule formation. Finally, a more precise model including covalent bond and simple harmonic oscillator (phonon) is proposed., Comment: 19 pages, 12 figures

Suggestion of modifications of finite-dimensional QED models are proposed for interpreting chemical reactions in terms of artificial atoms and molecules on quantum dots placed in optical cavities. Moving both photons and atoms is possible between the cavities. Super dark states of diatomic systems are described, in which the motion of atoms between cavities is impossible due quantum interference. Chemical processes with two level atoms and three level atoms with lambda spectrum are schematically modeled by solving the single quantum master equation with the Lindblad operators of photon leakage from the cavity and influx into it; association and dissociation reactions then differ only in the initial states. An example is given of the optical interpretation of the transition of an electron from atom to atom in terms of the multilevel Tavis-Cummings-Hubbard model with an estimate of the accuracy. Polyatomic chemical reactions are too complex for accurate modeling. Our method of rough interpretation helps to obtain their long-term results, for example, the form of stationary states of reagents, such as dark and super dark states.