Search

Hyperspectral imagery (HSI) classification may suffer the challenge of hyperspectral-monospectra, where different classes present similar spectra. Joint spatial-spectral feature extraction is a popular solution for the problem, but this strategy tends to inflate accuracy since test pixels may exist in training patches. Domain generalization methods show promising potential, but they still fail to distinguish similar spectra across varying domains, in addition, the theoretical support is usually ignored. In this paper, we only rely on spectral information to solve the hyperspectral-monospectra problem, and propose a Convergence and Error-Constrained Conditional Domain Generalization method for Hyperspectral Imagery Classification (C$^3$DG). The major contributions of this paper include two aspects: the Conditional Revising Inference Block (CRIB), and the corresponding theories for model convergence and generalization errors. CRIB is the kernel structure of the proposed method, which employs a shared encoder and multi-branch decoders to fully leverage the conditional distribution during training, achieving a decoupling that aligns with the generation mechanisms of HSI. Moreover, to ensure model convergence and maintain controllable error, we propose the optimization convergence theorem and risk upper bound theorem. In the optimization convergence theorem, we ensure the model convergence by demonstrating that the gradients of the loss terms are not contradictory. In the risk upper bound theorem, our theoretical analysis explores the relationship between test-time training and recent related work to establish a concrete bound for error. Experimental results on three benchmark datasets indicate the superiority of C$^3$DG.

Many studies have shown that high-energy $\gamma$-photon beams can be efficiently generated via nonlinear Compton scattering driven by laser pulses with intensities $> 10^{22}\rm{W/cm^2}$ recently available in laboratories. Here, we propose a laserless scheme to efficiently generate high-energy polarized $\gamma$-photon beams by collision of two ultrarelativistic electron beams. The self-generated field of a dense driving electron beam provides the strong deflection field for the other ultrarelativistic seeding electron beam. A QED Monte Carlo code based on the locally constant field approximation is employed to simulate the collision process, and the polarization properties of produced $\gamma$ photons are investigated. The simulation results and theoretical analysis indicate that the photon polarization, including both linear and circular polarizations, can be tuned by changing the initial polarization of the seeding beam. If an unpolarized seeding beam is used, linearly polarized photons with an average polarization of 55\% can be obtained. If the seeding beam is transversely (longitudinally) polarized, the linear (circular) polarization of photons above 3 GeV can reach 90\% (67\%), which is favorable for highly polarized, high-energy $\gamma$ photon sources., Comment: 12 pages, 8 figures

Third-order approximate solutions for surface gravity waves in the finite water depth are studied in the context of potential flow theory. This solution provides explicit expressions for the surface elevation, free-surface velocity potential and velocity potential. The amplitude dispersion relation is also provided. Two approaches are used to derive the third order analytical solution, resulting in two types of approximate solutions: the perturbation solution and the Hamiltonian solution. The perturbation solution is obtained by classical perturbation technique in which the time variable is expanded in multiscale to eliminate secular terms. The Hamiltonian solution is derived from the canonical transformation in the Hamiltonian theory of water waves. By comparing the two types of solutions, it is found that they are completely equivalent for the first to second order solutions and the nonlinear dispersion, but for the third order part only the sum-sum terms are the same. Due to the canonical transformation that could completely separate the dynamic and bound harmonics, the Hamiltonian solutions break through the difficulty that the perturbation theory breaks down due to singularities in the transfer functions when quartet resonance criterion is satisfied. Furthermore, it is also found that some time-averaged quantities based on the Hamiltonian solution, such as mean potential energy and mean kinetic energy, are equal to those in the initial state in which sea surface is assumed to be a Gaussian random process. This is because there are associated conserved quantities in the Hamiltonian form. All of these show that the Hamiltonian solution is more reasonable and accurate to describe the third order steady-state wave field. Finally, based on the Hamiltonian solution, some statistics are given such as the volume flux, skewness, and excess kurtosis.

Is chatbot able to completely replace the human agent? The short answer could be - "it depends...". For some challenging cases, e.g., dialogue's topical spectrum spreads beyond the training corpus coverage, the chatbot may malfunction and return unsatisfied utterances. This problem can be addressed by introducing the Machine-Human Chatting Handoff (MHCH), which enables human-algorithm collaboration. To detect the normal/transferable utterances, we propose a Difficulty-Assisted Matching Inference (DAMI) network, utilizing difficulty-assisted encoding to enhance the representations of utterances. Moreover, a matching inference mechanism is introduced to capture the contextual matching features. A new evaluation metric, Golden Transfer within Tolerance (GT-T), is proposed to assess the performance by considering the tolerance property of the MHCH. To provide insights into the task and validate the proposed model, we collect two new datasets. Extensive experimental results are presented and contrasted against a series of baseline models to demonstrate the efficacy of our model on MHCH., Comment: 9pages, 4 figures, accepted by AAAI 2021

The nonliteral interpretation of a text is hard to be understood by machine models due to its high context-sensitivity and heavy usage of figurative language. In this study, inspired by human reading comprehension, we propose a novel, simple, and effective deep neural framework, called Skim and Intensive Reading Model (SIRM), for figuring out implied textual meaning. The proposed SIRM consists of two main components, namely the skim reading component and intensive reading component. N-gram features are quickly extracted from the skim reading component, which is a combination of several convolutional neural networks, as skim (entire) information. An intensive reading component enables a hierarchical investigation for both local (sentence) and global (paragraph) representation, which encapsulates the current embedding and the contextual information with a dense connection. More specifically, the contextual information includes the near-neighbor information and the skim information mentioned above. Finally, besides the normal training loss function, we employ an adversarial loss function as a penalty over the skim reading component to eliminate noisy information arisen from special figurative words in the training data. To verify the effectiveness, robustness, and efficiency of the proposed architecture, we conduct extensive comparative experiments on several sarcasm benchmarks and an industrial spam dataset with metaphors. Experimental results indicate that (1) the proposed model, which benefits from context modeling and consideration of figurative language, outperforms existing state-of-the-art solutions, with comparable parameter scale and training speed; (2) the SIRM yields superior robustness in terms of parameter size sensitivity; (3) compared with ablation and addition variants of the SIRM, the final framework is efficient enough.