Search

Traditional recommender systems such as matrix factorization methods have primarily focused on learning a shared dense embedding space to represent both items and user preferences. Subsequently, sequence models such as RNN, GRUs, and, recently, Transformers have emerged and excelled in the task of sequential recommendation. This task requires understanding the sequential structure present in users' historical interactions to predict the next item they may like. Building upon the success of Large Language Models (LLMs) in a variety of tasks, researchers have recently explored using LLMs that are pretrained on vast corpora of text for sequential recommendation. To use LLMs for sequential recommendation, both the history of user interactions and the model's prediction of the next item are expressed in text form. We propose CALRec, a two-stage LLM finetuning framework that finetunes a pretrained LLM in a two-tower fashion using a mixture of two contrastive losses and a language modeling loss: the LLM is first finetuned on a data mixture from multiple domains followed by another round of target domain finetuning. Our model significantly outperforms many state-of-the-art baselines (+37% in Recall@1 and +24% in NDCG@10) and our systematic ablation studies reveal that (i) both stages of finetuning are crucial, and, when combined, we achieve improved performance, and (ii) contrastive alignment is effective among the target domains explored in our experiments., Comment: RecSys 2024 (Long Paper)

In this paper, we extend Chandrasekhar's method of calculating rotating black holes into $f(R)$ theory. We consider the Ricci scalar is a constant and derive the Kerr and Kerr-Ads metric by using the analytical mathematical method. Suppose that the spacetime is a 4-dimensional Riemannian manifold with a general stationary axisymmetric metric, we calculate Cartan's equation of structure and derive the Einstein tensor. In order to reduce the solving difficulty, we fix the gauge freedom to transform the metric into a more symmetric form. We solve the field equations in the two cases of the Ricci scalar $R=0$ and $R\neq 0$. In the case of $R=0$, the Ernst's equations are derived. We give the elementary solution of Ernst's equations and show the way to obtain more solutions including Kerr metric. In the case of $R\neq 0$, we reasonably assume that the solution to the equations consists of two parts: the first is Kerr part and the second is introduced by the Ricci scalar. Giving solution to the second part and combining the two parts, we obtain the Kerr-Ads metric. The calculations are carried out in a general $f(R)$ theory, indicating the Kerr and Kerr-Ads black holes exist widely in general $f(R)$ models. Furthermore, the whole solving process can be treated as a standard calculation procedure to obtain rotating black holes, which can be applied to other modified gravities.

We investigate the cosmological implications of a phantom dark energy model with bulk viscosity. We explore this model as a possible way to resolve the big rip singularity problem that plagues the phantom models. We use the latest type Ia supernova and Hubble parameter data to constrain the model parameters and find that the data favor a significant bulk viscosity over a non-constant potential term for the phantom field. We perform a dynamical analysis of the model and show that the only stable and physical attractor corresponds to a phantom-dominated era with a total equation of state that can be greater than $-1$ due to the viscosity. We also study the general effect of viscosity on the phantom field and the late time evolution of the universe. We apply the statefinder diagnostic to the model and find that it approaches a nearby fixed point asymptotically, indicating that the universe can escape the big rip singularity with the presence of bulk viscosity. We conclude that bulk viscosity can play an important role in affecting the late-time behavior as well as alleviating the singularity problem of the phantom universe., Comment: accepted by EPJC

We study the accretion of relativistic Vlasov gas onto a Kerr black hole, regarding the particles as distributed throughout all the space, other than just in the equatorial plane. We solve the relativistic Liouville equation in the full $3+1$ dimensional framework of Kerr geometry. For the flow that is stationary and axial symmetric, we prove that the distribution function is independent of the conjugate coordinates. For an explicit distribution that can approximate to Maxwell-J\"{u}ttner distribution, we further calculate the particle current density, the stress energy momentum tensor and the unit accretion rates of mass, energy and angular momentum. The analytic results at large distance are shown to be consistent with the limits of the numerical ones computed at finite distance. Especially, we show that the unit mass accretion rate agrees with the Schwarzschild result in the case of low temperature limit. Furthermore, we find from the numerical results that the three unit accretion rates vary with the angle in Kerr metric and the accretion of Vlasov gas would slow down the Kerr black hole. The closer to the equator, the faster it slows down the black hole., Comment: 28 pages, 11 figures, accepted for publication in Physical Review D

The bumblebee model is an extension of the Einstein-Maxwell theory that allows for the spontaneous breaking of the Lorentz symmetry of the spacetime. In this paper, we study the quasinormal modes of the spherical black holes in this model that are characterized by a global monopole. We analyze the two cases with a vanishing cosmological constant or a negative one (the anti-de Sitter case). We find that the black holes are stable under the perturbation of a massless scalar field. However, both the Lorentz symmetry breaking and the global monopole have notable impacts on the evolution of the perturbation. The Lorentz symmetry breaking may prolong or shorten the decay of the perturbation according to the sign of the breaking parameter. The global monopole, on the other hand, has different effects depending on whether a nonzero cosmological constant presences: it reduces the damping of the perturbations for the case with a vanishing cosmological constant, but has little influence for the anti-de Sitter case., Comment: accepted by Eur. Phys. J. C

Using various latest cosmological datasets including Type-Ia supernovae, cosmic microwave background radiation, baryon acoustic oscillations, and estimations of the Hubble parameter, we test some dark energy models with parameterized equations of state and try to distinguish or select observation-preferred models. We obtain the best fitting results of the six models and calculate their values of the Akaike Information Criteria and Bayes Information Criterion. And we can distinguish these dark energy models from each other by using these two information criterions. However, the $\Lambda $CDM model remains the best fit model. Furthermore, we perform geometric diagnostics including statefinder and Om diagnostics to understand the geometric behaviour of the dark energy models. We find that the six DE models can be distinguished from each other and from $\Lambda$CDM, Chaplygin gas, quintessence models after the statefinder and Om diagnostics were performed. Finally, we consider the growth factor of the dark energy models with comparison to $\Lambda $CDM model. Still, we find the models can be distinguished from each other and from $\Lambda $CDM model through the growth factor approximation., Comment: 23 pages, 6 figures, to be published in CPL

We propose a new model for the viscosity of cosmic matters, which can be applied to different epochs of the universe. Using this model, we include the bulk viscosities as practical corrections to the perfect fluid models of the baryonic and dark matters since the material fluids in the real world may have viscosities due to thermodynamics. Such inclusion is put to the test within the framework of $f(T)$ gravity that is proved to be successful in describing the cosmic acceleration, where $T$ denotes the torsion scalar. We perform an observational fit to our model and constrain the cosmological and model parameters by using various latest cosmological datasets. Based on the fitting result, we discuss several cosmological implications including the dissipation of matters, the evolutionary history of the universe, $f(T)$ modification as an effective dark energy, and the Hubble tension problem. The corresponding findings are (i) The late time dissipation will make the density parameters of the matters vanish in the finite future. Moreover, the density ratio between the baryonic and dark matters will change over time. (ii) The radiation dominating era, matter dominating era and the accelerating era can be recovered and the model can successfully describe the known history of the universe. (iii) The $f(T)$ modification is the main drive of the acceleration expansion and currently mimics a phantom-like dark energy. But the universe will eventually enter a de Sitter expansion phase. (iv) The Hubble tension between local and global observations can be significantly alleviated in our model., Comment: accepted by Eur. Phys. J. C

We investigate the nonrotating neutron stars in $f(T)$ gravity with $f(T)=T+\alpha{T}^2$, where $T$ is the torsion scalar in the teleparallel formalism of gravity. In particular, we utilize the SLy and BSk family of equations of state for perfect fluid to describe the neutron stellar matter and search for the effects of the $f(T)$ modification on the models of neutron stars. For positive $\alpha$, the modification results in a smaller stellar mass in comparison to general relativity, while the neutron stars will contain larger amount of matter for negative $\alpha$. Moreover, there seems to be an upper limit for the central density of the neutron stars with $\alpha>0$, beyond which the effective $f(T)$ fluid would have a steplike phase transition in density and pressure profiles, collapsing the numerical system. We obtain the mass-radius relations of the realistic models of neutron stars and subject them to the joint constraints from the observed massive pulsars PSR J0030+0451, PSR J0740+6620, and PSR J2215+5135, and gravitational wave events GW170817 and GW190814. For the neutron star model in $f(T)$ gravity to be able to accommodate all the mentioned data, the model parameter $\alpha$ needs to be smaller than $-4.295$, $-6.476$, $-4.4$, and $-2.12$ (in the unit of ${G}^2M_\odot^2/c^4$) for SLy, BSk19, BSk20, and BSk21 equations of state, respectively. If one considers the unknown compact object in the event GW190814 not to be a neutron star and hence excludes this dataset, the constraints can be loosened to $\alpha<-0.594$, $-3.5$, $0.4$ and $1.9$ (in the unit of ${G}^2M_\odot^2/c^4$), respectively., Comment: version accepted for publication in EPJC

An exact Kerr-like solution has been obtained recently in Einstein-bumblebee gravity model where Lorentz symmetry is spontaneously broken. In this paper, we investigate the superradiant instability of the Kerr-like black hole under the perturbation of a massive scalar field. We find the Lorentz breaking parameter $L$ does not affect the superradiance regime or the regime of the bound states. However, since $L$ appears in the metric and its effect cannot be erased by redefining the rotation parameter $\tilde{a}=\sqrt{1+L}a$, it indeed affects the bound state spectrum and the superradiance. We calculate the bound state spectrum via the continued-fraction method and show the influence of $L$ on the maximum binding energy and the damping rate. The superradiant instability could occur since the superradiance condition and the bound state condition could be both satisfied. Compared with Kerr black hole, the nature of the superradiant instability of this black hole depends non-monotonously not only on the rotation parameter of the black hole $\tilde{a}$ and the product of the black hole mass $M$ and the field mass $\mu$, but also on the Lorentz breaking parameter $L$. Through the Monte Carlo method, we find that for $l=m=1$ state the most unstable mode occurs at $L=-0.79637$, $\tilde{a}/M=0.99884$ and $M\mu=0.43920$, with the maximum growth rate of the field $\omega_{I}M=1.676\times10^{-6}$, which is about 10 times of that in Kerr black hole., Comment: 18 pages, 3 figures, 1 table, version accepted for publication in Physical Review D

General relativity can be formulated equivalently with a non-Riemannian geometry that associates with an affine connection of nonzero nonmetricity $Q$ but vanishing curvature $R$ and torsion $T$. Modification based on this description of gravity generates the $f(Q)$ gravity. In this work we explore the application of $f(Q)$ gravity to the spherically symmetric configurations. We discuss the gauge fixing and connections in this setting. We demonstrate the effects of $f(Q)$ by considering the external and internal solutions of compact stars. The external background solutions for any regular form of $f(Q)$ coincide with the corresponding solutions in general relativity, i.e., the Schwarzschild-de Sitter solution and the Reissner-Nordstr\"om-de Sitter solution with an electromagnetic field. For internal structure, with a simple model $f(Q)=Q+\alpha Q^2$ and a polytropic equation of state, we find that a negative modification ($\alpha<0$) provides support to more stellar masses while a positive one ($\alpha>0$) reduces the amount of matter of the star., Comment: 21 pages, 3 figures, typos corrected, matching the published version

In the framework of the new version of the gedanken experiments proposed by Sorce and Wald, we investigate the weak cosmic censorship conjecture (WCCC) for an Einstein-Maxwell-Dilaton-Axion (EMDA) black hole. Our result shows that no violations of WCCC can occur with the increase of the background solution parameters for this near-extremal EMDA black hole when the second order correction of the perturbations is taken into account. Namely, the near-extremal EMDA black hole cannot be over-charged or over-spun., Comment: 13 pages, to appear in Mod. Phys. Lett. A

In this paper we generalize the off-shell Abbott-Deser-Tekin (ADT) conserved charge formalism to Palatini theory of gravity with torsion and non-metricity. Our construction is based on the coordinate formalism and the independent dynamic fields are the metric and the affine connection. For a general Palatini theory of gravity, which is diffeomorphism invariant up to a boundary term, we obtain the most general expression for off-shell ADT potential. As explicit examples, we derive the off-shell ADT potentials for Einstein-Hilbert action, the most general $L(g_{\mu\nu}, R^{\lambda}{}_{\nu\alpha\mu}, T^{\lambda}{}_{\alpha\beta}, Q_{\alpha\mu\nu})$ theories and the teleparallel Palatini gravity., Comment: 20 pages, version to be published in Class. Quantum Grav

Is a deep learning model capable of understanding systems governed by certain first principle laws by only observing the system's output? Can deep learning learn the underlying physics and honor the physics when making predictions? The answers are both positive. In an effort to simulate two-dimensional subsurface fluid dynamics in porous media, we found that an accurate deep-learning-based proxy model can be taught efficiently by a computationally expensive finite-volume-based simulator. We pose the problem as an image-to-image regression, running the simulator with different input parameters to furnish a synthetic training dataset upon which we fit the deep learning models. Since the data is spatiotemporal, we compare the performance of two alternative treatments of time; a convolutional LSTM versus an autoencoder network that treats time as a direct input. Adversarial methods are adopted to address the sharp spatial gradient in the fluid dynamic problems. Compared to traditional simulation, the proposed deep learning approach enables much faster forward computation, which allows us to explore more scenarios with a much larger parameter space given the same time. It is shown that the improved forward computation efficiency is particularly valuable in solving inversion problems, where the physics model has unknown parameters to be determined by history matching. By computing the pixel-level attention of the trained model, we quantify the sensitivity of the deep learning model to key physical parameters and hence demonstrate that the inversion problems can be solved with great acceleration. We assess the efficacy of the machine learning surrogate in terms of its training speed and accuracy. The network can be trained within minutes using limited training data and achieve accuracy that scales desirably with the amount of training data supplied.

The spectral energy distribution (SED) sequence for type Ia supernovae (SN Ia) is modeled by an artificial neural network. The SN Ia luminosity is characterized as a function of phase, wavelength, a color parameter and a decline rate parameter. After training and testing the neural network, the SED sequence could give both the spectrum with wavelength range from 3000\AA~to 8000\AA~ and the light curve with phase from 20 days before to 50 days after the maximum luminosity for the supernovae with different colors and decline rates. Therefore, we call this the Artificial Neural Network Spectral Light Curve Template (ANNSLCT) model. We retrain the Joint Light-curve Analysis (JLA) supernova sample by using the ANNSLCT model and obtain the parameters for each supernova to make a constraint on the cosmological $\Lambda$CDM model. We find that the best fitting values of these parameters are almost the same as those from the JLA sample trained with the Spectral Adaptive Lightcurve Template 2 (SALT2) model. So we believe that the ANNSLCT model could be used to analyze a large number of SN Ia multi-color light curves measured in the current and future observational projects., Comment: 8 pages two columns, 17 figures

The phantom dark energy remarkably boosts our prehension of the accelerating Universe. Various models are widely discussed in the phantom Universe without bulk viscosity. From the hydrodynamics' point of view, it is natural to introduce the nonperfect fluid in the study of the Universe, as an ideal fluid just an approximation to the real world after all and using the generalized equation of state (EoS) with bulk viscosity, the early inflationary universe and the accelerated expansion of the late-time universe are described by many authors. In this paper, in order to investigate how the viscosity will influence the evolution of the Universe, we study a class of phantom dark energy models with bulk viscosity by the method of dynamical analysis technique. We show that there are different cosmic late-time behaviors and the stability also brings some constraints on the models. We also plot the evolutionary trajectories of this model in the statefinder parameter-planes to see the different behaviors of the models from the statefinder viewpoint., Comment: 14 pages, 7 figures

Recently, it was found that after imposing a helix boundary condition on a scalar field, the Casimir force coming from the quantum effect is linearly proportional to $r$, which is the ratio of the pitch to the circumference of the helix. This linear behavior of the Casimir force is just like that of the force obeying the Hooke's law on a spring. In this paper, inspiring by some complex structures that lives in the cells of human body like DNA, protein, collagen etc., we generalize the helix boundary condition to a more general one, in which the helix consists of a tiny helix structure, and makes up a hierarchy of helix. After imposing this kind of boundary condition on a massless and a massive scalar, we calculate the Casimir energy and force by using the so-called zeta function regularization method. We find that the Hooke's law with the generalized helix boundary condition is not exactly the same as usual one. In this case, the force is proportional to the cube of $r$ instead. So we regard it as a generalized Hooke's law, which is complied by a \emph{generalized quantum spring}., Comment: 10 pages, 1 figure

Using the solution phase space method, we investigate the thermodynamics of black holes in Einstein-aether-Maxwell theory, for which the traditional Wald method (covariant phase space method) fails. We show the first laws of thermodynamics and definitive entropy expressions at both Killing and universal horizons for some examples of exact black hole solutions, including 3-dimensional static charged quasi-BTZ black hole, two 4-dimensional static charged black holes and 3-dimensional rotating solution. At Killing horizons the entropies are exactly one quarter of the horizon area, but at universal horizons of 3-dimensional black holes, the entropies have a corrected term in addition to the one proportional to the horizon area., Comment: 23 pages, version appeared in Class. Quantum Grav

Conformal invariance can ameliorate or eliminate the singularities residing in the black holes, and may still exist in the strong gravity regimes close to these black holes. In this paper, we try to probe this conformal invariance by looking into the wave absorption and scattering by the nonsingular static spherical black holes. The partial and total absorption cross section, as well as the differential scattering cross section, are presented for black holes with different choices of conformal parameters. Although the photon trajectories are unchanged from the Schwarzschild case since the spacetimes are conformally related, the wave optics are affected by the conformal parameters. As a result, the absorption of waves generally increases with the conformal parameters, while the shadow of the black holes remains the same as the Schwarzschild case. Moreover, the peaks in the oscillatory pattern of scattering shift towards smaller observing angles as the conformal parameters grows, while the widths of the glory peaks do not show sensitive dependence. The unique signature of the wave absorption and scattering by the nonsingular static spherical black holes in conformal gravity thus can serve to distinguish themselves from the Schwarzschild in the low frequency regime, and from other spherical black holes of alternative gravities in the high frequency limit and glory peaks., Comment: 21 pages, 12 figures, title, abstract and some contents changed, an appendix added

While a scalar field with a pole in its kinetic term is often used to study the cosmological inflation, it can also play the role of dark energy, which is called the pole dark energy model. We propose a generalized model that the scalar field may have two or even multiple poles in the kinetic term and we call it the multi-pole dark energy. We find the poles can place some restrictions on the values of the original scalar field with non-canonical kinetic term. After transforming to the canonical form, we get a flat potential for the transformed new scalar field even if the original field has a steep one. It The late-time evolution of the universe is obtained explicitly for the two pole model, while dynamical analysis is performed for the multiple pole model. We find that it does have a stable attractor solution, which corresponds to the universe dominated by the potential of the scalar field., Comment: 11 pages, 7 figures

The restoration of spin connection clarifies the long known local Lorentz invariance problem in telelparallel gravities. It is considered now that any tetrad together with the associated spin connection can be equally utilized. Among the tetrads there is a particular one, namely proper tetrad, in which all the spurious inertial effects are removed and the spin connection vanishes. A specific tetrad was proposed in the literature for spherically symmetric cases, which has been used in regularizing the action, as well as in searching solutions in various scenarios. We show in this paper that the this tetrad is not the unique choice for the proper tetrad. We construct a new tetrad that can be considered as the proper one, and it will lead to different behaviors of the field equation and results in different solutions. With this proper tetrad, it is possible to find solutions to teleparallel gravities in the strong field regime, which may have physical applications. In the flat spacetime limit, the new tetrad coincides with the aforementioned one., Comment: 6 pages

In this paper we elevate the formalism for off-shell Abbott-Deser-Tekin (ADT) conserved charges in the presence of arbitrary matter fields to including the internal gauge transformation, when the gauge fields are present. For this purpose, we resort to exact symmetry, and the symmetry generator is combined by diffeomorphism and internal gauge transformation. For universality, we consider an apparently non-covariant Lagrangian, which contains a matter Chern-Simons term. We show that the elevated off-shell ADT formalism is equivalent to the covariant phase space method and the Barnich-Brandt-Comp\`ere (BBC) formalism. To check the validity of our method, we explicitly compute the conserved charges of general non-extremal rotating charged G\"odel black holes in minimal five-dimensional gauged supergravity, reproducing the previously known results., Comment: 23 pages; spelling corrected for some author's names in citations

Wormholes are hypothetical tunnels that connect remote parts of spacetime. In General Relativity, wormholes are threaded by exotic matter that violates the energy conditions. In this work, we consider wormholes threaded by nonexotic matter in nonminimal torsion-matter coupling $f(T)$ gravity. We find that the nonminimal torsion-matter coupling can indeed hold the wormhole open. However, from geometric point of view, for the wormhole to have asymptotic flatness, the coupling matter density must falloff rapidly at large radius, otherwise the physical wormhole must be finite due to either change of metric signature or lack of valid embedding. On the other hand, the matter source supporting the wormhole can satisfy the null energy condition only in the neighborhood of the throat of the wormhole. Therefore, the wormhole in the underlying model has finite sizes and cannot stretch to the entire spacetime., Comment: 6 pages,4 figures

Recently, a new kind of $f(z)$ theory is proposed to provide a different perspective for the development of reliable alternative models of gravity in which the $f(R)$ Lagrangian terms are reformulated as polynomial parameterizations $f(z)$. In the previous study, the parameters in the $f(z)$ models have been constrained by using cosmological data. In this paper, these models will be tested by the observations in the solar system. After solving the Ricci scalar as a function of the redshift, one could obtain $f(R)$ that could be used to calculate the standard Parameterized-Post-Newtonian (PPN) parameters. First, we fit the parametric models with the latest cosmological observational data. Then the tests are performed by solar system observations. And last we combine the constraints of solar system and cosmology together and reconstruct the $f(R)$ actions of the $f(z)$ parametric models., Comment: 15 pages

There is undetermined potential function $V(\phi)$ in the action of mimetic gravity which should be resolved through physical means. In general relativity(GR), the static spherically symmetric(SSS) solution to the Einstein equation is a benchmark and its deformation also plays a crucial role in mimetic gravity. The equation of motion is provided with high nonlinearity, but we can reduce primal nonlinearity to a frequent Riccati form in the SSS case of mimetic gravity. In other words, we obtain an expression of solution to the functional differential equation of motion with any potential function. Remarkably, we proved rigorously that there is a zero point of first order for the metric function $\beta(r)$ if another metric function $\alpha(r)$ possesses a pole of first order within mimetic gravity. The zero point theorem may be regarded as the generalization of Birkhoff theorem $\alpha\beta=1$ in GR. As a corollary, we show that there is a modified black hole solution for any given $V(\phi)$, which can pass the test of solar system. As another corollary, the zero point theorem provides a dynamical mechanism for the maximum size of galaxies. Especially, there are two analytic solutions which provide good fits to the rotation curves of galaxies without the demand for particle dark matter., Comment: 11 pages

It is known that a massive charged scalar field can trigger a superradiant instability in the background of a Kerr-Newman black hole. In this paper, we present a numerical study of such an instability by using the continued fraction method. It is shown that for given a black hole, the unstable scalar mode with a specific azimuthal index $m$ only occurs in a finite region in the parameter space of the scalar field. The maximum mass of the scalar cloud is exactly the upper bound of the mass of the unstable modes. We show that due to the electromagnetic interaction between the scalar field and the Kerr-Newman black hole, the growth rate of the instability can be $15.7\%$ larger than that of a scalar field in Kerr spacetime of the same rotation parameter. In addition, we find a maximum value of the growth rate $\tau^{-1}=1.788\times 10^{-7}M^{-1}$, which is about $4\%$ larger than that in the Kerr case., Comment: 8 pages, 5 figures, accepted for publication in PRD

The currently accelerated expansion of our Universe is unarguably one of the most intriguing problems in today's physics research. Two realistic non-minimal torsion-matter coupling $f(T)$ models have been established and studied in our previous papers [Phys. Rev. D92, 104038(2015) and Eur. Phys. J. C77, 504(2017)] aiming to explain this "dark energy" problem. In this paper, we study the generalized power-law torsion-matter coupling $f(T)$ model. Dynamical system analysis shows that the three expansion phases of the Universe, i.e. the radiation dominated era, the matter dominated era and the dark energy dominated era, can all be reproduced in this generalized model. By using the statefinder and $Om$ diagnostics, we find that the different cases of the model can be distinguished from each other and from other dark energy models such as the two models in our previous papers, $\Lambda$CDM, quintessence and Chaplygin gas. Furthermore, the analyses also show that all kinds of generalized power-law torsion-matter coupling model are able to cross the $w=-1$ divide from below to above, thus the decrease of the energy density resulting from the crossing of $w$ will make the catastrophic fate of the Universe avoided and a de Sitter expansion fate in the future will be approached., Comment: 9 pages,12 figures

We construct and train an artificial neural network called the back-propagation neural network to describe the evolution of the type Ia supernova spectrum by using the data from the CfA Supernova Program. This network method has many attractive features, and one of them is that the constructed model is differentiable. Benefitting from this, we calculate the absorption velocity and its variation. The model we constructed can well describe not only the spectrum of SNe Ia with wavelength range from $3500\AA$ to $8000\AA$, but also the light-curve evolution with phase time from $-15$ to $50$ with different colors. Moreover, the number of parameters needed during the training process is much less than the usual methods., Comment: v2 refs added, two columns, 7pages, 12 figures