Conversation generation as a challenging task in Natural Language Generation (NLG) has been increasingly attracting attention over the last years. A number of recent works adopted sequence-to-sequence structures along with external knowledge, which s
uccessfully enhanced the quality of generated conversations. Nevertheless, few works utilized the knowledge extracted from similar conversations for utterance generation. Taking conversations in customer service and court debate domains as examples, it is evident that essential entities/phrases, as well as their associated logic and inter-relationships can be extracted and borrowed from similar conversation instances. Such information could provide useful signals for improving conversation generation. In this paper, we propose a novel reading and memory framework called Deep Reading Memory Network (DRMN) which is capable of remembering useful information of similar conversations for improving utterance generation. We apply our model to two large-scale conversation datasets of justice and e-commerce fields. Experiments prove that the proposed model outperforms the state-of-the-art approaches.
MnMX (M = Co or Ni, X = Si or Ge) alloys, experiencing structural transformation between Ni2In-type hexagonal and TiNiSi-type orthorhombic phases, attract considerable attention due to their potential applications as room-temperature solid refrigeran
ts. Although lots of studies have been carried out on how to tune this transformation and obtain large entropy change in a wide temperature region, the crystallography of this martensitic transformation is still unknown. The biggest obstacle for crystallography investigation is to obtain a bulk sample, in which hexagonal and orthorhombic phases coexist, because the MnMX alloys will fragment into powders after experiencing the transformation. For this reason, we carefully tune the transformation temperature to be slightly below 300 K. In that case, a bulk sample with small amounts of orthorhombic phases distributed in hexagonal matrix is obtained. Most importantly, there are no cracks between the two phases. It facilities us to investigate the microstructure using electron microscope. The obtained results indicate that the orientation relationship between hexagonal and orthorhombic structures is [4-2-23]h//[120]o & (01-10)h//(001)o and the habit plane is {-2113.26}h. WLR theory is also adopted to calculate the habit plane. The calculated result agrees well with the measured one. Our work reveals the crystallography of hexagonal-orthorhombic transformation for the first time and is helpful for understanding the transformation-associated physical effects in MnMX alloys.
A variety is finitely based if it has a finite basis of identities. A minimal non-finitely based variety is called limit. A monoid is aperiodic if all its subgoups are trivial. Limit varieties of aperiodic monoids have been studied by Jackson, Lee, Z
hang and Luo, Gusev and Sapir. In particular, Gusev and Sapir have recently reduced the problem of classifying all limit varieties of aperiodic monoids to the two tasks. One of them is to classify limit varieties of monoids satisfying $xsxt approx xsxtx$. In this paper, we completely solve this task. In particular, we exhibit the first example of a limit variety of monoids with countably infinitely many subvarieties. In view of the result by Jackson and Lee, the smallest known monoid generating a variety with continuum many subvarieties is of order six. It follows from the result by Edmunds et al. that if there exists a smaller example, then up to isomorphism and anti-isomorphism, it must be a unique monoid $P_2^1$ of order five. Our main result implies that the variety generated by $P_2^1$ contains only finitely based subvarieties and so has only countably many of them.
Legal judgment prediction(LJP) is an essential task for legal AI. While prior methods studied on this topic in a pseudo setting by employing the judge-summarized case narrative as the input to predict the judgment, neglecting critical case life-cycle
information in real court setting could threaten the case logic representation quality and prediction correctness. In this paper, we introduce a novel challenging dataset from real courtrooms to predict the legal judgment in a reasonably encyclopedic manner by leveraging the genuine input of the case -- plaintiffs claims and court debate data, from which the cases facts are automatically recognized by comprehensively understanding the multi-role dialogues of the court debate, and then learnt to discriminate the claims so as to reach the final judgment through multi-task learning. An extensive set of experiments with a large civil trial data set shows that the proposed model can more accurately characterize the interactions among claims, fact and debate for legal judgment prediction, achieving significant improvements over strong state-of-the-art baselines. Moreover, the user study conducted with real judges and law school students shows the neural predictions can also be interpretable and easily observed, and thus enhancing the trial efficiency and judgment quality.
In this paper, we consider the Cauchy problem to the 3D MHD equations. We show that the Serrin--type conditions imposed on one component of the velocity $u_{3}$ and one component of magnetic fields $b_{3}$ with $$ u_{3} in L^{p_{0},1}(-1,0;L^{q_{0}}(
B(2))), b_{3} in L^{p_{1},1}(-1,0;L^{q_{1}}(B(2))), $$ $frac{2}{p_{0}}+frac{3}{q_{0}}=frac{2}{p_{1}}+frac{3}{q_{1}}=1$ and $3<q_{0},q_{1}<+infty$ imply that the suitable weak solution is regular at $(0,0)$. The proof is based on the new local energy estimates introduced by Chae-Wolf (Arch. Ration. Mech. Anal. 2021) and Wang-Wu-Zhang (arXiv:2005.11906).
Stalactic, taiga, sylvester and Baxter monoids arise from the combinatorics of tableaux by identifying words over a fixed ordered alphabet whenever they produce the same tableau via some insertion algorithm. In this paper, three sufficient conditions
under which semigroups are finitely based are given. By applying these sufficient conditions, it is shown that all stalactic and taiga monoids of rank greater than or equal to $2$ are finitely based and satisfy the same identities, that all sylvester monoids of rank greater than or equal to $2$ are finitely based and satisfy the same identities and that all Baxter monoids of rank greater than or equal to $2$ are finitely based and satisfy the same identities.
We report a mechanical point-contact spectroscopy study on the single crystalline NbGe$_2$ with a superconducting transition temperature $Trm_c$ = 2.0 - 2.1 K. The differential conductance curves at 0.3 K can be well fitted by a single gap s-wave Blo
nder-Tinkham-Klapwijk model and the temperature dependent gap follows a standard Bardeen-Cooper-Schrieffer behavior, yielding $Delta_0 sim$ 0.32 meV and 2$Delta_0$/$krm_{B}$$Trm_{c}$ = 3.62 in the weak coupling limit. In magnetic field, the superconducting gap at 0.3 K keeps constant up to $H_{c1}sim$150 Oe and gradually decreases until $H_{c2}sim$350 Oe, indicating NbGe$_2$ going through a transition from type-I to type-II (possible type-II/1) superconductor at low temperature.
The kagome lattice is host to flat bands, topological electronic structures, Van Hove singularities and diverse electronic instabilities, providing an ideal platform for realizing highly tunable electronic states. Here, we report soft- and mechanical
- point-contact spectroscopy (SPCS and MPCS) studies of the kagome superconductors KV$_3$Sb$_5$ and CsV$_3$Sb$_5$. Compared to the superconducting transition temperature $T_{rm c}$ from specific heat measurements (2.8~K for CsV$_3$Sb$_5$ and 1.0~K for KV$_3$Sb$_5$), significantly enhanced values of $T_{rm c}$ are observed via the zero-bias conductance of SPCS ($sim$4.2~K for CsV$_3$Sb$_5$ and $sim$1.8~K for KV$_3$Sb$_5$), which become further enhanced in MPCS measurements ($sim$5.0~K for CsV$_3$Sb$_5$ and $sim$3.1~K for KV$_3$Sb$_5$). While the differential conductance curves from SPCS are described by a two-gap $s$-wave model, a single $s$-wave gap reasonably captures the MPCS data, likely due to a diminishing spectral weight of the other gap. The enhanced superconductivity probably arises from local strain caused by the point-contact, which also leads to the evolution from two-gap to single-gap behaviors in different point-contacts. Our results demonstrate highly strain-sensitive superconductivity in kagome metals CsV$_3$Sb$_5$ and KV$_3$Sb$_5$, which may be harnessed in the manipulation of possible Majorana zero modes.
A crucial subroutine for various quantum computing and communication algorithms is to efficiently extract different classical properties of quantum states. In a notable recent theoretical work by Huang, Kueng, and Preskill~cite{huang2020predicting},
a thrifty scheme showed how to project the quantum state into classical shadows and simultaneously predict $M$ different functions of a state with only $mathcal{O}(log_2 M)$ measurements, independent of the system size and saturating the information-theoretical limit. Here, we experimentally explore the feasibility of the scheme in the realistic scenario with a finite number of measurements and noisy operations. We prepare a four-qubit GHZ state and show how to estimate expectation values of multiple observables and Hamiltonian. We compare the strategies with uniform, biased, and derandomized classical shadows to conventional ones that sequentially measures each state function exploiting either importance sampling or observable grouping. We next demonstrate the estimation of nonlinear functions using classical shadows and analyze the entanglement of the prepared quantum state. Our experiment verifies the efficacy of exploiting (derandomized) classical shadows and sheds light on efficient quantum computing with noisy intermediate-scale quantum hardware.
Recent work introduced progressive network growing as a promising way to ease the training for large GANs, but the model design and architecture-growing strategy still remain under-explored and needs manual design for different image data. In this pa
per, we propose a method to dynamically grow a GAN during training, optimizing the network architecture and its parameters together with automation. The method embeds architecture search techniques as an interleaving step with gradient-based training to periodically seek the optimal architecture-growing strategy for the generator and discriminator. It enjoys the benefits of both eased training because of progressive growing and improved performance because of broader architecture design space. Experimental results demonstrate new state-of-the-art of image generation. Observations in the search procedure also provide constructive insights into the GAN model design such as generator-discriminator balance and convolutional layer choices.
