Search

Understanding and comprehending video content is crucial for many real-world applications such as search and recommendation systems. While recent progress of deep learning has boosted performance on various tasks using visual cues, deep cognition to reason intentions, motivation, or causality remains challenging. Existing datasets that aim to examine video reasoning capability focus on visual signals such as actions, objects, relations, or could be answered utilizing text bias. Observing this, we propose a novel task, along with a new dataset: Trope Understanding in Movies and Animations (TrUMAn), with 2423 videos associated with 132 tropes, intending to evaluate and develop learning systems beyond visual signals. Tropes are frequently used storytelling devices for creative works. By coping with the trope understanding task and enabling the deep cognition skills of machines, data mining applications and algorithms could be taken to the next level. To tackle the challenging TrUMAn dataset, we present a Trope Understanding and Storytelling (TrUSt) with a new Conceptual Storyteller module, which guides the video encoder by performing video storytelling on a latent space. Experimental results demonstrate that state-of-the-art learning systems on existing tasks reach only 12.01% of accuracy with raw input signals. Also, even in the oracle case with human-annotated descriptions, BERT contextual embedding achieves at most 28% of accuracy. Our proposed TrUSt boosts the model performance and reaches 13.94% performance. We also provide detailed analysis to pave the way for future research. TrUMAn is publicly available at:https://www.cmlab.csie.ntu.edu.tw/project/trope, Comment: CIKM 2021. The first two authors contributed equally to this work

The human ability of deep cognitive skills are crucial for the development of various real-world applications that process diverse and abundant user generated input. While recent progress of deep learning and natural language processing have enabled learning system to reach human performance on some benchmarks requiring shallow semantics, such human ability still remains challenging for even modern contextual embedding models, as pointed out by many recent studies. Existing machine comprehension datasets assume sentence-level input, lack of casual or motivational inferences, or could be answered with question-answer bias. Here, we present a challenging novel task, trope detection on films, in an effort to create a situation and behavior understanding for machines. Tropes are storytelling devices that are frequently used as ingredients in recipes for creative works. Comparing to existing movie tag prediction tasks, tropes are more sophisticated as they can vary widely, from a moral concept to a series of circumstances, and embedded with motivations and cause-and-effects. We introduce a new dataset, Tropes in Movie Synopses (TiMoS), with 5623 movie synopses and 95 different tropes collecting from a Wikipedia-style database, TVTropes. We present a multi-stream comprehension network (MulCom) leveraging multi-level attention of words, sentences, and role relations. Experimental result demonstrates that modern models including BERT contextual embedding, movie tag prediction systems, and relational networks, perform at most 37% of human performance (23.97/64.87) in terms of F1 score. Our MulCom outperforms all modern baselines, by 1.5 to 5.0 F1 score and 1.5 to 3.0 mean of average precision (mAP) score. We also provide a detailed analysis and human evaluation to pave ways for future research., Comment: WWW 2021. The first two authors contributed equally to this work

The Zhaoshan long-baseline Atom Interferometer Gravitation Antenna (ZAIGA) is a new type of underground laser-linked interferometer facility, and is currently under construction. It is in the 200-meter-on-average underground of a mountain named Zhaoshan which is about 80 km southeast to Wuhan. ZAIGA will be equipped with long-baseline atom interferometers, high-precision atom clocks, and large-scale gyros. ZAIGA facility will take an equilateral triangle configuration with two 1-km-apart atom interferometers in each arm, a 300-meter vertical tunnel with atom fountain and atom clocks mounted, and a tracking-and-ranging 1-km-arm-length prototype with lattice optical clocks linked by locked lasers. The ZAIGA facility will be used for experimental research on gravitation and related problems including gravitational wave detection, high-precision test of the equivalence principle of micro-particles, clock based gravitational red-shift measurement, rotation measurement and gravito-magnetic effect., Comment: 20 pages, 15 figures

Vision is a common source of inspiration for poetry. The objects and the sentimental imprints that one perceives from an image may lead to various feelings depending on the reader. In this paper, we present a system of poetry generation from images to mimic the process. Given an image, we first extract a few keywords representing objects and sentiments perceived from the image. These keywords are then expanded to related ones based on their associations in human written poems. Finally, verses are generated gradually from the keywords using recurrent neural networks trained on existing poems. Our approach is evaluated by human assessors and compared to other generation baselines. The results show that our method can generate poems that are more artistic than the baseline methods. This is one of the few attempts to generate poetry from images. By deploying our proposed approach, XiaoIce has already generated more than 12 million poems for users since its release in July 2017. A book of its poems has been published by Cheers Publishing, which claimed that the book is the first-ever poetry collection written by an AI in human history.

We, basing on the quantum critical point (QCP)($p = 0.19 $), propose a phenomenological description on high-$T_c$ superconductivity in the cuprate superconductors, and suggest it divides the whole doping region into two parts: the underdoped region ($p < 0.19 $) and the overdoped region ($ p> 0.19 $). The electrons in the former are localized and form the localized Fermi liquid, a kind of the non-Fermi liquid, where the carriers are the holes; the electrons in the latter are itinerant and form the Fermi liquid, where the carriers are the electrons. We further argue that localization is a prerequisite to the pseudogap; the superconductivity gap and coherence forms at the same time, which both share the same energy scale; coherence induces phonon around the node., Comment: 12 pages, no figures

We investigate the pure penguin process $B^-\to \pi^-\phi$ using QCD factorization approach to calculate hadronic matrix elements to the $\alpha_s$ order in some well-known NP models. It is shown that the NP contributions in R-parity conserved SUSY models and 2HDMs are not enough to saturate the experimental upper bounds for $B\to \phi \pi$. We have shown that the flavor changing $Z^\prime$ models can make the branching ratios of $B\to \phi \pi$ to saturate the bound under all relevant experimental constraints., Comment: No figures

We have analyzed the ${\bar B}^0\to \pi^0\pi^0$ puzzle in three kinds of models beyond the standard model (SM). It is shown that the minimal flavor violation (MFV) models, the minimal supersymmetric standard model (MSSM), and the two Higgs doublet models (2HDM) I and II can not give an explanation of the ${\bar B}^0\to \pi^0\pi^0$ puzzle within $1 \sigma$ experimental bounds and the model III 2HDM can explain the puzzle without a conflict with other experimental measurements. If the constraint on $C_{8g}$ from $b\to s g$ is not imposed, for all kinds of insertions considered there are regions of parameter space, where the scalar quark mass is larger (much larger) than the gluino mass in the case of LR or RL (LL or RR), in which the puzzle can be resolved within $1 \sigma$ experimental bounds., Comment: 7 pages, 3 eps files. Add some more explicit analysis version appeared in PLB

We investigate the branching ratios and direct CP asymmetries of $B_c^+\to D^0\pi^+$ and $B_c^+\to D^+\pi^0$ decays in the PQCD approach. All the diagrams with emission topology or annihilation topology are calculated strictly. A branching ratio of $10^{-6}$ and $10^{-7}$ for $B_c^+\to D^0\pi^+$ and $B_c^+\to D^+\pi^0$ decay is predicted, respectively. Because of the different weak phase and strong phase from penguin operator and two kinds of tree operator contributions, we predict a possible large direct CP violation: $A^{\rm dir}_{cp} (B_c^\pm \to D^0 \pi^\pm)\approx -50%$ and $A^{\rm dir}_{cp} (B_c^\pm \to D^\pm \pi^0)\approx 25%$ when $\gamma=55^\circ$, which can be tested in the coming LHC., Comment: The numerical results have been updated and some errors have been corrected. (10 pages, 16 figures, revtex4)

We study the $B\to \phi K_S$ and $B\to \ep K_S$ decays in MSSM by calculating hadronic matrix elements of operators with QCD factorization approach and including neutral Higgs boson (NHB) contributions. We calculate the Wilson coefficients of operators including the new operators which are induced by NHB penguins at LO using the MIA with double insertions. We calculate the $\alpha_s$ order hadronic matrix elements of the new operators for $B\to \phi K_s$ and $B\to \eta^\prime K_s$. It is shown that the recent experimental results on the time-dependent CP asymmetries in $B\to \phi K_S$ and $B\to \ep K_S$, $S_{\phi K}$ is negative and $S_{\eta^{\prime} K}$ is positive, which can not be explained in SM, can be explained in MSSM if there are flavor non-diagonal squark mass matrix elements of 2nd and 3rd generations whose size satisfies all relevant constraints from known experiments ($B\to X_S\gamma, B_s\to \mu^+\mu^-, B\to X_s \mu^+\mu^-, B\to X_s g, \Delta M_s$, etc.). In particular, we find that one can explain the experimental results with a flavor non-diagonal mass insertion of any chirality when $\alpha_s$ corrections of hadronic matrix elements of operators are included, in contrast with the claim in the literature. At the same time, the branching ratios for the two decays can also be in agreement with experimental measurements., Comment: 21 pages, 23 figures. Some changes have been made, which include comments on contributions to $B\to \phi K_S$ and $B\to \eta' K_S$ from RR and RL insertion and some typing errors

We have studied the neutral Higgs boson (NHB) contributions to the pure penguin process $B\to \phi K_S$ in MSSM with middle and large $\tan\beta$ (say, $>$ 8). We show that the $\alpha_s$ order hadronic matrix elements of NHB induced operators can make sizable effects on both the branch ratio and time dependent CP asymmetry $S_{\phi K}$. Under the all relevant experimental constraints, the Higgs mediated contributions to $S_{\phi K}$ alone can provide a significant deviation from SM and, in particular, lead to a negative $S_{\phi K}$ which is reported by BaBar and Belle., Comment: 7 pages and 6 figures

We investigate the pure penguin decays $B\to \pi\phi$ in the Standard Model(SM) and in the Constrained Minimal Supersymmetric Standard Model (CMSSM) using the QCD factorization approach and consider the Sudakov effects in the twist-3 contribution. We find ${\rm Br}(B^-\to \pi^-\phi) = (1.95-5.70)\times 10^{-9}$ in SM and $(1.1-2.4)\times 10^{-8}$ in CMSSM with large $\tan\beta $ which is about one order of magnitude larger than that in SM., Comment: 10 pages, 1 figure (some errors corrected and results refreshed)

the author considered it unmatured and withdraw it., Comment: the author considered it unmatured and withdraw it